EPE_07-1999

7/23/2019 EPE_07-1999

http://slidepdf.com/reader/full/epe07-1999 1/69

EPE Online, Febuary 199 9 - www.epemag.com - XXX

http://0799b.pdf/

http://0799b.pdf/

http://0799b.pdf/

http://0799b.pdf/

7/23/2019 EPE_07-1999


EPE Online, July 1999 - www.epemag.com - 661EPE Online, July 1999 - www.epemag.com - 661

PROJECTS AND CIRCUITS

PRACTICAL OSCILLATOR DESIGN - Part 1Hartley and its variants by Raymond Haigh

Worked examples and circuit info for hands-on constructors.

700

CIRCUIT SURGERY by Alan Winstanley and Ian Bell Gate Post - SCR and triac triggering; Stripboard and High Voltages; Current and Dielectrics 695

NETWORK - THE INTERNET PAGE surfed by Alan Winstanley It’s a Hoax; Fall-out; Stuck on the Ramp; Links.

716

REGULARS AND SERVICES

INNOVATIONS - Barry Fox Highlights technology’s leading edge.

Plus everyday news from the world of electronics.718

READOUT - John Becker addresses general points arising. 724

SHOPTALK with David Barrington The essential guide to component buying for EPE Online projects.

722

EDITORIAL 662

SERIES AND FEATURES

LED STROBOSCOPE by Robert Penfold Another low-cost Starter Project - how to freeze the motion of small machines.

670

NEW TECHNOLOGY UPDATE by Ian Poole New processes provide improved thick film performance.

693

INTRUDER DETERRENT by Bart Trepak Is there anyone at home? Would-be intruders may never be sure if thelights keep switching.

686

12V Battery Tester by Terry de Vaux-Balbirnie Give a probing health check to your lead-acid batteries. 664

EPE MOOD PICKLER by Andy Flind Oh for good night’s sleep! Insomniacs rejoice - your wakeful nights could beover with this mini-micro under the pillow.

678

INGENUITY UNLIMITED hosted by Alan Winstanley Superior Heads-Tails indicator; Home Alarm System.

675

PRACTICALLY SPEAKING by Robert Penfold Making better connections with sockets and switches.

712

http://0799m.pdf/

http://0799m.pdf/

http://0799m.pdf/

http://0799m.pdf/

http://0799l.pdf/

http://0799l.pdf/

http://0799l.pdf/

http://0799l.pdf/

http://0799o.pdf/

http://0799o.pdf/

http://0799o.pdf/

http://0799p.pdf/

http://0799p.pdf/

http://0799p.pdf/

http://0799p.pdf/

http://0799p.pdf/

http://0799p.pdf/

http://0799r.pdf/

http://0799r.pdf/

http://0799r.pdf/

http://0799r.pdf/

http://0799r.pdf/

http://0799r.pdf/

http://0799q.pdf/

http://0799q.pdf/

http://0799q.pdf/

http://0799q.pdf/

http://0799q.pdf/

http://0799q.pdf/

http://0799d.pdf/

http://0799d.pdf/

http://0799g.pdf/

http://0799g.pdf/

http://0799g.pdf/

http://0799k.pdf/

http://0799k.pdf/

http://0799k.pdf/

http://0799j.pdf/

http://0799j.pdf/

http://0799j.pdf/

http://0799j.pdf/

http://0799f.pdf/

http://0799f.pdf/

http://0799f.pdf/

http://0799i.pdf/

http://0799i.pdf/

http://0799i.pdf/

http://0799i.pdf/

http://0799h.pdf/

http://0799h.pdf/

http://0799h.pdf/

http://0799n.pdf/

http://0799n.pdf/

http://0799n.pdf/

http://0799d.pdf/

http://0799q.pdf/

http://0799r.pdf/

http://0799p.pdf/

http://0799d.pdf/

http://0799o.pdf/

http://0799n.pdf/

http://0799m.pdf/

http://0799l.pdf/

http://0799k.pdf/

http://0799j.pdf/

http://0799i.pdf/

http://0799h.pdf/

http://0799g.pdf/

http://0799f.pdf/

http://0799j.pdf/

http://0799i.pdf/

http://0799g.pdf/

http://0799f.pdf/

7/23/2019 EPE_07-1999


EPE Online, Febuary 1999 - www.epemag.com - XXX

As publishers of an Internet-based magazine, we watch with interest how variousgovernments around the world are trying to come to terms with the Internet and itsimpact -- both positive and negative -- on the social structure of the population.

Australia this week took steps to censor the Internet by passing a bill in the Senate. Thisbill outlines a complaints-based scheme under which the Australian BroadcastingAgency (ABA) can force Australia-based Internet providers to remove any material thatwould be considered offensive or illegal under their film and video guidelines.

If the material, which is considered offensive, is not removed within one working day,ISP’s face penalties of tens of thousands of dollars.

With the current proposal, International sites (estimated to be some 90million) would notbe blocked. Adult sites based in Australia would effortlessly evade the bill by movingoffshore or underground, thereby circumventing the purpose of the new legislation.

Angry Internet service providers have turned on the Austrian Government by divertingrequests from Government computer users to a protest page, which makes these userswait 120 seconds before they can reach their desired destination, while anti-censorshipgroups have organized national rallies in the real world.

In contrast, the Canadian Broadcast Regulation Agency (CRTC) wrapped up hearingsinto Internet censorship and Canadian cultural representation in cyberspace. Thecommission decided that it would be pointless to attempt any Internet regulation at thistime. The US has also grappled with this subject over the past several years withoutcoming to any meaningful conclusions.

The challenge to free speech has always been central to the argument of any formcensorship. How we balance a free exchange of ideas using a global network and cometo terms with cultural and social standards that vary by country and continent will be thesubject of much debate as we enter the new millenium.

EPE Online, July 1999 - www.epemag.com - 662

http://0799e.pdf/

7/23/2019 EPE_07-1999



8-CHANNEL ANALOG DATA LOGGER

There are three reasons why the author designedthis Data Logger:

o) To get to know more about Microchip's newPIC16F87x family.

o) To monitor the EPE Online Musical Sundial(June '99) and record sunlight conditions.

o) Several readers had suggested that one shouldbe published

The PIC16F87x family are much more powerfulthan the familiar '84 devices: up to eight channels ofADC; serial communications I/O at controllable baudrates; enlarged program and on-chip EEPROM mem-ories; 20MHz maximum clock rate; external serial

data memory read/write.

It is some of these attributes that are put to use inthis Data Logger, whose extensive specs include:

o) Up to eight channels of analog data input and 10-bit digital conversion.

o) Data storage/retrieval using on-board serialmemory (2 megabits max).

o) Sampling rates: 0 5 secs to 62 secs, plus external

clock option Automatic non-volatile storage ofcurrent sampling count value and rate when log-ging session ends.

o) LCD display of sampled data, elapsed time andcount value.

o) Serial transfer of recorded data to PC at 9600baud.

o) Transferred data formatted and stored on Drive Cas:

Eight uncorrected binary files.

Eight ASCII-converted numeric-value textfiles.

Microsoft Excel tabbed composite numeric-text file.

Four dual-channel binary files for the EPE Vir- tual Scope (Jan & Feb '98).

o) Prototype has been run under DOS and Windows3.1/95, on a 20MHz '386 up to 120MHz Pentium(Win '98 compatibility unknown).

o) Analog signal input range 0V to 5V

A follow-up article (PIC16F87x Mini Tutorial )takes a closer look at how the PIC16F87x family canbe programmed to implement the type of functionsoffered by the Data Logger.

Microchip tell us that the long-awaitedPIC16F87x devices will be on sale from the end ofJune.

FROM PIPELINES TO PYLONS

Readers in the UK are fortunate enough to enjoyvirtually uninterrupted electricity, provided by theworld's largest interconnected electrical system,which links our power stations together to form theNational Grid. The high quality of Britain's electricitysupply is taken for granted by the British, althoughfor both the micro-electronics enthusiast as well asthe general public there is much mystique surround-ing the way in which electrical power is created anddelivered safely to their homes.

In this two-part feature, supported by the exper-tise of the international power generation companyNational Power plc, Alan Winstanley describes someof the high technology involved in generating power

from a gas pipeline to the turbines and generators

and then to the electricity pylon and beyond! We alsoexamine in close-up some of the techniques relatedto the provision of a 230V supply directly to housing

and industry.If ever you have wondered what “neutral” really

means, why the earth plays such a vital role insafety, or why an electricity power station would everneed gas, or if you just want to brush up on somefundamental theory, this article provides backgroundwhich is essential reading for electronics users andconsumers everywhere.

SMT ULTRASONIC PUNCTURE FINDER

This device will f ind most bicycle punctures inrecord time, eliminating the messy business of im-

mersing the tube in water. All but the very slowestpunctures produce a strong ultrasonic signal and canbe located to within a few millimeters. It has manyother uses as a low cost ultrasound receiver.

http://0799f.pdf/

7/23/2019 EPE_07-1999



Twelve-volt lead-acid batter-ies are now found in many walksof life. The most familiar is, ofcourse, the type used in cars.However, smaller ones are usedfor mobile radio rigs, alarm sys-tems, solar-charged circuits, and

other specialized applications.The project to be described

here is an instrument that will de-termine the charge state of sucha battery.

OVERVIEW The 12V Battery Tester is

built in the style of a logic probe,which gives a neat appearance(see photographs) and also al-lows for easy use with one hand.

The display is provided by a rowof LEDs (l ight-emitting diodes)giving a simple “Low”, “Medium”,and “High” readout.

There is also a “Crank Test”display LED for checking car bat-teries under a high load. Thisidentifies a failing battery andprovides a similar result to thedevice used for the purpose inservice centers. It also doubles asa “danger level” signal.

It is important to note thatthis circuit is only suitable fortesting 12V lead-acid batteries.It will not give accurate resultswith any other type.

If you are a motorist and do alot of start-stop driving (thereforeusing the starter motor exces-sively), drive mostly with theheadlights switched on, and use

In this was a sealed glasscapsule often marked with red,yellow, and green colored sec-tors. The higher the capsulefloated in the liquid, the greaterwas its density (“specific grav-ity”) and the higher was thestate of charge. The coloredbands were arranged to give asimple “poor”, “medium”, and“good” indication.

Hydrometers are still avail-able, but they cannot be usedwith most batteries because theelectrolyte is not accessible.This device aims to give a read-out which is just as simple butdoes the job with much lessfuss.

CHARGE CHECK

The state of charge of abattery may be determined bymeasuring its terminal voltage.This falls with loss of charge ina more-or-less linear way. Fig.1shows a graph of the voltage ofa nominal 12V lead-acid batteryagainst charge state. Althoughthere will be small variations,

A “health” check for lead-acid batteries.

the heated rear windscreen forlong periods, the battery maysoon lose its charge. This prob-lem will be aggravated if it isnearing the end of its servicelife. It would then be a goodidea to use this instrument ev-ery so often and re-charge thebattery before it lets you down.

The user of a small batteryfor some special application of-ten does not know, until theequipment fails, that it has gone“flat”. Not only is this inconve-nient but it may damage thebattery.

This is because leaving it ina poor state of charge (even ina half-charged condition) willcause deterioration over time.

This will result in reduced ca-pacity and service life. Havingthis instrument available will en-able you to check the conditionof the battery as frequently asyou wish.

MEASURING CHARGE Many lead-acid batteries

used today are of the “sealed-for-life” variety. This makestopping-up with distilled water

unnecessary and (depending ontype) allows them to be used inany orientation.

Some years ago, every self-respecting user had a “batteryhydrometer”. This consisted of arubber bulb which, whensqueezed and released, alloweda sample of battery acid to bedrawn into a glass tube.

.

.

.

.

.

.

Fig.1. Graph of the voltage of a nominal 12V lead-acid battery against charge state.

7/23/2019 EPE_07-1999



this is largely independent ofthe physical size or-manufacturer.

Note that, over the usefulrange of charge, the difference

is only 1V or so. It is importantto also note that “zero” charge

means the practical end point

not true zero! A battery dis-charged below this level is likelyto suffer irreversible damageeven though it would still be ca-pable of delivering current.

In this circuit, the voltage(hence charge) is measured us-

ing a set of LEDs one each,

Green, Yellow, Orange, andRed. These show though holes

in the top of the box. The firstthree are grouped together butthe red is slightly displaced be-cause it performs a differentfunction (the crank test or dan-ger level signal).

To use the instrument, aflying lead is clipped on to the

negative ( ) battery terminal.

Holding the box in one hand,the probe is now touched on tothe positive (+) terminal and theLEDs observed.

Some, or all of them, shouldlight up according to the state ofcharge. Thus, all LEDs on signi-fies “high”, all except green“medium”, only orange and red“low”.

Disregarding the crank testfor the moment, the red LEDwill always be on unless the bat-tery is seriously discharged. If itis off (“danger level”) the batterymust be charged urgently and,

depending on type, it may neverrecover its full capacity.

OPERATING POINTS Rather than to indicate

“full”, “half”, and “zero” charge,it was decided to provide the“good” point at around 80 per-cent, but the “low” one towards

the end of the useful remainingcharge. The orange LED there-fore represents a “charge now”signal.

However, the green one will

be on even when the battery isslightly discharged. This pre-vents it going off almost straightaway after a period of use. Ex-perience with the particular ap-plication will soon show howthese operating points need tobe interpreted.

Some manufacturers statethat re-charging must be carriedout on their batteries when thevoltage falls to 12V. Others al-

low it to fall to, say, 11 7V. Tak-

ing this into account, these arethe selected operating points:

High 12·6V

Medium 12·3V

Low 12·0V

Crank Test 9·8V

For ease of construction,these operating points are pre-set and cannot be altered un-

less you are competent at re-calculating the resistor values ina potential divider chain.

HOW IT WORKS The complete circuit dia-

gram for the 12V Battery Testeris shown in Fig.2. The principlecomponent is a quadruple bipo-lar opamp (operational ampli-fier), IC2, which contains fouridentical units in a single 14-pin

package. The individualopamps are referred to in thetext and in the diagram as IC2ato IC2d. The power supply tothe circuit is obtained from thebattery “on test” via fuse FS1.

All four opamp inverting in-puts (pins 2, 6, 9 and 13) areconnected together and, in turn,connected to the “regulation”

(Reg) pin of IC1, a 5V voltagereference device. This behavesrather like a Zener diode, and isconnected and drawn as such.

However, it behaves with

much greater precision. A voltageequal to or very close to 5V willtherefore appear at all the invert-ing inputs of IC2. IC1 requires acertain small reverse current toflow through it to allow regulationto take place and this is the pur-pose of resistor R6. Correct oper-ation will be maintained down to alevel much lower than that of abattery at its end point.

CHAIN REACTION The non-inverting inputs of

IC2a to IC2d (pins 3, 5, 10 and 12respectively) are connected tovarious points along the potentialdivider chain, made up of resis-tors R1 to R5. The ends of thechain are connected across thesupply so that, according to theindividual resistor values, thevoltage appearing at these pointswill be a known fraction of thebattery voltage.

They have been selected toprovide 5V when the terminalvoltage of the battery is at the op-erating points. Thus, 5V will ap-pear at the non-inverting input of

IC2d with a supply of 12 6V, at

IC2c with 12

and IC2a with 9

3V, IC2b at 12 0V

8V.

With increasing supply volt-age (and therefore batterycharge), the non-inverting supplyvoltage of IC2a to IC2d will there-fore exceed that at the inverting

ones. As this happens, the corre-sponding output will go high andthe LED (D1 to D4) associatedwith it will operate. Each LED hasits operating current limited byone of resistors R7 to R10 to anominal 15mA.

Since the operating pointsmust be known with a fair degreeof accuracy, the resistors used in

7/23/2019 EPE_07-1999



the potential divider chain are 1per cent tolerance types. Notonly do these have accurately-known values, but they willmaintain them over time.

A fuse, FS1, is included toprotect the tester in the event of

an error which may result in ashort-circuit to the supply. A lead-acid battery can deliver a veryhigh current under such condi-tions and this could result inwiring or PCB tracks melting. Thefuse should protect the circuit in

the event of in-correct connec-tion and morewill be saidabout this later.

The usualmethod of in-cluding a diodefor reverse-polarity protec-tion is inappro-priate here dueto its forwardvoltage drop.This, beingslightly voltage-

dependent, would change theoperating characteristics of thecircuit in an unpredictable wayas the various LEDs came on.

CONSTRUCTION All the components for the

12V Battery Tester , except thefuse, are mounted on a singlesmall printed circuit board(PCB), which has been de-signed to fit the specified probebox. The component layout and(approximately) full-size copperfoil master are shown in Fig. 3and photographs. This board isavailable from the EPE Online Store (code 7000234) at

www.epemag.com

Referring to the componentlayout in Fig.3, drill the twomounting holes then solder IC2

Ω

Ω

Ω

Ω

Ω

Ω

–

Fig.2. Complete circuit diagram for the 12V Battery Tester.

NOTES ON USE

In any particular application,experience will best decide howto interpret the operating points.However, the following hintsshould help those using the in-strument for the first time.

1. Always connect the unit withthe correct polarity.

1. Do not perform a test with acharger connected. In fact,allow the battery to rest forat least one hour after charg-ing to allow the voltage tostabilize. This is because itfalls slightly with time.

1. Remove any load and allowthe battery to rest for a whilebefore testing. This is be-cause the voltage risesslightly, so failing to do thiswould result in an incorrectreading.

Most of the change happensin the first few minutes. How-ever, best results will be ob-tained if the battery is left idlefor at least one hour.

http://www.epemag.com/




7/23/2019 EPE_07-1999



socket in position (but do not insert the IC itself at thisstage). Follow by soldering all other components inposition except IC1 and the LEDs. Take special careto solder each resistor R1 to R5 into its correct posi-tion or the operating points will be incorrect.

Taking extra care with the polarity, solder theLEDs in position so that their tops stand about10mm above the circuit board. They should belevel and in a straight line or it will result in a poorappearance to the finished instrument. Note thatthe cathode end (k) is labeled in each case (thishaving the slightly shorter end lead).

Before handling IC1, observe some precau-tions to prevent possible damage to it by static

charge, which mayexist on the body.The simplestmethod is to touch

a water tap imme-diately before un-packing it andtouching the pins.This will earth thebody and allowany charge to flowaway harmlessly.

Refer to thepinout details inFig.3 and look atthe flat face ofIC1. Cut off the

left-hand end wire

COMPONENTS Resistors

R1 22kR2 4k3R3 470 ohmsR4 430 ohmsR5 18kR6 3k9R7, R8, R9, R10 680 ohms (4 off)

See also theSHOP TALK Page!

R1 to R5 must be of the 0.6W metalfilm type having 1% tolerance. Allother resistors may be of the 5%0.5W carbon film type.

SemiconductorsD1 3mm red LEDD2 3mm orange LEDD3 3mm yellow LED

D4 3mm green LEDIC1 REF50Z 5V reference voltageIC2 LM324N quad opamp

MiscellaneousFS1 20mm chassis mounting fuseholder, with 500mA 20mm quickblow fuse.

Printed circuit board availablefrom the EPE Online store , code7000234 (www.epemag.com );logic probe box (with probe), size130mm x 34mm x 30mm approx;14-pin IC socket; multi-strandedconnecting wire; extra flexible wire;

self-locking cable tie; crocodile clip(with black insulation cover) to suitthe application; 12.7mm plasticspacers (2 off); solder, etc.

$24Approx. Cost Guidance Only

–

REF

Fig.3. Printed circuit board copper foil master pat-

tern and topside component layout,

together with the two “test”

leads.

Layout of components on the completed circuit board. The method of mounting the PCB on the removable panel and posi-

tioning of the fuseholder is shown below.

http://0799q.pdf/

http://0799q.pdf/








http://0799q.pdf/

http://0799q.pdf/


7/23/2019 EPE_07-1999



(this is connected internally tothe substrate and does nothing).Solder the remaining wires inposition with the flat face of theIC towards the left-hand edge ofthe PCB. (An alternative TO92pinout outline is included as thisversion may be offered to read-ers.)

Solder a 5cm piece ofstranded connecting wire to thepoint labeled +V on the copper track side of the PCB and 30cmof black extra flexible wire to the0V point on the topside (this willbe the flying lead).

Finally, insert IC2, takingcare over its orientation, into its

socket. This is a bipolar deviceand requires no special handlingprecautions.

BOXING UP The PCB may now be pre-

pared for mounting on the re-movable side panel of the spec-ified box. Note that this is donewith the copper track side facing

outwards see photographs.

Check the position of the

PCB and, by careful measure-ment, mark the mounting andLED holes on the panel. Drill allthese holes.

In the prototype, those forthe LEDs were made slightlysmaller in diameter than that ofthe LEDs themselves. This isbecause they did not actuallyprotrude through the holes andthis was thought to give a betterappearance to the finished unit.

Having finalized the posi-tioning of the PCB, the nexttask is to establish the positionof the fuseholder in the case.This is mounted on the boxbase and a small hole should bemarked and drilled to take theholder fixing bolt.

Now attach the PCB on theremovable panel, using small

nuts and bolts with spacers be-tween to allow the LEDs to takeup their correct positions. Thespacers are trimmed to the re-quired length by cutting them tosize with a small hacksaw. Se-cure the fuseholder in positionand solder the +V wire leading

from the PCB to one end of it.

PROBING QUESTION Prepare the probe, which is

supplied with the specified box.This needs care and patience.Cut off a 5cm piece of strandedwire and strip a small section(about 2mm) of insulation fromthe end. Solder “tin” the baredwires and insert them into thehole in the probe.

Now “feed” solder into it un-til it is almost level with the rim

of the hole it must not bulge

above it. It is very easy to make

a poor joint which parts easily

check carefully.

When satisfied, push thelarger plastic bush over theprobe. Pass the wire throughthe smaller bush and engage itwith the larger one. This leavesa recess, which will be grippedin the larger of the two holes in

the ends of the case when thisis assembled.

If the recess is too wide, thesoldered joint is probably toothick. Solder the end of the wireto the free terminal of the fuse-holder. Check the position ofthe flying lead in the other re-cess and, leaving a l ittle slack,

place a tight cable tie around itto provide strain relief.

Assemble the case usingthe plastic bushes and screwsprovided with it. Make sure theprobe and flying lead are cor-rectly located in the holes andcheck that the flying lead cannot

be pulled free by any reason-able amount of force.

Make a label to cover therectangular opening in the faceof the front panel of the box. Fita crocodile clip appropriate tothe size of battery to be testedto the end of the flying lead.

TESTING Testing is not absolutely

necessary. You could use a“proof is in the pudding” ap-proach. However, some kind ofbasic check is probably a goodidea and, if you have a digitalvoltmeter, you could confirm theoperating points. In the proto-type unit, these were all within

0

values.

05V (50mV) of the nominal

Always take care to connect the circuit with the correct polarity. In two tests on

the prototype, incorrect polarityresulted in the fuse blowing, buteverything else survived.

However, this MUST NOTbe relied on and damage couldoccur. The value of the fuse ishigher than is strictly necessarybecause the very low valueshave a significant resistance,which would result in an unac-

Completed “checker” with the removable side panel, holding the PCB, slotted into position ready for closing-up the box.

7/23/2019 EPE_07-1999



ceptable voltage drop.

If you are using the instru-ment to check the state ofcharge of a small battery, it willbe convenient to use a direct

method of testing. Begin with afully charged battery and run itdown using the piece of equip-ment with which it will be usedor a small bulb of comparablepower rating.

Aim to discharge to the lowpoint over a period of 10 to 15

hours do not try to hurry the

job by applying a large load. Forexample, if the capacity of thebattery is 7Ah, it will dischargeto the end point if 500mA is

drawn for about 14 hours.This could be obtained us-

ing a 5W or 6W car-type bulb.In practice, the capacity may beconsiderably less than the nomi-nal value depending on theamount of service the batteryhas given and how carefully ithas been used.

TAKING THE LOAD Connect the load (bulb) for

15 minutes, disconnect it andallow the battery to rest forabout 15 minutes. Connect theTester and note that all LEDscome on.

Connect the load again andcontinue discharging for set in-tervals (say, 30 minutes), allow-ing a 15-minute rest in eachcase before making the check.Keep track of the total elapseddischarge time.

You should find that thegreen, yellow, and orange LEDsgo off at reasonably spaced in-tervals. Do not wait until the red

LED goes off you could

damage the battery . If it hap-pens by accident, re-chargepromptly.

If you are using the unit totest a car battery, it would not

be advisable to run it down inthe way described above, be-cause the car is likely not tostart afterwards! Either the op-erating points will need to betaken on trust or you could test

the unit with a bench power sup-ply having a continuously vari-able output voltage. Again, adigital voltmeter would deter-mine the operating points accu-rately.

A BIT CRANKY As stated previously, the

red LED will normally remain onduring the test unless the bat-tery is seriously discharged

(below a terminal voltage of9 8V). You could regard this as

corresponding to the minimumstate of charge, which could bereached by accident and whichwould demand immediate re-charging. It also provides the“Crank Test” (see below) whichis only applicable to car batter-ies.

A lead-acid battery hassuch a low internal resistancethat any normal load placed on

it will result in only a small volt-age drop (very small indeedwith a physically large batterysuch as a car battery). Thus,virtually the full available volt-age will appear across it.

However, when the batteryis subjected to a very heavyload, even this small internalresistance will result in a sig-nificant voltage drop andthis is subtracted from thesupply to give the “terminalvoltage”. In this case, it doesnot mean that the battery isnecessarily low on charge.

In a car, the heaviestload is imposed by thestarter motor while turningthe engine (“cranking”). Thevoltage may then fall to some10V.

With a battery nearing theend of its service life, the inter-nal resistance tends to rise andthe voltage will fall still further.Eventually, it will reach thepoint where the starter motor

fails to turn quickly enough tostart the engine. It may noteven turn at all.

To perform a crank test,first ensure that the battery is ina good state of charge. It will behelpful if you can prevent thecar from starting (consult theworkshop manual to make surethis is safe/possible). The testeris then applied to the batterywhile an assistant operates thestarter. Take great care to avoidhot or moving parts!

The red LED should remainon (although it may flicker a lit-tle). If it goes off immediately orvery quickly afterwards, the bat-tery should be renewed.

Do not run the test for more than a few seconds since the battery will run down rapidly and may not start the car next time. There could also be problems with

the engine flooding making it more difficult to start.

http://0799g.pdf/

7/23/2019 EPE_07-1999



These days the word stroboscope probably conjures up im-ages of high power flashing lightsat a disco rather than a scientificinstrument. However, the oldstyle stroboscope still lives on,and units of this type can producesome fascinating results

As many readers will nodoubt be aware, the scientific pur-pose of a stroboscope is to“freeze” moving machinery. Thebasic idea is to synchronize theflashing light of the stroboscopewith the machine so that the lightflashes at precisely the samepoint in each cycle of the ma-chine.

The flash of light must bemuch brighter than the ambientlight level so that any onlookers

only see the machine during thepulses of light. Because they onlysee the machine at the samepoint in each cycle it seems to bestationary.

In fact, things become moreinteresting if the stroboscope andthe machine are slightly out ofsynchronization. With the lightflashing slightly later in each cy-cle the machine appears asthough it is operating in slow mo-tion. A lack of synchronization in

the opposite direction makes itseem as though the machine isgoing slowly in reverse!

By carefully adjusting theflash-rate it is therefore possibleto move to any point in the oper-ating cycle of the machine, and toeffectively make the machine op-erate at the desired speed in ei-

It is adequate for use withthings such as watches, smallengines, or any small mecha-nisms that have a repeating ac-tion. The use of LEDs keeps thecost to a minimum, enables anextremely simple circuit to beused, and permits safe opera-tion from a low voltage batterysupply.

DESIGN CONSIDER- ATIONS

On the face of it, this appli-cation requires nothing morethan a low frequency oscillatordriving one or more LEDs. Inpractice the oscillator must pro-vide very brief output pulses ifthe desired action is to be pro-vided. To be more precise, it isthe ratio of the on time of theLEDs to the off time that is of

importance.If the LEDs were simply to

be switched on for 50 per centof the time, the machine wouldgo through half a cycle duringthe course of each flash of light,

Freeze the action with the second of our low-cost,easy-build practical starter projects.

ther direction. This makes itpossible to closely analyze theaction and see precisely howeverything operates.

BRIGHT LEDs In order to “freeze” large

pieces of machinery it is neces-sary to use a high power strobo-scope. Such a device is notnecessarily very complicated,but it requires the use of rela-tively expensive flash tubes thatneed high operating voltages.Not a project that is usually con-sidered suitable for beginners.

However, as the title sug-gests, the LED Stroboscope featured here is based on ultra-bright light-emitting diodes(LEDs) that provide compara-tively small light levels and is alow-budget project that is idealfor the newcomer to electronics.Consequently this unit must beused in a darkened room and itwill only illuminate a small area.

7/23/2019 EPE_07-1999



providing a very blurred imageto the users. Practical experi-ments suggest that an off to onratio of at least 100 to 1 isneeded in order to obtain a rea-sonably sharp “frozen” image.

This brings a second prob-lem, which is a lack of bright-ness from the LEDs when they

are only switched on for one percent of the time. A modernultra-bright LED will providegood brightness from a currentof about 20mA, but pulsing it atthis current for one per cent ofthe time this gives an averagedrive current of only about

0 2mA. This gives nothing more

than a faint glow from even themost efficient of LEDs.

In order to overcome this,the LEDs must be pulsed at a

much higher current than nor-mal, and should ideally bepulsed at about two amps (2000milliamps). This gives an aver-age current of 20mA and goodbrightness.

Most LEDs are rated to takecontinuous currents of up to20mA or 50mA, but they canwithstand much higher currents

provided they are supplied forshort periods and the averagecurrent consumption is withinthe permitted maximum current.The required high currentpulsed operation is thereforeacceptable provided the mini-mum frequency is not made toolow, which would leave theLEDs switched on for too longduring each pulse.

CIRCUIT OPERATION The full circuit diagram for

the LED Stroboscope appearsin Fig.1. The basis of the circuitis a (more or less) conventionaloscillator circuit based on IC1.This is a form of relaxation os-cillator and it operates by firstcharging capacitor C2, and thendischarging it.

Integrated circuit IC1 is anoperational amplif ier (opamp),but it is used here as a voltagecomparator. Its output (pin 6)goes high when the invertinginput (pin 2) is at a lower volt-age than the non-inverting input(pin 3). Reversing the relativestates of the two inputs results

in the output going low.

Resistors R1 and R2 form apotential divider that biases thenon-inverting input of IC1 tohalf the supply potential, but thecoupling through resistor R3and potentiometer VR1 to theoutput of IC1 modifies this po-tential. When the output of IC1is high the bias potential ispulled higher, and when it is lowthe bias is taken lower. Theamount of change depends onthe setting of VR1, and be-comes greater as the resistance

Fig.1. Complete circuit diagram for the LED Stroboscope. The two “strobe light” LEDs can be housed remotely from the main unit (see photograph).


R1, R2 15k (2 off)R3 10kR4 1kR5 100kR6, R7 4.7 ohms (2 off)


All 0.25W 5% carbon fi lm

CapacitorsC1 220u radial electrolytic, 16VC2 470n polyester, 10mm lead spacing

SemiconductorsD1 1N4148 signal diodeD2, D3 ultra-bright red LED (see text) (2 off)TR1 TIP121 or TIP122 np n power Darlington transistor.

IC1 LF351N bi-FET opamp

Miscel laneousS1 s.p.s.t. miniature toggle switchB1 12V battery pack (8xAA size cells in holder)

Multi-project PCB available from theEPE Online store , code 7000932(www.epemag.com ); medium sizemetal or plastic case, size to choice(see text); small box for LEDs(optional); control knob; battery clip(PP3 type); 8-pin DIL socket; multi-strand connecting wire, solder pins,solder, etc.

$15Approx. Cost Guidance Only (Excluding Batts & Case)

Potent iometerVR1 100k rotary carbon, linear

http://0799q.pdf/

http://0799q.pdf/






http://0799q.pdf/


7/23/2019 EPE_07-1999



of VR1 is reduced.

Initially there is no chargeon capacitor C2 and the outputof IC1 goes high. Capacitor C2then charges from the output of

IC1 via resistor R5 and themuch lower resistance path pro-vided by R4 and diode D1. Thelow resistance of R4 results in arapid increase in the charge po-tential until it goes above thebias voltage at the non-invertinginput. The output of IC1 thengoes low, and capacitor C2starts to discharge via resistorR5.

There is no discharge paththrough resistor R4 and diode

D1 because D1 blocks any flowof current in this direction. Ca-pacitor C2 therefore dischargesat a relatively slow rate throughR5 alone.

The waveform producedacross C2 is a form of sawtoothwave (Fig.2a). More impor-tantly, the output of IC1 pro-duces brief positive pulses, asin Fig.2b. In fact, the pulses areeven briefer than those shownin Fig.2b due to the massive

difference in the values of R4and R5. This gives the requiredmark-space ratio of about 1 to100.

STROBE RATE For a stroboscope to be of

any practical value, it must bepossible to vary the output fre-quency over a reasonable span.This is the purpose of poten-

tiometer VR1, which is shownconfigured as a variable resistor.

Conventionally, the frequencyof an RC oscillator is controlledby varying the resistance in the

timing circuit, but that would bedifficult in this case as there aretwo resistors in the timing circuit(R4 and R5). Simply varying thevalue of R5 would produce sub-stantial changes in the mark-space ratio of the output signal.Altering the value of the positivefeedback resistance gives the re-quired changes in frequency with-out significantly altering the out-put waveform.

With VR1 at a high value

there is little change in the biaspotential at the non-inverting in-put when IC1’s output changesstate. The charge on capacitor C2therefore has to change by only asmall amount to move from onethreshold to the other, and thistakes relatively little time.

With VR1 set at minimum re-sistance the charge and dis-charge threshold voltages arepulled several volts apart, greatlylengthening the charge/discharge

cycle. The flash rate can be var-ied from approximately 17 to 100per second. This corresponds torotation speeds from about 1000RPM to 6000 RPM.

CURRENT DRIVER Only output currents of a

few milliamps can be providedby IC1, and a large amount ofamplification is needed to pro-

vide the LEDs (D2 and D3) withsuitably high drive currents.This is provided by TR1, whichis a Darlington power transistorused as an emitter followerbuffer stage.

A Darlington transistor is re-ally two transistors connected sothat the output current of one de-vice drives the input of the sec-ond. This effectively gives a su-per high gain transistor having acurrent gain equal to the product

of the current gains of the individ-ual transistors. The current gain ofTR1 is typically several thousand,and this enables it to provide out-put currents of a few amps.

A current of a little over oneamp is driven through eachLED, which produces an aver-age current of about 10mA perLED. This is high enough togive good brightness but lowenough to avoid operating theLEDs close to the point where

they are in serious danger ofbeing destroyed. The averagecurrent consumption of the cir-cuit as a whole is about 23mA.

Fig.2. (a) Example waveform at pin 2 of IC1 and (b) at pin 6 of IC1.

Component layout on the multi-project PCB.

7/23/2019 EPE_07-1999



CONSTRUCTION The EPE and EPE Online

multi-project printed circuitboard (PCB) forms the basis ofthis project, and it utilizes the

component layout, copper foilmaster and wiring shown inFig.3. This board is availablefrom the EPE Online Store (code 7000932) atwww.epemag.com

Although this project is ex-tremely simple, the usual warn-ings about the multi-projectPCB have to be repeated here.Unlike an ordinary customprinted circuit board, the multi-board has numerous “leftover”

holes that tend to confusethings slightly when fitting thecomponents. Take extra care toavoid misplaced componentswhen building this board anddouble-check the completed

board very carefully for errors.

In all other respects con-struction of the board is largelystraightforward. The LF351N

used for IC1 is not static-sensitive, but it is still advisableto use a holder for this device.

Fitting TR1 is slightly awk-ward because the leadouts of

the device do not match upproperly with the board layout.Things would be much easier ifthe base (b) and collector (c)

terminals of the Darlington tran-sistor were the other way round.In order to fit the device into thislayout it is necessary to crossover the base and collectorleads, but this is not difficultprovided TR1 is given the orien-tation shown in Fig.3.

The pinout wires of TR1 canbe fitted with short pieces ofPVC sleeving to ensure thatthere are no accidental short-circuits. Fig.4 shows the leadout

configuration for TR1, andshould help you to avoid errorswhen connecting this compo-nent.

Although the TIP122 is apower transistor it only operatesat very low average power lev-els in this circuit, and so noheatsink is required. A TIP122

is used for TR1 onthe prototype, butsome suppliersstock the TIP121

instead, and this isequally suitable.

LEDs Virtually any

medium size plasticor metal box should accommo-date this project, but bear inmind that the battery pack con-

Fig.3. PCB layout copper foil master and interwiring details.Check component positions as not all holes are used.

Fig.4. Pinout details for the TIP121/122 Darlington power

transistor.




7/23/2019 EPE_07-1999



sisting of eight AA size cells isfairly bulky. Remember to allowfor this factor when selectingthe case. Mount the printed cir-cuit board on the base panel ofthe case using either plasticstand-offs or 6BA bolts plusspacers.

The two “strobe” LEDs canbe mounted on the front panelof the main case, but the unit iseasier to use if they are fitted ina separate much smaller case.The LEDs are then connected tothe main unit via a piece of

three-way cable about 0 5 to two

meters long.

The current flow to the

LEDs is not sufficient to warrantany form of heavy-duty cable.Twin-screened cable or a three-way lead peeled from a piece ofribbon cable will suff ice. Therear panels of both cases mustbe drilled with holes to take thecable, and if metal cases areused the holes should be fittedwith PVC grommets to protectthe cable.

Practically any ultra-bright LEDs should work well in this

circuit, and the higher their effi-ciency the brighter the pulses oflight produced. However, even-ness of illumination is also im-portant in this application, and itis worth experimenting with afew LEDs to find the ones thatgive the best results.

In general, larger LEDsseem to give more even beamsof light, but the latest 5mm di-ameter types give the highestlight levels. The 3mm types do

not seem to give high enoughlight output levels to be of usein this application.

TESTING Once the small amount of

hard wiring has been added theunit is ready for testing. It is es-sential to thoroughly check the

finished unit before connectingthe battery and switching on be-cause mistakes could easily re-sult in the LEDs being fed con-tinuously with a high current.This would destroy them in a

fraction of a second.

When initially testing thecircuit it is not a bad idea toconnect a 220 ohm resistor inthe cathode (k) connection tothe LEDs. This will limit the cur-rent to a safe level in the eventof a fault.

With control VR1 set atmaximum resistance (fullycounter-clockwise) the LEDsshould flash at a low enough

rate for the pulsing action to beseen. Advancing VR1 in aclockwise direction will soon in-crease the operating frequencyto the point where the flashing istoo fast to be perceived.

A quick way of checkingthat the LEDs are still strobing isto simply wave them around inthe air. This should draw a sortof dotted line of light in the air.

IN USE Although this unit is only

suitable for use with smallitems of machinery it is stillessential to use it with duecare so that accidents areavoided. Most ultra-bright LEDsproduce fairly tight beams oflight, so it should not be neces-sary to use the unit at veryclose ranges.

Finding the right setting forfrequency control VR1 is largely

a matter of trial and error. If therequired “freezing” is obtained,but with the machine in two ormore positions simultaneously,the flash rate is too high result-ing in more than one flash percycle of the machine. The“freezing” effect will be obtainedif the stroboscope is set to flashon every other cycle, every third

cycle, etc., but the “frozen” im-age will be rather blurred. Opti-mum results are produced atthe highest flash rate that givesa single “frozen” image.

The maximum rate of100Hz (6000 RPM) might not besufficient for some small piecesof machinery. The maximumoutput frequency can be in-creased to about 300Hz (18,000RPM) by increasing the value ofVR1 to 220 kilohms, but accu-rate adjustment of the flash ratewill then be more tricky. Using avalue higher than 220k is notrecommended as it could resultin the oscillator stalling.

NEXT MONTHS

STARTER PROJECT - 3

FREEZER ALARMDon’t miss out!

http://0799e.pdf/

http://0799e.pdf/

http://0799e.pdf/

http://0799e.pdf/

http://0799h.pdf/

http://0799e.pdf/

7/23/2019 EPE_07-1999



This Heads-Tails Indicator(see Fig.1) is superior to otherdesigns in several ways. It israndom to a fault, needs no

setting up, and operates with asingle pushswitch to display eitherred or green on a single LED.

In the circuit diagram ofFig.1, the oscillator based aroundIC1a and IC1b, feeds a rapid trainof pulses via ICIc to IC2a, a D-type flip-flop. As each pulse isreceived, IC2a changes statefrom “heads” to “tails” and viceversa. Within 70 milliseconds,IC2a receives about 500 pulses,then timer IC1c stops the clock.Within a further 20ms, IC1dswitches on transistor TR1, whichcauses the tricolour LED D1 todisplay either red or green.

The purpose of IC1d/TR1 isto prevent the differing loadsrelated to D1 from biasing thetimer IC1c in favor of eitherheads or tails. Various factors

combine to ensure that theSuperior Heads-Tails Indicatoris virtually completely random.

Rev. Thomas ScarboroughFresnaye, Cape Town, R.S.A.

Home Alarm System

The circuit diagram in Fig. 2is a simple home alarm systempanel designed to provide some

of the more useful features ofcommercial panels at a fractionof the cost. The loops markedas “Zone 1”, “Zone 2” and“Entry/Exit” (S4, S5, S7) arenormally-closed alarm sensors.The “Tamper Loop” (S6) is alsoa normally-closed tamperprotection circuit.

If Zone 1, for example, wasopened then pin 2 of the OR

ROLL-UP, ROLL-UP!

Ingenuity is our regular round-up of readers' owncircuits. We pay between $16 and $80 for all materialpublished, depending on length and technical merit.We're looking for novel applications and circuit tips, notsimply mechanical or electrical ideas. Ideas must be thereader's own work and must not have been submittedfor publication elsewhere. The circuits shown haveNOT been proven by us. Ingenuity Unlimited is open toALL abilities, but items for consideration in this columnshould preferably be typed or word-processed, with abrief circuit description (between 100 and 500 wordsmaximum) and full circuit diagram showing all relevantcomponent values. Please draw all circuit schematics

as clearly as possible.Send your circuit ideas to: Alan Winstanley,

Ingenuity Unlimited , Wimborne Publishing Ltd., AllenHouse, East Borough, Wimborne, Dorset BH21 1PF.They could earn you some real cash and a prize!

Win a Pico PC-Based Oscilloscope

•

50MSPS Dual Channel Storage Oscilloscope

• 25MHz Spectrum Analyzer

• Multimeter

• Frequency Meter

• Signal Generator

If you have a novel circuit idea whichwould be of use to other readers, then a PicoTechnology PC based oscilloscope could beyours.

Every six months, Pico Technology will be

awarding an ADC200-50 digital storage oscil-loscope for the best IU submission. In addi-tion, two single channel ADC-40s will be pre-sented to the runners up.

Superior Heads-Tails Indicator

µ

Ω

Fig.1. Circuit for a superior Heads-Tails indicator

7/23/2019 EPE_07-1999



gate ICla would be brought highvia resistor Rl, thus causingLED D5 to light, and themonostable IC5 would also betriggered via transistor TR5.

The relay contacts RLA1 willclose for a period set by VR2.

A triggering circuit,comprising of TR5 andcapacitor C8, is requiredbecause if the sensor (S4)connected to the OR gate wereto remain permanently open(e.g. a broken window strip),then the alarm would soundindefinitely. Resistor R16 allowsC8 to discharge a few secondsafter the zone (e.g. a PIR) has

been closed again thus resettingthe alarm.

The three key switches S1to S3 may take many forms, butto disarm the alarm they mustbe closed and to arm thesystem and trigger the exit timerthey must be open. On theprototype, the “Lock” was a 36-way Centronics socket and the“Key” simply a matching plugwith six pins jumpered.

Exit time is provided bymeans of IC4, another NE555monostable timer. With “keyswitch” S1 closed (systemdisarmed) pin 2 of IC4 isgrounded (0V) thus output pin 3of the timer is held high. Thisoutput is inverted by IC3c andfed to one input of IC2a (a 4081AND gate), the output of whichconnects to the set input of the

bistable latch formed by IC3a andlC3b. Assuming the Exit timer IC4has timed out then pin 2 of IC2awill be high but pin l will be low(Exit/Entry zone closed).

Upon opening of the entry/exitzone the bistable will be set, thusbringing its Q output (pin 4 of IC3b)high. This begins to chargecapacitor C3 to give an entry timeof about 17 seconds before pin 5 ofICla is brought high. If the key isinserted during this interval thenthe Alarm Bell Timer (IC5) isdisabled, the exit t imer re-established and the bistable isreset. If not then the alarm soundsas before. Capacitor C4 provides

power on reset for the bistable andthe LEDs. Dl to D4 provide visualindication of zone status.

A power supply circuit capableof charging a lead acid battery isshown separately in Fig.3b. Thealarm circuit could easily bemodified to accept more zones bythe use of the second OR gate inthe 4072 package.

Damien MaguireCo. Wicklow, Ireland

Fig.2. Suggested (UK) power supply/charger circuit for the Home Alarm System

7/23/2019 EPE_07-1999



Fig.2. Control circuit diagram for the Home Alarm System

http://0799i.pdf/

7/23/2019 EPE_07-1999



One of the more unusualprojects published in the printedissue of EPE last year was theEPE Mood Changer featured inthe June ’98 issue, which wasan adjustable low-frequencymagnetic field generator. Suchfields are thought to encourageelectrical activity of similar fre-quency in the brain, which inturn may promote associatedmoods and sensations such asdeep relaxation, creativity oreven sleep.

STATE OF MIND It has been known for many

years that the human brain ex-hibits electrical activity at vari-

is thought to accompany mentalimagery and creativity.

The lowest band, 1Hz to3Hz, is known as Delta and isnormally only found in deepsleep. Whilst this might notsound very useful, many insom-niacs will disagree, arguing thatanything that helps them to get

into the Delta state is very use-ful indeed.

ENLIGHTENMENT Various techniques for us-

ing these brainwaves have beendeveloped over the years. Theearliest method was to detectthem with electrodes placed onthe head, so that users wouldknow when they were presentand could therefore learn to

generate them at will and in-crease their intensity.

Known as EEG Biofeed-

In an age of stress and insomnia, let the “force” be with you.

ous frequencies and that tosome extent these are related tocurrent moods. For instance,normal waking consciousness,such as that emitted by readersof this article (we hope!) is gen-erally accompanied by activityat about 20Hz, known as theBeta frequency.

Lower frequencies havecorresponding states, such asthe band between 8Hz and13Hz which is generally sup-posed to accompany a deeplyrelaxed but aware state. Thisband is called Alpha and widerinterest in it began some yearsago when researchers discov-ered high levels in Zen masters

during meditation. More recentlythere has been interest in theTheta band, 4Hz to 6Hz, which

WARNING NOTICE

It is known that photic stim-ulation at Alpha frequencies cancause seizures in persons suf-fering from Epilepsy. We wouldtherefore suggest that it is notwise for such people to try thisproject.

A user who is not a knownepileptic, but when using theEPE Mind PICker begins to ex-perience an odd smell, sound,or other unexplained effects,should TURN IT OFF IMMEDI-ATELY and seek professionalmedical advice.

YOU MUST TREAT THIS UNIT

WITH DUE RESPECT

7/23/2019 EPE_07-1999



back, this was a difficult andsometimes messy technique,requiring high levels of amplifi-cation, complex filtering andusually some sort of gel toachieve good electrode to scalpcontact through hair. As a resultthe more pro-active“entrainment” became morepopular, an example being theEPE Mind PICkler featured inthe December ‘98 issue.

Instead of passive detec-tion, these use flashing lightsand special sounds to encour-age the production of the de-sired frequencies in the user’sbrain. It is interesting to notethat some of the early commer-cial devices of this type in-cluded a magnetic field as partof the stimuli provided.

More recently a small andrelatively expensive device wasmarketed which was claimed toreduce the stress of modern liv-ing by providing a weak mag-netic field close to the user atfrequencies corresponding tothose of brain activity. Giventhe connection between mag-

netism and electricity it seemsquite possible that this couldwork by simple electromagneticinduction.

This will undoubtedly pro-duce minute currents in conduc-tive brain tissue, which may wellencourage it to settle into suchfrequencies of it’s own. Themere fact that devices using thetechnique have been made andsold on a commercial basis sug-gests that it is worth investiga-

tion, especially at home con-struction costs.

DOES IT WORK The big question is “Does it

work?” All results from some-thing like this are highly subjec-tive and may vary from user touser. However, the author finds

that using one of these devicesfor a while at Alpha frequencydoes seem to produce a senseof relaxation, and popping it un-der the pillow with a Delta set-ting often results in an excellent

night’s sleep.

Correspondence from con-structors following publication ofthe original EPE Mood Changer suggest that it works for themtoo, especially for insomniacs.Whilst there can be no guaran-tee that it will be effective forany particular reader, this newdesign does make it much eas-ier to test.

The use of a PIC microcon-

troller for frequency generationhas simplified construction andeliminated the need for calibra-tion and adjustments. At aroundhalf the size and weight of itspredecessor, it fi ts easily into asmall pocket and the lighterweight makes it more robust,the prototype has even beendropped without sustaining -damage.

It is cheaper to build andrunning costs are lower since it

should operate for over twohundred hours from just twoAAA cells. Readers wishing todiscover if it will work for them,for insomnia, as a meditationaid, or even simply as a stressreliever, now have no excusenot to build one.

HOW IT WORKS The working principle of the

device is straightforward. It pro-

duces a low frequency sinewavesignal, which is applied to anair-cored inductor to create analternating magnetic f ield. Thesignal is generated digitallysince this eliminates the ampli-tude control problems often as-sociated with low frequencygeneration and it is easier to -produce precise frequencies us-

ing digital techniques.

In this design a PIC16F84microcontroller is programmed toproduce the sequence of outputsshown in Fig.1 from the eight out-puts of port B. These are com-bined by a network of resistorswith values chosen to produce asingle output which is approxi-mately sinusoidal in shape, albeitslightly “jagged” as shown inFig.2.

It can be seen that the out-puts from port B remain in their

highest (all “on”) and lowest (all”off”) states for two cycles, to sim-ulate the flatter parts of the sinecurve at these points, so a com-plete sinewave cycle is simulatedwith a total of eighteen steps. Thesignal then goes to an outputstage which drives the coil. The“bridge” configuration of this ef-fectively doubles the coil drive

Fig.1. Output pattern from PIC port B.

7/23/2019 EPE_07-1999



voltage to increase output.

To minimize battery con-sumption it was decided to op-erate the PIC in LP (low power)

mode using a 32

768kHz watch

crystal for the oscillator. Thesecrystals are readily available atlow cost from many componentsuppliers.

Readers with knowledge of

PIC programming will know thatthe PIC divides its oscillator byfour for the internal clock, so theprogram chunters along at justover 8kHz, which raises inter-esting problems for the pro-grammer. At this low frequencyevery program step must becounted as the final output fre-quency depends entirely on the

total number of steps in the pro-gram loop generating it.“Structured” programming withtidy modular subroutines wasquickly abandoned and instead astraightforward “top-down” ap-

proach is used as far as possible.

FLOW CHARTS The main flow diagram of the

program is shown in Fig.3. It be-gins by configuring the inputs andoutputs and setting output RA0high for two seconds to turn on a“battery check” LED.

Two interesting points arosefrom this. The original intentionwas to drive the LED with RA4,which is an “open drain” outputand can therefore only sink cur-rent. However, it was found thatthe LED continued to glow slightlywhen this output was set “high” toturn it off so it would appear im-practical to use RA4 in this way.Subsequently, RA0 was used forLEDdriving and RA4 became one

of the four inputs for f requencysetting.

The other point to note isthat the original data for the

Fig.2. Example of digitally

generated “sinewave”.

Fig.5. Complete circuit diagram for the EPE Mood PICker.

Fig.3. Flow diagram for the EPE Mood PICker main

program.

7/23/2019 EPE_07-1999



PIC16C84 states that togglingRA0 when configured for LP op-eration may result in a spuriousinternal clock cycle. Althoughthis did not appear to happenduring design with the “F” ver-sion, to be on the safe side acouple of “NOP”s (no operation)have been inserted into the pro-gram after each change of thisbit.

Following the battery testthe program stores the states ofthe four frequency settingswitches (see S3 to S6 in Fig.5)in a file. These give a total ofsixteen possible combinations,which are used to select the de-sired output frequency. It thencontinues with an eighteen-steprepeating loop, which generatesthe output sequence from theeight pins of port B.

The flow of one of theseeighteen steps is shown inFig.4. It begins by setting orclearing the port bit currentlydue for updating. The switchstates are then compared withthe value currently held in thefile and, if they are unchanged,a table is called to select a de-

lay associated with the filevalue.

The delay is one of sixteenwhich set the overall output fre-quency. If the switch and file

values are at odds the programis sent right back to the start,where it performs the batterycheck again and stores the newswitch values before returningto the frequency generation rou-tine.

Halfway through the eigh-teen steps, when all the outputsare high, the delay is calledtwice, and at the final step whenthey are all low it is again calledtwice before the process is re-

peated. A few “NOP”s are in-serted as required to ensure theduration of these two steps isexactly twice that of the othersin the sequence.

CIRCUIT DESCRIPTION The full circuit diagram for

the EPE Mood PICker project isshown in Fig.5. Starting at theleft, the top two switches S1 andS2 both perform the on-off func-

tion. DIL switches are not nor-mally available in units of five,so a six-way version is em-ployed, and rather than leaveone unused it is connected inparallel with the power switch.

The next four switches, S3

to S6, are connected to thePIC’s (IC1) port A bits 1 to 4,with pull-down resistors R1 toR4 to give logic-level inputs.The battery indicator LED D1 isconnected to RA0 through cur-

rent limiting resistor R5.

Most red LEDs have a for-ward voltage of about two volts,so when the supply voltage ap-proaches this it will becomemuch dimmer, warning the userthat the batteries are reachingthe end of their useful life. Asmentioned earlier the LED lightsfor about two seconds eachtime the unit is switched on orwhen the frequency settingswitches are altered.

The MCLR function of IC1is not used and is rendered in-active by connection to the posi-tive supply through resistor R6.

A crystal, X1, is connectedacross the oscillator terminalswith 68pF capacitors C3 and C4to ground. This capacitor valueappears high for this type ofcrystal, which is normally usedwith around 12pF. The PIC datasheets suggest 68pF however,

and the PIC oscillator will notwork with 12pF.

Incidentally, if a crystal ofthis type is used with a 5V sup-ply, a resistor should be placedbetween OSC2 (pin 15) and thecrystal to prevent overdriving. A

Fig.4. Flow diagram for a sin-

gle step of the frequency generation loop.

7/23/2019 EPE_07-1999



value of 56 kilohms appearsgenerally suitable.

SINEWAVE - GENERATION

Resistors R7 to R20 com-bine the outputs and shapethem into a sinewave. Six of therequired values consist of two

resistors in series giving calcu-lated errors of less than 0 1 per

cent at every step, and the re-sulting waveform looks verygood indeed on an oscilloscope.If you have a ‘scope, it's wellworth checking this.

The total output impedanceof the resistor network is a frac-tion under 22 kilohms (22k). So,loading it with a 100k resistor,R21, reduces the output peak-to-peak voltage to a little less

than the supply of 3V at about2 5V p-p.

The output stage is builtaround IC2, an AD8532 dualopamp. This device appearedon the market fairly recently andhas some very useful features.Its inputs and outputs are bothrail-to-rail and the outputs can

source or sink up to 250mA. De-spite this high output current

capability it draws only 1 6mA of

quiescent current and is guaran-teed to operate from supplies

down to 2 7V.

In practice, this circuit oper-ates down to 2V without prob-lems. It is fast, up to 4MHz, anddistortion is low enough for au-

dio applications.

The only drawback appearsto be an upper supply voltagelimit of 7V, which makes directoperation in circuits with 9Vsupplies impossible. For lowervoltages though, it opens upnew horizons.

In this circuit, IC2a is con-nected as a non-inverting ampli-fier and IC2b as an inverter. Be-tween them they provide a

“bridge” output to the coil L1,which therefore has a peak-to-

peak drive of about 5V, or 1 77V

RMS. Three decoupling capaci-tors are used, C1, C2, and C5which is a 100nF ceramic typeplaced close to IC2 to preventany tendency to instability.

CONSTRUCTION Construction of this project

is fairly straightforward since allthe components are fitted on asingle printed circuit board(PCB). The topside componentlayout together with the under-side copper foil master trackpattern is shown in Fig.6. Thisboard is available from the EPE

Online Store (code 7000233) atwww.epemag.com.

If the specified case isused, the first task is to checkthat the PCB fits correctly onthe mounting pillars provided.Following this a square holeshould be cut in the case for theDIL switches, see photographs.

In the prototype, this wasdone by holding the PCBagainst the inside front of the

case with the same gap fromthe top as it had when securedto the pillars to act as a tem-plate whilst drilling twelve 1mmholes through the case in theDIL switch pin positions. Theswitch was then popped intothese holes on the other side toposition it accurately whilst

Fig.6. PCB component layout and (approximately) full-sized copper foil master pattern for the

EPE Mood PICker. Note that the DIL switch should be mounted on the board using a wire-wrap socket (see text).







7/23/2019 EPE_07-1999



marking its outline with a knife.

Careful work with a drill, rat-tail files, and a sharp knife thenresulted in an accurately placedsquare hole into which the

switch fitted neatly. The PCBwas also used to mark the posi-tion of the hole for LED D1.

The remaining componentsshould be fitted and soldered inthe usual way with the excep-tion of LED D1 and the switch.The layout of all the compo-nents is shown in Fig.6. CrystalX1 lies flat as shown and shouldbe secured with a drop of glueas it is delicate and has thinleads. A socket is recom-

mended for IC1.

COIL DETAILS The coil L1 should be air-

cored as mentioned earlier.There are several possibili-ties here, but the PCB hasbeen designed specificallyto use the coil taken froman “Eagle” LT44 audiodriver transformer, whichshould be dismantled asfollows.

The outer metal casingcan be simply pulled apartat the bottom and prisedoff. Beneath this will befound an adhesive papertape around the core, whichis easily removed. The lam-inations appear to be heldby wax impregnation andcan be carefully removedusing a sharp knife to breakthe adhesion between them

and small pliers or sidecut-ters to pull them out.

In usual transformerfashion some of them areE-shaped and some are juststrips. The more lamina-tions are removed, the eas-ier it becomes to prise outthe remaining ones.

This operation will leave just the bobbin with its primaryand secondary windings. Onlythe primary is used in this pro-

ject, although the secondaryconnections are soldered to the

board to secure it.

An alternative to the LT44bobbin is one taken from a “10Henry choke”, which can beseen in the photographs. Thethinking here was as follows.Both the choke and the LT44primary have about the sameDC resistance and are physi-cally about the same size. Thetransformer has two windingsbut the choke has only one so itprobably consists of more turnsof thicker wire. More turns,same current, more magneticfield for the same power!


R1 to R4, R13, R14, R21 to R24

100k (10 off)R5 22 ohms

R6 4k7R7, R9, R17, R19 220k (4 of f)R8, R12, R16, R20 68k (4 of f)R10, R18 33k (2 of f)R11, R15 120k (2 of f)

See also the

SHOP TALK Page!

All 0.6W 1% carbon f i lm

CapacitorsC1, C5 100n resin-dipped ceramic (2 off)C2 100u radial electrolytic, 10V

C3, C4 68p resin-dipped ceramic

(2 off)

SemiconductorsD1 3mm red LED, 10mA type

IC1 PIC16F84 pre-programmed microcontrollerIC2 AD8532 dual opamp

MiscellaneousX1 32.768kHz watch type crystal

L1 air-cored coil (LT44, see text)S1 to S6 s.p.s.t. 6-way sub- miniature DIL switch

B1 1.5V AAA cell (2 off) with 2-cell AAA holder.

Printed circuit board available

from the EPE Onl ine store , code7000233 (www.epemag.com );handheld plastic box, with batterycompartment, s ize 105mm x 61mmx 28mm; 14-pin wire-wrap socket(see text); 18-pin DIL socket; batteryconnector with leads,solder, etc.

$30Approx. Cost Guidance Only (Excluding Batteries)

Completed unit showing the battery compartment and positioning of the

circuit board.

http://0799q.pdf/

http://0799q.pdf/







http://0799q.pdf/


7/23/2019 EPE_07-1999



In practice the chokeproved far more difficult to dis-mantle as the laminations ap-peared to be stuck together withsome kind of glue instead of thewaxy substance of the LT44.Prising them out was a realtouch-and-go operation. Thebobbin has only two connec-tions so it was glued to theboard though the connectionsdo fit directly into the sameholes.

Following all this effort therewas only a tiny increase in thefield output. Considering the dif-ficulty of dismantling it and thefact that it costs nearly threetimes as much, constructors arerecommended to stick with theLT44 audio transformer as thesource of the coil.

SWITCHED-ON Some options are available

when fitting the DIL switch. Forsimplicity it could be soldereddirectly to the PCB and oper-ated through a hole in the case,but this might prove awkward.

A socket could be used toraise it a bit, but the same diffi-culty would apply. It could beglued to the case and thenwired to the board.

However, for the prototypea socket with wire-wrap pinswas used to raise it so that itssurface was flush with that ofthe case. Wire-wrap sockets do

not seem to be as common asthey once were, and 12-pin onesmay well not exist. A 14-pin onewas modified for the prototype bypulling out the unwanted pins andthen cutting off the unused bit of

the body with a sharp knife. Thiswould also work for a 16-pinsocket.

However, even 14- and 16-pin wire-wrap sockets are becom-ing rare. A couple of alternativesthat are still readily available arestrips of wire-wrap socket that canbe cut to length, and 28-pin sock-ets that can be cut to provide twosuch strips.

To fit the socket the DILswitch was first f itted into it, then

the socket was inserted into theboard with a large dollop of Blu-Tack beneath it. The PCB wasthen screwed to the mounting pil-lars and the top half of the casewas held in place whilst theswitch was pressed down againstthe Blu-Tack into the exact re-quired position. The board wasthen carefully removed and thesocket pins soldered, then theBlu-Tack was removed.

This admittedly fiddly opera-tion results in a switch that ismounted directly onto the PCBbut is perfectly f lush with the casefront when the unit is assembled,providing easy operation and areally neat appearance. A furtheradvantage is that if the switchshould fail a replacement can besimply plugged in. A similar pro-cedure was used to fit LED D1,using a large blob of Blu-Tack tohold it in position for soldering.

The batteries are fitted in a 2-cell AAA holder, fixed into thecase with double-sided adhesivestrip. They have a slight tendencyto pop out of the clip if the unit isknocked, but this can be pre-vented by filling the spare spacein the battery compartment with apiece of firm polyurethane foam.

TESTING Very little testing is required

with this project. When switchedon the LED should light for twoseconds, then the output should

start running. Whilst D1 is lit thecircuit draws about 10mA but innormal operation the circuitdraws less than 4mA.

Changing any of the fourfrequency setting switchesshould result in another opera-tion of the LED. With all fourswitches in the “off” position theunit will operate at its lowest fre-

quency of just 1 5Hz so it can be

checked with an analog meter,as the pointer should be able to

follow the sinewave outputsfrom IC2a and IC2b to the coil.

All eight port B outputs ofIC1 should be alternating highand low at the same frequency.If an oscilloscope is available ahigher frequency can be se-lected so that the output wave-form can be observed. With allfour switches “on” the frequencyshould be just under 22Hz.

STAY CALM To use the unit, an appropri-ate frequency should be se-lected. Those available areshown in Fig.7 together withtheir switch settings and associ-ated effects. The binary equiva-lent of the number for each fre-quency is set with the switches

will know how to do this!

most readers of EPE Online

All four switches “off” iszero and all four “on” is fif teen,

a total of 16 possible combina-tions. The bottom switch S6 isthe least significant bit or “LSB”.

The unit should be keptclose to the body for maximumeffect. For insomnia one of thetwo lowest frequencies shouldbe selected and it can be placedunder the user’s pillow, which

Fig.7. Switch settings, frequencies, and effects.

7/23/2019 EPE_07-1999



seems to be particularly effec-tive.

Switching the unit off andthen on again immediatelyshould be avoided as this mayresult in the LED failing to oper-

ate, probably because capacitorC2 has insufficient time to dis-charge so the circuit never actu-ally powers down. This is not aserious problem as tests show itto be generating the magneticfield correctly even when thisoccurs.

SPACED OUT One of the available fre-

quencies is 7 85Hz, which is as

near as could be programmedto the 7 83Hz “Schumann Reso-

nance”. This intriguing phenom-ena is one of the naturally oc-curring magnetic fields thathave always surrounded us.

It appears that the space be-tween the earth’s surface and theionosphere forms a gigantic reso-nant cavity with physical dimen-sions that give it a frequencysomewhere between 7Hz and

8Hz. Excitation by phenomenasuch as lightning starts oscillationand very low attenuation at thesefrequencies allows it to keep go-ing more or less continually.

Enthusiasts of the effects offields at this frequency say thatmodern man is missing out on itssupposed beneficial effects be-cause it tends to be masked bymore powerful f ields from theelectrical equipment and wiringwhich nowadays surrounds us all.

There is even a story that NASAinstalled Schumann frequencyfield generators in spacecraft af-ter finding that space sicknesswas in part due to the astronautstravelling beyond the range ofthis field.

If any reader knows thetruth of this story we would bevery interested hear from them.Meanwhile, in an age of stressand insomnia, it is hoped thatthis project will bring some ben-

eficial calm and relaxation to itsconstructors.

COMING SOON

Micro-Tesla!A sensitive magnetic field sensorcapable of indicating the tinyfields generated by the EPEMood PICker. It is also capable of

detecting an ordinary permanentmagnet at a range in excess ofthree meters, which could lead toa variety of interesting experi-ments and applications for theenthusiast.

http://0799j.pdf/

7/23/2019 EPE_07-1999



At night, an intruder alarmbell box may not be noticed and awould-be intruder might attempt abreak in if the house appears un-occupied. However, a light burn-ing in an upstairs room is a gooddeterrent and it is quite a simplematter to build a light sensitiveswitch to switch the lights on au-

tomatically when it gets dark.The deterrent effect can be

made even greater by automati-cally switching the lights on andoff every so often, so that it ap-pears that there is somebodythere to do the switching.

Obviously too rapid an on-offswitching cycle would be unrealis-tic and a period of a few minuteswould seem ideal. Since the lightbeing on would be more of a de-terrent than if the potential in-

truder happened to come alongwhile the light was off, the circuitshould keep the lights on forlonger than the period for whichthey are switched off.

DISCO LIGHTS Building a light sensitive

switch is easy enough, but settingone up to switch on the lights inthe same room is not quite sosimple. Great care must be taken

to place the sensor in a positionwhere it will sense the externallight level but not respond to thelight being switched on in theroom.

Failure to ensure this will re-sult in a feedback effect with thelight switching on when it getsdark, resulting in it being lightthus causing the light to switch off

the lamp being controlled. Toallow this, it samples the lightlevel only when the lamp beingswitched is off. When the lampis on, the brightness level is ig-nored so that feedback cannotoccur.

As the light level falls, thecircuit begins switching thelights on and off at predeter-mined intervals and, when in

the off state and the light levelhas increased above a set level,this action stops and the lightsremain switched off.

SCHMITT NAND The circuit is based on a

quad 2-input Schmitt triggerNAND gate. To understand howthe circuit works, it is important,therefore, to know what aSchmitt NAND gate does.

The truth table for a 2-inputNAND gate is shown in Fig.1.When both inputs are high, theoutput will be low, but if eitheror both of the inputs go low, the

Is there anyone at home?

again, and so on. The effect islikely to be far from realistic,although the would-be intrudermay be led to think that there isa disco party going on inside!

However, the device de-scribed here has been designedin such a way that it can beplugged into any convenient

wall socket and the lampplugged or connected to it with-out regard as to which way thesensor is facing or the power of

Fig.1. NAND gate and truth table.

CC

CC

CC

CC

Fig.2. Response of (a) “ordinary” and (b) Schmitt trigger gates to slowly rising and falling input levels.

7/23/2019 EPE_07-1999



output will go high.

A Schmitt trigger NANDgate differs from an “ordinary”NAND gate in the way it re-sponds to input signals, as illus-

trated in Fig.2. Assuming thatone input of the NAND gate isconnected to the positive rail(logic high) and a slowly risingramp voltage is applied to theother, the ordinary NAND gateoutput will be high while the in-put is below the lower logicthreshold.

For most CMOS gates, thelower logic threshold belowwhich a logic 0 value is guaran-teed by the manufacturer lies at

about one-third of the supplyvoltage. Similarly, the guaran-teed logic 1 condition lies aboveabout two-thirds of the supplyvoltage. The region between thetwo levels is considered indeter-minate, in that neither logiclevel can be guaranteed.

When the input voltagelevel enters the indeterminateregion, the circuit begins to be-have like a very high gain am-plifier so that any noise super-

imposed on the input(particularly when it is near the

middle of the indeterminate re-gion) will cause the output toswitch between states givingrise to the output waveformshown in Fig.2a.

Once the input level isabove this indeterminate region,it will now be considered to beat a logic 1 level, and the outputwill settle into its low state asdetermined by the NAND gate’sinversion function.

A similar action will happenwith a falling input voltage. Theindeterminate region does notnormally cause problems inwell-designed digital circuits be-cause the input voltages switch

so fast that the output has notime to oscillate. It can be agreat problem, though, if slow orpoorly defined logic signals areapplied.

The waveform in Fig.2bshows how a Schmitt triggergate responds to the sameslowly rising or falling input andit can be seen that no oscillationoccurs. The circuit achieves thisby having two well-defined inputthresholds and no indeterminate

region.As the input voltage rises,

nothing happens until the upperthreshold is reached whereuponthe output switches to a low level.Any small amount of noise on theinput is ignored and the circuit willremain in this state until the input

has fallen below the lower thresh-old, when the output will switchhigh.

SCHMITT OSCILLA- TOR

Not only is this characteristicextremely useful in this intruderdeterrent application, where thevoltage from the sensor will in-evitably change slowly as dark-ness falls or day breaks, but italso allows the construction ofsimple oscillators, as shown inFig.3.

When the circuit in Fig.3 isswitched on, the capacitor C willbe discharged so the input volt-age will be low, resulting in theoutput being high. Current willnow flow via resistor R and thecapacitor voltage will rise expo-nentially until the upper thresholdis reached.

At this point, the output willswitch low, causing the capacitorto discharge exponentially via re-sistor R until the lower thresholdis reached, when the output willgo high again, and so on.

The output will be approxi-mately a square wave (except forthe initial cycle) because the RCtime constant will be the samewhether the capacitor is chargingor discharging. This can be easilymodified, however, by connecting

a diode in series with another re-sistor across resistor R as showndotted in Fig.3.

Depending on the direction ofthe diode, this will effectively re-duce either the charging or dis-charging resistance, causing acorresponding change in themark space ratio of the outputwaveform. Alternatively, a resis-

CC

CC

CC

Fig.3. Schmitt trigger oscillator and its waveforms.

7/23/2019 EPE_07-1999



tor in parallel with a diodeplaced in series with resistor Rwould achieve the same result.

The circuit of Fig.3 is shownwith one input tied to the posi-tive rail, but the gate still retains

its NAND characteristics. Byconnecting the spare input to0V, the output will switch highand oscillation will stop. Thisinput can, therefore, be used toswitch the oscillator on or off ifrequired.

CIRCUIT DETAILS The operation of the In-

truder Deterrent circuit shown inFig.4 can now be described. As-

suming for the moment that thepoint marked “A” in the circuit ishigh, it will be seen that the out-put of IC1a will oscillate at arate determined by C1, R2, R3and D1.

Because of the orientationof diode D1, the charging period(during which the output of this

gate is high) will be shorter thanthe discharging period so that theoutput will be low for longer thanit is high. As this oscillator con-trols the time for which the light isswitched on and off, the compo-

nents have been chosen to give a“lights on” period of around 540seconds and a “lights off” periodof about 100 seconds.

The output of this oscillator isfed to IC1b, which simply invertsthe oscillator output. The outputof this is in turn connected to an-other oscillator, based aroundIC1c, C2, R4, R5 and D2.

In this oscillator, the values ofthe capacitor and resistors aremuch lower than with IC1a so thatthe circuit oscillates at a muchhigher frequency, of about

2 5kHz. This gate oscillates when

the output of IC1b is high and D2,R5 ensure that the transistor TR1is switched on for only a short pe-riod (around 30ms) while remain-ing off for most of the cycle(about 400ms).

This is done to reduce theaverage current drawn by thecircuit and enable a simple ca-pacitive dropper to be used inthe mains-derived power sup-ply. Transistor TR1 is used to

trigger the triac CSR1, which inturn switches the light on.

From this it should be clearthat the lights are only switchedon when the control input ofIC1c is high, because, when thisinput is low, the gate’s outputwill be forced high causing TR1(a pnp device) and the triac toswitch off.

With the control input lowand the lights off, the potentialdivider formed by R1, VR1 andphototransistor TR2 is effec-tively connected across the sup-ply rails and comes into play. Inhigh light levels, TR2 will con-duct and if the light level sensedby TR2 is such that the voltageat point A falls below the lowerthreshold of IC1a, this oscillatorwill be disabled and its outputwill stay high, causing IC1b out-

Ω

µ

µ

Fig.4. Complete circuit diagram for the Intruder Deterrent.

7/23/2019 EPE_07-1999



put to stay low ensuring that thelights stay off.

As darkness falls, TR2 willbegin to switch off and the volt-age at point A will rise. When

the upper threshold is reached,the output of IC1a will go low,causing the output of IC1b to gohigh, thus switching the lightson. (Note, of course, that capac-itor C1 will have been chargedsince the output of IC1a re-mains high when it is disabledso that this input will also behigh.)

Both ends of the potentialdivider formed by R1, VR1 andTR2 will now be high so point A

will also be high, irrespective ofthe light level being sensed byTR2. Only when the lightsswitch off again at the end ofthe cycle (when IC1a outputgoes high) will the ambient lightlevel be sampled again, ensur-ing that the lights beingswitched are ignored.

POWER SUPPLY Power for the circuit is de-

rived from the normal house-hold mains supply. (Note that this article assumes a UK mains supply, Ed.) Capacitor C4 andZener diode D4 drop the mainsvoltage to around 8V with D3rectifying and C3 smoothing theoutput to give a DC supply of

just over 7V. This value willvary depending on the currentdrawn by the circuit, especiallywhen the triac is switched on,but will have no effect on thecircuit operation.

Resistor R7 is included toprotect the Zener diode from thehigh current, which could flowwhen the circuit is first con-nected to the mains with capaci-tor C4 discharged. Resistor R8serves to discharge C4 whenthe circuit is disconnected fromthe mains, ensuring that high

and potentially lethal voltagesare not stored by the capacitor.

The LED D5 is included toshow when the triac is beingtriggered, to assist in setting up

the circuit if no load is con-nected.

An on/off switch may beconsidered an unnecessary lux-ury, as there will normally beone fitted to the socket intowhich the device will beplugged. However, an s.p.s.t.switch connected across thetriac (as shown in Fig.4) couldbe useful to select On or Auto-matic mode allowing the unit tobe plugged permanently into a

socket and by-passed if not re-quired, allowing the light to beswitched on and off manually.The switch should, of course, berated for mains applications.

CONSTRUCTION Because this circuit is

mains powered, it is not suitablefor beginners. If you are in anydoubt about connecting it cor-rectly, consult a qualified electri-

cian.The circuit is assembled on

a small printed circuit board(PCB), whose layout and trackdetails are shown in Fig.5. Thisboard is available from the EPE Online Store (code 7000235) atwww.epemag.com

The PCB should be assem-bled and checked carefully be-fore making any connections to

the mains supply any mis-

takes could easily result in the

instant destruction of your workand many of the components.

Construction should follownormal practice, with low profilecomponents such as resistorsand diodes being mounted firstbefore progressing to the tallerdevices such as capacitors.Note that capacitor C4 rests

above resistors R7 and R8. Asocket should be used for IC1.

There is one wire link on theboard between IC1 and the ter-

minal block

a short piece ofcomponent lead trimmed from aresistor will serve.

Take care to mount all po-larity sensitive componentssuch as diodes, electrolytic ca-pacitors and semiconductors theright way around. TransistorTR1 may be almost any smallsignal pnp type, such as aBC212, BC327, 2N3702,ZTX500 etc., but make sure thatthe type chosen has the same

lead configuration as the devicespecified (see Fig.5).

Phototransistor TR2 ishoused in a clear 2-pin 5mmLED-type package with the baseconnection being omitted. Thecollector connection is the oneadjacent to the flat portion onthe rim of the package (asshown in Fig.5).

Other phototransistors couldpossibly be used, although thishas not been tried. The onespecified has the great advan-tage of being cheap and easy tomount in a panel as a standardLED clip can be used. Althoughit is shown mounted on thePCB, it will probably be bestmounted in the side of the boxand connected to the PCB byflying leads.

POWER RATINGS Two components worthy of

special note are capacitor C4and triac CSR1. Both of thesemust be rated for mains opera-tion, which means that C4should have a voltage rating of250V AC and CSR1 a rating ofat least 400V AC.

Note that on many capaci-tors the voltage ratings usually





7/23/2019 EPE_07-1999



refer to DC values and thebreakdown voltage on AC sup-plies may be very much lower.A Class X device rated for con-necting directly across themains supply must be used forC4.

Because of the limited cur-rent which this kind of powersupply circuit can provide, it isimportant to use a sensitive-gate triac to ensure that the de-

vice will trigger in all situations.The TIC206D specified has amaximum trigger current re-quirement of 5mA and any re-placement should also have thisvalue or less.

In practice, this will meanthat devices with a load currentrating of 3A or less will be suit-able, limiting the maximum loadpower to 750 watts in the caseof the device specified.

Even this can only beachieved with the devicemounted on a suitably largeheatsink and a more realisticmaximum power of 200W to300W should be respected if no

heatsink is fitted. This should inany case be more than suffi-cient for the intended applica-tion where a 40W or 60W lampwill probably be used. No provi-sion for a heatsink has beenmade on the PCB or in the sug-gested plug-box.

VARIABLE TIMING Variable resistor VR1 is

specified as a preset compo-nent and determines the ambi-ent light level at which the cir-cuit will begin to switch. Thereshould not be any need to keepadjusting this, especially if theunit is always mounted in thesame location, but it can be re-placed by a panel mounted po-tentiometer of the same value ifrequired.

If you decide to do this,choose a device with a plasticspindle and connect it to the

PCB with flying leads. As men-tioned earlier, the On time isdetermined by capacitor C1 to-gether with R2 and R3 while theOff time depends on the valueof resistor R2.

As these times are not criti-cal, other values for these com-ponents could also be used, al-though for realism the resulting

times should be measured inminutes rather than seconds,and the Off time should be rela-tively short compared to the Ontime.

The nature of the circuit en-sures that the On time will al-ways be longer than the Offtime (provided diode D1 is notreversed) and with the compo-nent values specified these willbe around nine minutes on andalmost two minutes off. Longertimes may be achieved by mak-ing C1 and/or R2 and R3 larger.

There is little advantage inmaking these times variablebut, again, panel-mounted po-

tentiometers (with suitable se-ries resistors) could be used ifrequired.

The relatively long timeconstants can make testing thecircuit a bit of a chore and it istherefore recommended that,initially, the circuit is built with

C1 equal to (say) 4 7mF and

once circuit operation is provedthe larger specified value canbe substituted.

FIRST TESTS With safety in mind, it is

probably best to temporarily by-pass capacitor C4 with a wirelink and power the circuit from a9V battery or a low voltagepower supply by connecting thepositive to terminal L and thenegative to N. This part of thetesting should be done beforemounting the unit in i ts plug-box.

No load should be con-nected to the triac as the LEDcan be used to indicate the out-put status.

This approach will enablethe circuit to be checked and setup at a low safe voltage beforeit is connected to the mains sup-ply for a final check. Rememberto remove the C4 bypass link


R1 10kR2, R8 1M (2 off)R3 4M7

R4 22kR5, R7 1k (2 off)R6 470 ohms


All 0.25W 5% carbon film

CapacitorsC1 100u radial electrolytic, 16VC2 100n ceramic plate, 63VC3 470u radial electrolytic, 16VC4 100n 250V AC Class X

SemiconductorsD1 to D3 1N4148 signal diodeD4 8V2 400mW Zener diodeD5 red LEDTR1 BC558 pn p transistorTR2 BP103B phototransistorCSR1 TIC206D sensitive-gate triac, 400VIC1 4093 quad 2-input Schmitt NAND gate

MiscellaneousS1 s.p.s.t. insulated switch, mainsrated (see text);

Printed circuit board availablefrom the EPE Online store , code7000235 (www.epemag.com);LED clip; 3-way terminal block;14-pin DIL socket; plug-box with

integral mains outlet socket, in-linefuseholder and fuse (see text);mains-rated connecting wire;solder, etc.

$16Approx. Cost Guidance Only (Excluding Case)

PotentiometerVR1 470k min. horizontal preset

http://0799q.pdf/

http://0799q.pdf/







http://0799q.pdf/


7/23/2019 EPE_07-1999



ENCLOSURE The unit is intended for

fitting into a plug-box, whichalso has an integral mainsoutlet socket. Since the cir-cuit requires both live andneutral mains connections, itis not suitable as a replace-ment for a wall l ight switch. Ifan alternative case is used, itmust be fully insulated andno metal parts (i.e. PCB fix-ings) should pass through thecase walls.

The interwiring betweenthe PCB and the plug-box isshown in Fig.6. Mains rated

cable of an amperage suitableto the lamp being poweredshould be used.

The phototransistor can be

mounted in the box by drilling asuitable hole and using a panelmounting LED clip. A suitablehole should also be cut for

switch S1, positioned so thatthere is no danger of the switchtouching the plug or socket ter-minals. Ensure that the switch issecurely mounted. Note that nometal parts should pass through

the case walls.

For safety, an in-line fuse

(rated to suit the lamp used

e.g. 1A or 2A) should be in-cluded in the live lead if one isnot already fitted in the mainsplug.

FINAL CHECKS Remember that this circuit

operates at mains potential.

Check all connections carefullybefore switching on the supplyand do not touch any part of thecircuit when it is on.

Having satisfied yourselfabout the correctness of yourcomplete assembly, plug thecircuit into the mains. Using amains-insulated screwdriver,adjust VR1 to set the level ofbrightness (or rather darkness)at which operation will start.

When IC1a output switches

high in high ambient light condi-tions, as well as switching thelight off, capacitor C1 will con-tinue to charge to the supply

235

Fig.5. Component layout and (approximately) full-size PCB track details. Plus pin-outs for

TR1 and TR2.

Constituent parts of the specified plug-box (left) and (right) two halves of the box, with outlet socket fitted, together with the completed PCB.

before mounting the unit in itsbox and connecting it to themains.

7/23/2019 EPE_07-1999



rail. This means that the first Onperiod following a reduction inthe ambient light level will bemuch longer than normal, be-cause C1 has to discharge from100 per cent of the supply railinstead of 60 per cent of the

supply as occurs on subsequentoccasions.

IN USE The unit can be positioned

anywhere in the room (using amains extension cable and con-

nector if necessary), preferablyfacing a window and notshielded by furniture or otherobstacles so that the light leveloutside can be sensed, althougha darker location can be com-

pensated for to a certain extentby suitably adjusting VR1.

http://0799k.pdf/

7/23/2019 EPE_07-1999



The frequencies that arebeing used in today’s radio andelectronics circuits are alwaysincreasing. It was not manyyears ago that anything above afew hundred megahertz wasconsidered very high, and spe-cial techniques were required. Anumber of developments sincethen have enabled frequenciesat UHF and higher to be used

far more easily, and at muchless expense. One of thedrivers for this has been thephenomenal growth in the cellu-lar telephone business. How-ever, this is not the only growtharea. There are many new wire-less applications being intro-duced, many of which we do notsee directly in everyday life, yetdespite this we still reap thebenefits of them.

The success of radio and

wireless products means thatthe lower frequencies in the ra-dio spectrum are becoming verycongested. This has resulted inhigher frequencies having to beemployed. One very good ex-ample of this is the cellular tele-phone systems that operate atfrequencies around 1800MHzinstead of around 900MHz whenthe first systems in the UK werepositioned.

THICK AND THIN To be able to manufacture

equipment cheaply that wouldoperate at these frequencieshas forced the development ofa number of new techniques.The level of integration that canbe achieved in a cellular phonethese days is considerably

greater than was possible evena few years ago.

The frequencies that can behandled by ICs has also risen.However, with frequencies insome services rising evenhigher new techniques are re-quired now more than ever.This requires further improve-ments to be made. Often cir-cuits built on what are termed

thick and thin films are needed.Both of these techniques havebeen around for many yearsand involve the printing of cir-cuits onto low loss substrates.

In terms of performance,thin film technology has beenthe preferred choice for manyRF applications because of itssuperior performance. Thesedesigns are usually generatedusing sputtering and etchingtechniques on alumina sub-

strates. In contrast, thick f ilmtechnology has the advantagethat it is substantially less de-manding on both equipment andthe environment than thin filmand it is therefore muchcheaper. As a result there hasbeen a need to develop newtechniques to improve the per-formance of thick film technolo-gies.

Great care must be takenwith these circuits if the required

performance is to be realized.To accommodate the high fre-quencies being used, signalpaths must be very carefully de-signed otherwise performancewill be degraded. Not only mustthe circuit itself be correct, butalso many aspects like the ma-terials used, the conductor sizeand material and many other

associated conditions. If theseare not optimized correctly thenthe signal will suffer from noise,crosstalk, attenuation, propaga-tion delays and other problems.

REQUIREMENTS

To enable the circuits toperform to their requirements,there are a number of condi-tions that need to be fulfilled.The connections between pointson the circuit require a high con-ductivity, their thickness andwidth need to be controlledcarefully as they determine thecharacteristics of the way thesignal travels. Also the sub-strate must have a very lowloss.

To enable the correct char-acteristics to be achieved, ultra-fine lines are required. This,combined with the other require-ments is often difficult toachieve. The ultra-fine l ines,combined with the low levels ofresistive loss means that ex-tremely high levels of conduc-tivity are required.

To meet these require-ments, researchers at the multi-national company DuPont havedeveloped a high density thickfilm gold conductor process.The tracks that this can create

provide low losses at frequen-cies up to 20GHz whilst beingcapable of being etched tothicknesses of less than 20 mi-crometers.

The new material for thisprocess is designated QG150,and it is compatible with a widerange of substrates and thickfilm dielectrics. The new mate-

NEW PROCESSES PROVIDE IMPROVED PERFORMANCE FOR THICK FILM

CIRCUITS, REPORTS IAN POOLE.

7/23/2019 EPE_07-1999



rial is based on gold. In itself theuse of gold is not new. Themetal is used in a number ofareas, including on connectorsand within ICs themselves. Itoffers very high levels of con-ductivity, and is very stablechemically. This makes it anideal metal for use in electron-ics and as a result it is widelyused despite its cost.

Gold is also used in thin filmtechnology where it is also ca-pable of producing very fineconductors, but the lower costof thick fi lm technology is at-tractive to manufacturers whoare now coming under increas-ing pressures to improve perfor-mance, reduce prices and re-main competitive.

NEW PROCESS

There are three key areasthat were addressed to enablethe new process to succeed withthick films. The metal that wasused had to be in a suitablepowdered form. An organicbinder was needed to keep thepowder in place, and finally an

organic material was needed tocarry the material during theprocess.

The gold powder requirednew development work. It is anew form of powder for whichthe size and composition hasbeen carefully chosen to givethe required properties duringthe process, and afterwards inuse. The actual grains are ex-tremely small, being only frac-tions of a micron in diameter.With micro-grains of this size itis possible to achieve the re-quired line size, thickness, andetching characteristics.

A material was required tobind the micro-grains togetherin the process. Experimentswere undertaken to ensure ithad the correct properties, en-

abling the required bondstrength so that the wire ad-hered to the substrate whilst stillproviding the correct pre- andpost-etching capabilities.

The third and final con-stituent was the carrier. Thiswas an organic material and ithad to provide the right viscos-ity for enabling the gold micro-grains and the organic binder topass through the processingcorrectly. All three constituentsneeded to be optimized sepa-rately and together to producethe correct result.

RESULTS

Naturally in proving thetechnique many tests havebeen performed. It has beenshown that losses for the newprocess are consistently lowerover frequencies up to about10GHz than a comparable thinfilm process. Above this the re-sults are reasonably similar.However, as thick film technol-ogy is preferred because of thereduced costs, it means that thenew technique is likely to be

very successful.

In another demonstration ofthe success of the product, acompany named Micro-Precision Technologies ofSalem New Hampshire USAdeveloped a single substratesurface mount attenuator andphase shifter based on thick filmtechnology for use in cellularradio systems.

Using the DuPont’s QG150

on an alumina substrate alongwith other materials to enablethe specialized surface mountICs to be placed down onto thesubstrate and to enable solder-ing with the through holes re-quired for interconnectivity.

It was found that the newdesign showed significant im-provements over earlier ver-

sions manufactured using thinfilms, and in addition to this themanufacturing costs were signif-icantly reduced.

With the use of higher fre-

quencies increasing, along withincreased requirements arisingfrom the high levels of complex-ity and performance required,new processes like the DuPontQG150 will be in much greaterdemand. As many more wire-less applications are being intro-duced into the home, the costconstraints are also becomingmuch tighter. Any processesgiving improvements in theseareas are likely to be taken upvery quickly.

http://0799l.pdf/

7/23/2019 EPE_07-1999



Let us sift through thismonth’s post-box, starting withthis query concerning the use ofthyristors:

Gate Post

How can you calculate the

required pulse and duty cycle width for triggering the gate of an SCR or triac? Satish Kumar (by email).

An SCR (silicon controlledrectifier) or thyristor is a four-layer pnpn device (p and n typesemiconductor) as shown inFig.1a. Compare this with abipolar junction transistor (BJT)

which is three-layer either pnp

or npn .

The thyristor can be thoughtof as two overlapping transistors

the np of a pnp transistor

overlapping the np of an npn type, as indicated by the dottedboxed in Fig.1a. This leads tothe transistor equivalent circuitin Fig.1b.

The circuit symbol of athyristor is shown in Fig.1c. It isbased upon that of a rectifier,which is a unidirectional device,

meaning that current will onlyflow one way. The symbol for atriac is shown in Fig.1d.

The illustration in Fig.2shows a basic thyristor circuit.Assume the thyristor, CSR1, isnot conducting (no anode-cathode current), and also thatno gate current is flowing. Whenthe control unit produces a trig-

ger pulse, a small gate currentcauses a much larger anode-cathode current to flow.

However, unlike a transis-tor, which turns off again if thebase or gate current is re-moved, the thyristor’s anode-cathode current continues to

flow. It will only stop when theload supply voltage is re-moved or if the anode-cathodecurrent drops below a certainminimum level. Thus the trig-ger current causes the thyris-tor to latch on.

We can understand thisbehavior by looking at theequivalent circuit of the thyris-tor in Fig.1b. The “trigger” gatecurrent turns on transistor

TR1. The collector current ofTR1 provides a base current forTR2, turning it on too. In a simi-lar manner the collector currentof TR2 provides more base cur-rent for TR1 turning it on evenmore. This is a positive feed-back effect that quickly ensures

that both transistors are on.Once this condition has

been triggered by the gate it isself-sustaining, so gate currentis no longer needed. The thyris-tor can only be turned off by re-ducing its anode-cathode cur-

Our team of “surgeons” tackles a further range of readers’ queries sent in by post or email. We look at thyristor basics, re-visit the topic of mains earthing, check out capacitor fundamentals, and more besides in our monthly help-desk round-up

by ALAN WINSTANLEY

Fig.1. (a) Four layer pnpn construction of a thyristor

(SCR), (b) equivalent circuit,(c) circuit symbol for an SCR

and (d) for a triac.

Fig.2. (a) Basic thyristor circuit. Thyristor CSR1 remains in conduction once a suitable

triggering signal has been received. (b) The addition of external RC “snubbers” helps

avoid false triggering.

7/23/2019 EPE_07-1999



rent below some critical point,known as the holding current.

The voltage across a thyris-tor when it is ON typically has a

minimum value of around 1V,but may be higher (2V to 3V) forhigh current devices; the ONcurrent can be very high (tensof amps in high power devices).

The OFF current is very small

a leakage current. The maxi-mum OFF voltage (supply volt-

age) can be very high hun-

dreds of volts in high power de-vices.

Consultation

To answer your questionspecifically, we have to consultthe data sheet for the device wewant to use. There are a largenumber of thyristors to choosefrom, but we will use theBT149B as an example. This isa general-purpose low powerdevice from Philips, intended tobe interfaced directly to logicICs and other low power triggercircuits. The BT149B can han-dle OFF voltages of about 200V

and average ON currents ofabout 0 5A.

We need to consult the

switching information turn-on

and turn-off times. For theBT149B, the turn-on time is typ-ically 2ms and the turn-off timeis typically 100ms. The triggerpulse must be long enough toallow the device to turn on (2msin this case).

The duty cycle depends on

when we expect the device toturn off the trigger should

have been removed by this timeor the device will re-trigger. Onthis basis the shortest cyclepossible would be 102ms (2mson, plus 100ms off), but this as-sumes the thyristor is switchedoff at the same time the triggeris removed, which is not usually

terminals (Main Terminals 1 and2). Similar arguments would ap-ply to the switching times of the

trigger pulse it helps to con-

sult the data sheet for details.

Note that a trigger pulse has tobe provided for each phase ofthe waveform.

Look out also for a diac a

bi-directional diode with specialcharacteristics, which may beused for triggering triacs intoconduction. IMB .

Stripboard and High Voltages

Stripboard prototyping

board, of which Veroboard isthe best-loved brand, is theideal medium for developingsimple discrete circuits. Al-though many prefer to developprinted circuit boards using aCAD system, there is still muchto be said for the convenienceand adaptability of good oldstripboard, especially if the cir-cuit has not been finalized.

Provided the circuit is nottoo complex, stripboard makes

it easy to add components ormake other changes during thedevelopment process. It’s alsoideal for beginners and is per-fectly adequate for constructingmany projects in their final form,and has been for thirty years orso.

It does raise questions ofsafety, though, when used withmains voltages:

I sometimes use stripboard for projects rather than spending time making printed circuit boards. These can involve 230V AC mains voltages. Can anyone provide details of the maximum load capacity of copper strips? (Posted by Murray Cameron in the EPE web site Chat Zone.)

the case.

Usually the trigger is a shortpulse that is long enough toguarantee triggering and the

time between pulses is deter-mined by the control circuitrythat generates the trigger, whichof course depends on the appli-cation.

Note that thyristor switchingtimes are different for different

devices compare the 2N5064

from Motorola (also rated at

about 200V, 0 5A) for which

turn-on takes about 3ms andturn-off 30ms.

The gate current is, unfortu-nately, not the only way to turnon a thyristor. A sufficiently fastrising anode-cathode voltagecan also trigger the device, dueto the capacitances inherent inthe thyristor’s structure. To pre-vent this, RC “snubber” circuitscan be used to reduce the rise-time of voltages across thethryristor, as shown in Fig. 2band similar treatment may bemade on the gate terminal insome cases.

Incidentally, unwanted (orparasitic ) thyristor structurescan occur in CMOS chips. Inte-grated circuit fabricators andchips designers take care tomake sure that these unwantedthyristors cannot switch on, be-cause if they do they “crow-bar”or short-circuit the power supplywith disastrous consequences.The problem is known aslatchup.

Triac By placing two thyristors

“back-to-back” (in inverse paral-lel), a triac is formed, which issuitable for use with AC controlcircuits (AC load control, sound-to-light systems, etc.) becausecurrent can flow through it ineither direction between the two

7/23/2019 EPE_07-1999



Stripboard is now made byseveral manufacturers of whichVero is the most famous, but italso comes from Far Easternsources. This means there is no

across-the-board (sorry) stan-dard specification. However for

typical 0

1-inch matrix strip-

board, assume a maximum cur-rent of no more than roughly5A. The breakdown voltage be-tween strips is said to be about800V peak, absolute maximum.

You can use stripboard withmains voltages provided the fol-lowing precautions are used:

1. Ensure all stand-off pillars

are non-conductive typeswith no possibility of achassis or mounting screwbeing in contact with themains voltage. Use fully insulated nylon mounting- hardware.

2. If a break needs to bemade in mains-carryingcopper strips, make atleast three continuous breaks adjacent to eachother to completely re-

move all copper from thatsection of the board andfully isolate the mains strip.

3. Mains-carrying strips canbe reinforced by solderinga length of tinned copper

wire along the strip.

4. It is best to use PCBscrew-terminal blocks tomake a safe mains-voltage“flying lead” connection.

It is preferable to use aquality branded product for cir-cuitry involving mains currents.It is probably far safer to keepany heavy mains currents offthe board, and use suitablyrated fully insulated hook-upwire instead. ARW.

Live Supplies

It’s back to the topic of Live,

Neutral and Earth. What dothese designations actuallymean?

There are three parts to an AC socket labeled “hot” or Live,Neutral, and Ground. I’m trying to make sense of what these mean. Can someone provide a good explanation of the concept of “Ground”? asks HMB via theInternet.

The meaning of “Ground”depends on its context. In anelectronic circuit “ground” nearlyalways means “0V” and any-thing “grounded” is connectedto 0V.

Elsewhere in the circuit dia-gram (say in the power supplysection) the 0V rail will beshown as connecting to“ground”, in the same way that

the negative pole of a car bat-tery may be depicted as con-nected to the “chassis.” It savescluttering up wiring diagramswith lots of lines. In a circuit, aground symbol usually justmeans that everything with theground symbol is connected to-gether.

This does not usually meanthat the 0V rail is directly con-nected to physical earth (soil),which is the second meaning of

“grounding”, although the 0V railmay sometimes be connectedto the ground as well. The term“ground” as used in the USA issynonymous with “earth” in theUK and Europe. In an electricalinstallation, grounding means,physically connecting to earth(soil).

Regarding why the mainssupply is labeled the way it is,Fig. 3 shows a typical mainssupply which would be fed to a

domestic installation by an un-derground cable. At the “powersupply” end (e.g. the trans-former/sub-station), the neutralis firmly connected to the earth.This is “grounding” in the electri-cal sense of the word, and it im-plies that the neutral and earthwires are linked. In the USA, theneutral is the grounded conduc-tor whilst the earth wire is thegrounding wire: live is the un- grounded wire.

The live wire alternates be-tween +325V peak and -325Vpeak. A quick calculation of 325divided by the square root of 2gives a value of 230V RMS (the“official” UK mains value). Be-cause of the way in which three-phase electricity is generated,the neutral point has no voltageand the live wire is the one that

Fig.3. How the Live, Neutral, and Earth terminals of a domestic installation in the UK are connected to the incoming supply.

The “Protective Earth” may not be present in overhead supplies.

7/23/2019 EPE_07-1999



carries a potential. Appliancesare therefore connected be-tween live and neutral at thedomestic installation.

Hot Line The term “hot” is American

slang for “live”. If you, as a hu-man being, were to touch a livewire, then assuming that youare in contact with the ground, apotentially lethal current wouldflow from the “hot” or live,through your body and to earth.This is because the earth is con-nected to the neutral as shownin Fig.3 and so the “escaping”current will seek to complete a

circuit back to the neutral pointthrough the earth.

If an RCD (Residual Current

Device) also known as an

ELCB (Earth Leakage CircuitBreaker) or GFCI in the USA(Ground Fault Circuit Inter-

rupter) is installed, then the

current flowing to earth will bedetected and a high-speed tripswitch will operate to hopefullyprevent injury. Without the ben-efit of such a device, a fatal

shock may be received.

The purpose of the ground/ earth connection is to provide avery easy path for leaking cur-rent to “escape” through. If aninsulation fault arises and a livewire happens to touch exposedmetal work, then instead of theuser receiving a shock when hetouches it, a large current willflow straight to ground and ei-ther melt a fuse or trip an RCD.

(The LM1851 is a NationalSemiconductor chip whichforms the heart of a GFCI,should advanced readers careto consider making their own,perhaps for incorporation into amains-powered project.)

It is obvious that there aremany misconceptions surround-ing the way in which electricity

is generated and delivered. Infact many of us know how to use the mains supply, but a greatmany more are not entirely surewhere electrical power comesfrom to start with.

Next month I intend to set therecord straight with the start of atwo-part feature showing howpower is generated and transmit-ted to the home. The author hasspent many days as a guest ofthe renowned international powergroup National Power PLC(www.national-power.com) whogenerously provided the authorwith unrestricted access to an en-tire power station (NationalPower’s Killingholme “A” gas-firedplant near Immingham in Lin-colnshire, UK), as well as severalNational Power engineers all ofwhom helped enthusiastically withthe research.

The entire process from gaspipeline, to the turbines and gen-erators all the way through to theprovision of the domestic “230V”AC is described and illustrated infull. The feature will answer theabove questions and more be-sides in greater depth: readerswill find it fascinating stuff, so besure to read From Pipelines to Pylons starting in the August 99issue! ARW.

Currents and Dielectrics

I’m a frequent reader of EPE,and one of my favorite columns is Circuit Surgery. It’s amazing how much one can learn from such articles, and fantastic that someone is ready to read and possibly answer queries that we newcom- ers to the world of electronics might have. (Thanks. The words“flattery” and “everywhere” springto mind! We try to answer everyquery but we can’t alwayspromise a personal reply unfortu-nately.)

I’ve got three basic questions to ask:

1. Everyone knows that a capacitor conducts current briefly, but I don’t under-

stand how current can flow between two plates if they are separated by a dielectric.

2. Can you help with the physical significance of logic gates? In particular,AND and OR gates. These are constructed with transistors, but how is it that one can say that a particular circuit is an AND gate.

3. Finally, what is the significance of “double insulation”?

Regards to all at EPE! KarlVassallo Grant from Malta, via the Net.

There’s quite a lot to go at,so let’s answer your queries inorder. First, you are quite cor-

rect the two capacitor plates

are separated by an insulating

layer, called a dielectric. Cata-logs classify their capacitors bydielectric, because differenttypes of dielectric (polyester,polypropylene, silver mica andso on) are more suited in someapplications than in others.

Let us use the tried andtested hydraulic analogy; a bat-tery becomes a tankful of water.Imagine an electrical circuit asbeing a sealed system , usingwater travelling via the tank and

through a hosepipe. Imaginealso that the capacitor consistsof a rubber diaphragm, which isplaced in the circuit. There is“solid” water in the hosepipe onboth sides of the diaphragm,see Fig. 4.

Water cannot pass through the diaphragm, but it can

7/23/2019 EPE_07-1999



“stretch” it. If we squeeze thehosepipe to compress the wa-

ter, this forces the water upagainst the diaphragm, whichstretches “outwards”.

Because it is a sealed sys-tem, the water on the other side

has to go somewhere so it’s

forced further around thehosepipe by the movement ofthe diaphragm. Although no water passed through the diaphragm, never the less, water was seen to move around the circuit for a short time.

A capacitor operates inroughly the same sense. In the“sealed system” of an electriccircuit, electrons have to comefrom somewhere and go some-where! A flow of electrons con-stitutes an electrical current.Adding a charge onto one plateof a capacitor (say, by “sucking”

it from one pole of a battery)causes a corresponding charge tobe “sucked” from the other plate,and piled back on to the remain-ing pole of the battery.

It is a sort of electronics opti-cal illusion, and is the reason whywe say that capacitors can allowa current or signal to “pass”through. What we really mean isthat a charge flows onto the ca-pacitor on one plate and a corre-sponding charge flows off theother side. Thus a current is seento flow, even though there is adielectric breaking the circuit. It’snot possible to analyze this ingreater detail without looking atthe physics of the capacitorthough.

In Truth

With regard to your secondquery: I guess you are referring tothe “Truth tables” of basic logicgates. These summarize how alogic gate will react with a particu-lar combination of inputs. For ex-ample, an AND gate output goeslogic high when all its inputs arehigh; an OR gate goes high whenany of its inputs are high. Othergates such as NAND, NOR andEXOR have their own uniquetruth tables as well.

These were fully described in

our series Teach-In 98 An Intro-

duction to Digital Electronics ,which was co-written with the Uni-versity of Hull and which ap-peared in the November 1997 toSeptember 1998 issues of EPE.Logic gates and how truth tableswork, were explained in Part Four

(February 1998). Truth tables ex-ist for more complex logic de-vices and will be found in manu-facturers’ data sheets, whichthese days are commonly avail-able on the World Wide Web.

Double insulation

Finally, “double insulation”merely implies that extra precau-

tions have been taken by amanufacturer to fully insulateany live parts, and to make surethat there is no possibility of theuser coming into contact withlive wires. For reasons ex-

plained in the previous question(Live Supplies ), the earth isused to provide a low-impedance route for any faultcurrents to flow. Extra insulationand the widespread use of plas-tic moldings removes the possi-bility of any external parts everbecoming live (e.g. through alive wire coming adrift), so thereis no longer a need to “earth”the equipment.

Double-insulated units carrya symbol of two concentricsquares (check any mainsadapter to see) and have a two-core power cord. Typical exam-ples of double-insulated equip-ment include power and gardentools, consumer video, TV andaudio units, and most kitchenappliances. However, any con-structional projects that aremains-powered almost alwayshave a compulsory earth wire,which will be connected to the

mounting frame of the mainstransformer and elsewhere.

All exposed metal parts,including switches andbezels, metal panels andmounting screws, which canaccess the interior of amains-powered project,MUST be properly connectedto earth to guard against thepossibility of them ever be-coming live in the future. This

protects you and other users from electric shock. ARW.

Fig.4. Using the tried and tested hydraulic analogy

where the battery becomes a tankful of water and a capac-

itor the diaphragm. Water,sealed in a hydraulic “circuit”,

causes the diaphragm to stretch, which displaces water in the circuit. It appears

that water has flowed “through” the diaphragm.

http://0799m.pdf/

7/23/2019 EPE_07-1999



Oscillators are crucial to theworking of all but the mostmundane of electronic

equipment. Like beating hearts,they send out the impulses thatgive life to radio transmittersand receivers, computers,frequency counters, calculators,timers, oscilloscopes, and signalgenerators of all kinds.

DEFINING AND CLASSIFYING

Exploiting the negativeresistance of an arc (“spark”transmitters), or the use of analternator, or some otherelectro-mechanical device togenerate oscillations, has beenrelegated to history. For ourpurpose, therefore, an oscillatorcan be defined as an electroniccircuit that converts directcurrent (DC) to alternatingcurrent (AC), and an oscillationis the periodic variation ofcurrent or voltage in the circuit.

Defining oscillators is onething, classifying them is quiteanother. Some authorities placethem in two categories:feedback or negative resistance .Whilst there are circuits whichclearly depend on thephenomena of negativeresistance, any oscillator can beanalyzed from either standpoint,

and this approach can be -confusing.

Another method of

classification depends onwhether the circuit action is ofthe harmonic or the relaxation type. Oscillators in the formercategory generate more or lesssinusoidal waveforms, whilstthose in the latter usuallyproduce a square or sawtoothoutput. Harmonic and relaxationoscillators have fundamentallydifferent circuits, and this leadsus to a more useful method ofclassification.

All oscillators comprise twobasic elements: a frequencydetermining section and anamplifier or negative resistancedevice, which maintains theoscillations. The passive,frequency determiningcomponents are of three kinds:an inductor and capacitorforming a tuned circuit; a quartzcrystal or ceramic resonator; orresistors and capacitors. In thisseries of articles, the various

circuits will be considered underthese three headings.

BRIEF HISTORY The invention of the triode

valve was crucial to thedevelopment of all three classesof oscillator, and most of thecircuits in use today were

developed during the valve era.

In 1906, Dr Lee de Forestproduced the f irst electronic

amplifying device by adding acontrol grid to Fleming's diodevalve. Lee de Forest named hisinvention The Audion , and itwas Eccles (of f lip-flop fame)who subsequently called it atriode.

When attempts were madeto use the new device for RFsignal detection andamplification, its ability tomaintain oscillations in a tunedcircuit soon became apparent.

This phenomena was exploitedby Armstrong, de Forest,Meissner, and others, and, from1913, several patents for valveoscillator circuits were filed bothin Europe and America.

In 1924, after seven yearsof acrimonious litigation, controlof oscillating valve patents inAmerica was awarded, on alegal technicality, to de Forest.In Europe, patents were held bya German, Alexander Meissner,and an Englishman, Captain H.J. Round. It was during theseearly years that Colpitts,Franklin, Hartley, and othersdevised their classic LC (inductance/capacitance)circuits which, together withlater variants, have survivedinto the transistor age.

This new series is prepared with the electronics enthusiast and experimenter very much in mind and is intensely practical. Tried and tested circuits are fleshed out

with component values, and their vices and virtues are exposed.

PART 1: THE HARTLEY OSCILLATOR AND ITS VARIANTS

By RAYMOND HAIGH

7/23/2019 EPE_07-1999



The multivibrator , a two-valve circuit in which resistorsand capacitors fix the operatingfrequency, was first describedby Abraham and Bloch in 1918.Gill and Morrel published details

of their work on negativeresistance oscillators in 1922.The following year, W. G. Cadydemonstrated to G. W. Pierceof Harvard University the way inwhich a quartz crystal can

determine the frequency of avalve oscillator. Within a fewmonths, Pierce had developedhis own circuits of this kind.

All of the basic oscillators

Fig.1. Simplified circuits for series and shunt (parallel) fed Hartley oscillators: (a) series fed, (b)alternative series fed, and (c) shunt (parallel) fed. With series feed, power is supplied to the

transistor via the feedback winding.

Table 1 and Table 2 (later) give details of coil and variablecapacitor combinations which resonate over a 150kHz to30MHz range. Some experimenters will, of course, wish todesign their own tuned circuits to resonate at particular spotfrequencies, to cover different ranges, or to make use ofcomponents that are to hand. The basic formula relatingfrequency to inductance and capacitance is:

As a rough guide, assume the following tuning capacitances whencalculating the inductance required for a particular spot frequency:

Tuned Circuit Calculations

where f is in Hertz, L is in Henries, and C is in Farads.Inserting the value of pi gives:

Henries and Farads are unwieldy for our purpose, and theformula is more conveniently expressed in smaller units.Accordingly, when f is in kHz, L is in mH, and C is in mF:

and when f is in MHz, L is in mH and C is in pF:

f =1

2 LC

f =0.159

L =0.025

f2CC =

0.025

f2LLC

f =159.155

L =25.33

f2CC =

25.33

f2LLC

f =159155

L =25330

f2CC =

25330

f2LLC

Capacitance should be as high as possible consistent withsatisfactory circuit operation.

Inductors for frequencies lower than 50kHz are best wound onferrite pot cores. This ensures a high Q factor, and a knowninductance value per turn simplifies coil design. Resistor andcapacitor (RC) tuned oscillators are generally more suitable forthese lower frequency ranges (these will be covered later in theseries).

The tapping point for the Hartley oscillator can range from 10percent to 50 percent of the total number of turns on the tuningcoil. The usual figure is 20 percent to 25 percent, and this shouldalways be adopted as a starting point when hand-winding coils.Increasing the percentage increases the signal available at thetapping point, and the quoted RMS output voltages assume theuse of the specified coils.

Fixed tuning capacitors should have a polystyrene, mica, or, for thelarger values, Mylar film dielectric. High value ceramic capacitors

have a comparatively low Q, and this can inhibit oscillation.

F or Fr eq ue nc ie s i n t he re gi on of : V al ue of C

100 Hz1 kHz

20 kHz100 kHz

1 MHz10 MHz20 MHz

0.47 uF0.1 uF

0.01 uF4700 pF1000 pF

470 pF220 pF

7/23/2019 EPE_07-1999



we use today were, therefore,conceived within two decades ofthe invention of the triode valve.

VALVES AND TRANSISTORS The gradual shift from

valves to transistors during thesixties and seventies has notbeen without its drawbacks.Valves, for the most part, arefragile, microphonic, power-hungry devices, which generatecomparatively large amounts ofheat. But, despite these failings,they have advantages overtransistors, particularly when

used as the maintaining devicein an oscillator circuit.

Valves are highly linear,and their characteristics andinter-electrode capacitances donot vary much with changes intemperature and signal levels,or moderate shifts in supplyvoltage. The characteristics andinternal capacitances oftransistors, on the other hand,display significant variations.This can make biasing

arrangements morecomplicated and frequencystability more difficult toachieve, despite the minimalheat generation.

The introduction of the fieldeffect transistor (FET) with itshigh gate resistance and morevalve-like characteristics madeit easier to adapt the classicoscillator circuits. However, asfrequency increases throughVHF and into UHF, the gate

impedance of a FET reducesand it becomes more like itscurrent-driven bipolarcounterpart.

A limited range of valves isstill retailed. Whilst some aresuitable for use in oscillatorcircuits, they are fairlyexpensive and require high-

voltage power supplies.Because of this, all of thecircuits described in this seriesincorporate semiconductors asthe active devices. Any

reference to valves is in apurely historical context.

MAINTAINING OSCILLATION

The sudden application of adirect voltage to an LC tunedcircuit initiates, or shockexcites, oscillations. Because ofresistive losses, the oscillationsfade away, their duration andmagnitude being directly related

to the “Q” factor of the tunedcircuit (the higher the Q thelower the resistive losses).

Oscillations can bemaintained by connecting anamplifier to the tuned circuit insuch a way that energy is fedback in phase (i.e., bycontinuing the excitation). Thispositive feedback can beapplied via an additionalwinding inductively coupled tothe tuning coil (Armstrong,

Meissner), or v ia a tapping onthe coil itself (Hartley).

Sometimes the feedback iscapacitively coupled (Butler,Colpitts, Franklin), and it can bea mixture of both (Reinartz).The feedback eliminates orcancels out the resistive losses.Viewed in this way it can besaid to create “negativeresistance”.

It follows that devices which

display negative resistance canalso be used to maintainoscillation, despite the fact thatthey do not amplify. An electricarc, a neon lamp, the Esaki ortunnel diode, all have acharacteristic which exhibitsrising voltage with fallingcurrent; i.e. negative resistance.

They function by canceling

out the resistance in the tunedcircuit so that the oscillationscan continue. When circuits ofthis kind are analyzed from thefeedback standpoint, the

positive feedback is said toexist within the device itself.

TUNED CIRCUITS Before we begin to consider

our first group of oscillators,those where a tuned circuitdetermines the operatingfrequency, a brief review of therelationship betweeninductance, capacitance andfrequency may prove helpful.

A tuned circuit resonates atthe frequency at which thereactances of the inductor (coil)and capacitor are equal. Withincreasing frequency, inductivereactance rises whilst capacitivereactance falls. Resonance atany particular frequency can,therefore, be achieved by arange of inductance andcapacitance combinations; i.e.more L and less C , or more C and less L produces the same

result. When parallel LC circuitsare used in oscillators, it isdesirable to keep the ratio ofcapacitance to inductance ashigh as possible, consistent withreliable operation.

Q-FACTOR The performance of a tuned

circuit is defined by a f igure ofmerit known as the Q factor.High Q tuned circuits resonate

sharply, and oscillatorsincorporating them produce apurer waveform with lowerharmonic content. Moreover,less power has to be suppliedby the maintaining device tokeep the circuit oscillating, andthis is conducive to reliablestarting, stable operation andreduced drift.

7/23/2019 EPE_07-1999



The Q factor of most tuningcapacitors is extremely high,and only the Q of the inductorneed be considered. Valuesbetween 50 and 100 or moreare usual. Unfortunately, someof the ways in which the Q of acoil can be increased havedrawbacks as far as oscillators

are concerned. More will besaid of this later.

The oscillators illustrated allhave parallel tuned circuits.With this arrangement,increasing the ratio ofcapacitance to inductancemakes the Q of the circuit riseand its impedance, atresonance, fall. This last factorimproves the matching to thetransistor, thereby reducingdamping, particularly when a

bipolar device is used.Moreover, the large amount oftuning capacitance in circuitswamps the effect of smallchanges in the internalcapacitances of the transistorand this can help to reduce drift.

HARTLEY OSCILLATOR

Separate feedback windingsare not required with the Hartleyoscillator, and this simplifies coilconstruction and band switchingwhen several frequency rangeshave to be covered. Because

the maintaining amplifier is soeffectively coupled to the tunedcircuit, the arrangementoscillates readily and is veryeasy to get working and toadapt for a variety of purposes.The output can be madereasonably constant over a widefrequency range, and waveformpurity is good enough for mostradio receiver and signalgenerator applications.

With these advantages in

mind, the Hartley oscillator, andits Lampkin and Dow variants,will be the first to be discussed.

SERIES AND SHUNT POWER FEEDS

The Hartley circuit (andothers for that matter) can bearranged in either series or

shunt fed modes. Simplifiedcircuit diagrams depicting thealternatives are given in Fig.1.

Series feeds, so calledbecause the transistor and the

feedback winding are connectedin series across the powersupply, are depicted in Fig.1aand Fig.1b. The shunt (parallel)fed, arrangement, with thepower supply connected acrossthe active device, is given inFig.1c.

Versions Fig.1b and Fig.1cenable the moving vanes of avariable capacitor to beconnected to the negativesupply rail. These are,

therefore, usually the preferredoptions when continuouslyvariable tuning is desired.

SPOT FREQUENCY OSCILLATOR

A bipolar transistor versionof the series fed Hartley circuitis illustrated in Fig.2. It thriveson a high ratio of capacitance inthe tuned circuit. Indeed, above15MHz or so it becomesreluctant to oscillate unless the“tuning” capacitor C3 has avalue of 100pF or more.(Reducing the capacitance in aparallel tuned circuit lowers Q and increases impedance atresonance. The latter makes themismatch with the bipolartransistor even worse, andoscillation becomes less reliableas the active device is operatedcloser to its fT).

Turning to the circuitdiagram of Fig.2, capacitor C2is a DC blocking capacitor,which couples the base oftransistor TR1 to the tunedcircuit formed by coil L1 andcapacitor C3. The coil andcapacitor combination showncan be adjusted to resonate at100kHz. Note that the tapping is

– –

µ . .

. .

Fig.3. Circuit diagram for a simple DC-to-DC Converter using a series-fed Hartley oscillator.

7/23/2019 EPE_07-1999



located towards the end of thecoil furthest from the collector(c) of the transistor.

Resistor R1 and presetpotentiometer VR1 are base (b)

bias components, and presetVR2 sets the bias voltage onthe emitter (e). They enable thebiasing to be optimized for thebest waveform.

The preferred signal take-off point is the transistor baseend of the tuned circuit, as this

minimizes damping. An outputtaken from this point can be fedinto impedances as low as 4kilohms. (The specified coil istapped at less than 20 per cent

of the total turns: larger tappingpercentages will require ahigher load impedance).

This version of the Hartleycircuit oscillates so vigorouslythat the signal can also beextracted from the “hot”, orcollector end of the coil. Thevoltage developed here isgreater, but it has to be fed intoan impedance of 470 kilohms ormore or the circuit may stoposcillating.

Capacitor C1 decouples thepower supply. At lowfrequencies (10kHz down) a100mF electrolytic may have tobe substituted for thiscomponent, and C2 should beincreased to 100nF.

At 1MHz and below, thecircuit is sufficiently drift-free forspot frequencies or markers (at,say, 10kHz, 100kHz and 1MHz),

and their harmonics, to be usedfor receiver calibrationpurposes. Although not asstable as a crystal oscillator, LC instruments of this kind can

produce sufficiently accuratedial calibrations for mostpurposes. The markers must, ofcourse, be set to zero beat witha known standard, or set with afrequency counter.

SIMPLE DC-TO-DC CONVERTER

DC-to-DC convertersrequire an oscillator of somekind to enable the voltage

transformation to be effected.An example of the series fedHartley circuit being used in thisway is given in Fig.3. No claimscan be made for the conversionefficiency of the circuit, but itdoes have the advantage ofgreat simplicity. There can befew experimenters who couldn’tassemble it from their sparesboxes.

. . . .

. . . .

–

Fig.2. Circuit diagram for a series fed Hartley for generating spot or marker frequencies between 100Hz and 10MHz. The LC combination shown can be set to 100kHz (see Fig.3. for

alternative transistors).

Basic Series Fed

Hartley Oscillators

Two examples of the basic

series fed circuit are given inFig.2 and Fig.3.

The arrangement shown inFig.2 works reliably from be-low 100Hz to more than10MHz. Above 15MHz, tuningcapacitors have to be at least100pF or the circuit displays areluctance to oscillate. Above20MHz or so, operation canbecome erratic.

Neither end of the tunedcircuit is connected to ground,

and this makes the arrange-ment unsuitable for continu-ously variable tuning. It is,however, useful for generatingspot, or marker, frequencies,and the sinewave output is ofgood purity. Providedimpedance limitations are ob-served, this oscillator will de-liver an output in the region of4V RMS when connected to a9V supply.

How the series fed circuit

can be used as the basis of asimple DC-to-DC Converter isshown in Fig.3. The iron-coredinductor T1 with its 50 percenttapping point results in awaveform which is a series ofspiky pulses. Frequency of os-cillation ranges between 3kHzand 6kHz or more, dependingon the output current drawn.

7/23/2019 EPE_07-1999



The low voltage winding(secondary) of a miniaturemains transformer is used asthe inductor in a tuned circuitbrought to resonance by

capacitor C3. Positivefeedback, applied via the centertap (CT) makes the circuitoscillate, and a voltage increaseis achieved by the step-up ratioof the transformer (primarywinding). In this form, the circuitgenerates a chain of spikypulses, and the frequency ofoscillation ranges from around3kHz to 6kHz depending on theloading. (DC output currentflows through the transformer

mains winding. As current flow

rises, tuned circuit inductancefalls, and frequency of

oscillation increases.)

Blocking capacitor C2couples the base of thetransistor TR1 into the feedbackcircuit, and bias is applied byresistor R1. Diode D1 limits thenegative going voltage swingson the base of TR1 and greatlyincreases the voltage deliveredby the circuit. Diode D2 andcapacitor C4 rectify and smooththe output and C1 is the DCsupply decoupling capacitor.

Tuning capacitor, C3, needsto be selected for optimum

performance. Its value dependson the actual transformer used,

but will almost certainly liewithin the range 10nF to 100nF.

Miniature mainstransformers appear to give thebest results. With a 9V supplyand a 6V-0V-6V transformer,around 300V will be developedacross capacitor C4. This cangive a very unpleasant shock,and the unit should be handledwith care. Try reversing theconnections to the transformermains winding if the output

voltage seems low (there is anoptimum way of feeding the

. .. . . .

–

Fig.4. Circuit for an add-on BFO Unit using a parallel (shunt)fed Hartley oscillator (see Fig.3. for alternative transistors).

Shunt FedHartley Oscillators

Circuit diagrams Fig.4 andFig.5 show how a shunt fedHartley circuit can be used as aBFO (beat frequency oscillator)unit for radio receivers, and toform a simple metal detector.The arrangement works well be-

tween 10kHz and 1

good sine wave is produced.

5MHz, and a

Operation becomes uncer-tain when the tuning capaci-tance is reduced below 100pF,and also at frequencies above

1 5MHz or 2MHz. If a wide oper-

ating range is required, the se-ries fed circuits illustrated inFig.6 and Fig.7 are more suit-able.

The signal voltage availableat the tapping point (see S1b inFig.6) is dependant upon the

tapping ratio. With the specifiedinductors, it is approximately 2VRMS.

Oscillation may cease if thevalue of the capacitor couplingthe transistor emitter to the coiltap is reduced. Together withthe emitter resistor, this compo-nent forms a potential divider inthe feedback path.

Ω

. .

. . .

–

7/23/2019 EPE_07-1999



spiky pulses to the rectifier).

Output voltage falls rapidlywhen currents in excess of a

few tens of microamps ( A) are

drawn, making the unit safe tohandle despite the high voltage.

With an output current of 0 5mA

the voltage falls to around 100Vand the current drawn from a 9Vsupply is approximately 20mA.Oscillation is maintained withDC inputs down to 3V or less,but output is, of course,

correspondingly reduced. Ifstarting becomes uncertain atvery low supply voltages,reduce resistor R1 to about 22kilohms.

BFO UNIT An add-on beat frequency

oscillator (BFO) will make

beats with Morse to produce anaudible tone, or reinstates themissing carrier in SSBtransmissions so that the set’sdiodedetector can make themintelligible.

The injected signal has to betuned across the receiver’s IFpassband. This is achieved byexploiting the “varicap” action of areverse biased rectifier diode, D1.Potentiometer VR1 functions asthe tuning control by varying thereverse bias. Capacitor C1couples D1 across the tunedcircuit, R1 acts as an RF blocker,and C5 eliminates potentiometernoise. Signal output is taken fromTR1 emitter via capacitor C4.

TUNING-IN Use a short length of wire to

couple the BFO unit to thereceiver. Place it near the set,wrap it around the rod aerial, or

even connect it to it: experimentto find out which arrangementgives the best results.

Tune in a steady signal onthe receiver, set control VR1 tomid-travel, and adjust the ferritecore (slug) of coil L1 until astrong whistle is heard from thereceiver’s speaker, gradually

Ω

. . . .

–

Fig.6. Circuit diagram for an RF oscillator, covering 150kHz to 30MHz, using a bipolar transistor series fed Hartley oscillator.

See Tables 1 and 2 for suitable coils.

carrier wave (Morse) andamateur single-sideband (SSB)transmissions audible on adomestic shortwave radio. Thecircuit diagram of Fig.4 showshow a shunt fed Hartleyoscillator can be used in thisway.

Inductor coil L1 is aminiature 450kHz to 470kHz IFtransformer, coupled to thebase of transistor TR1 bycapacitor C2. Feedback is takenfrom TR1 emitter (e), via DCblocking capacitor C3, to atapping on the tuned winding.Resistors R2 and R3 bias thetransistor, and C6 bypasses thepower supply.

The unit functions byinjecting an RF signal into thereceiver at its IF frequency. This

. . .

. . .

–

Fig.5a. Circuit diagram for search head and

BFO for a simple metal detector.

Continuous Tuning

Circuit diagrams Fig.6 andFig.7 depict bipolar and fieldeffect transistor versions ofHartley oscillators, which can becontinuously tuned over a widefrequency range; i.e., from150kHz to 30MHz. Coil switch-ing arrangements are simple,and the circuits can be used asthe local oscillators in radio re-ceivers, or as the basis of a sig-nal generator. With appropriateinductor and capacitor combina-tions, the circuits will oscillatedown to 100Hz.

7/23/2019 EPE_07-1999



fading to a slow fluttering aszero beat is achieved. Tune thereceiver to one of the amateurbands (7MHz or 14MHz).

If the band is active

(sometimes they are completelydead), careful adjustment ofVR1, which acts as a fine tuningcontrol, should clarify theamateur SSB transmissions. Tryconnecting an extra 3m or 4mof wire to the set’s rod aerial toimprove reception. Althoughinexpensive and easy toconstruct, the unit will enable adomestic portable to produceresults comparable to thoseachieved by some of the

simpler amateur-bands receiverdesigns.

Almost any 450kHz to470kHz transistor radio IFtransformer will work in thiscircuit. If a salvaged item isavailable, use an ohmmeter todetermine the lower resistanceportion of the tapped tunedwinding, and connect this to thenegative rail.

METAL DETECTOR Simple metal detectors

work on the beat frequencyprinciple. An oscillator, tuned bya search coil, is kept close tozero beat with a secondoscillator located within theinstrument.

When the search coil isbrought near to metal itsinductance changes and thefrequency of oscillation shifts(lower for ferrous metals, higher

for non-ferrous). This results inan audible change in the beatnote. The extent of the changedepends on the proximity andsize of the metal object. Fig.5shows how a shunt fed Hartleycircuit can be used for both thesearch head and the beatfrequency oscillator.

Under current UKlegislation, the maximumoperating frequency ofequipment of this kind is

148

5kHz. In order to tune to

this frequency, all of thewindings of a 450kHz to 470kHzIF transformer, L1, areconnected in series to increasethe inductance, and additionalcapacitors C1 and VC1, arewired across it. These measureslower the resonant frequency toaround 120kHz.

Capacitors C2 and C3couple transistor TR1 to thetuned circuit. Resistors R1 andR2 set bias levels, capacitor C5

and resistor R3 de-couple thecircuit from the power supply,and the RF output is taken fromTR1 emitter via C4.

The frequency of the beatoscillator has to be kept within akilohertz or so of the frequencyof the search oscillator in orderto produce a clearly audibletone. The actual terrain overwhich the search coil is moved(wood floors, dry concrete, wetgrass, wet sand, etc.) affects

the frequency of the searchoscillator. To accommodatethis, variable capacitor VC1 isbrought out as a panel controlso that the beat oscillator canbe set close to it under allconditions of use.

ALMOST IDENTICALThe circuits of the two

oscillators are virtually identical.The only differences being the

omission of variable capacitorVC1 from the search coil circuit,and its larger, hand-woundinductor, which is detailed inFig.5b.

How signals from the twooscillators are combined, in asimple product detector, toobtain an audible tone is shownin Fig.5c. Capacitor C4 should

have the lowest possible value,consistent with an adequateoutput, so as to inhibit locking.Any small audio amplifier IC willbe suitable: the TBA820M, the

TDA7052, or, if headphonelistening is acceptable, just acouple of transistors.

Although the arrangement isextremely simple it does workquite well. Indeed, a muchgreater degree of complexity isrequired before any realimprovement in sensitivity isobtained.

CONTINUOUSLY

VARIABLE TUNING A bipolar transistor version

of a series fed Hartley oscillatoris given in Fig.6. Switchedinductors, L1 to L5, tuned by a10pF to 365pF air-spacedvariable capacitor, VC1, coverthe frequency ranges quoted inTable 1 and Table 2.

Maintaining oscillation overthe full swing of a 365pF tuningcapacitor becomes more

difficult as frequency risesthrough the HF region. Also,biasing becomes more criticaland more specific to individualtransistors. Accordingly, presetpotentiometers VR1 and VR2are used so that the voltagescan be optimized.

Capacitor C1 couples thetuned circuit to the base oftransistor TR1. Feedback is viaemitter preset resistor VR2 andits bypass capacitor C2.

Resistor R2 and C4 de-couplethe unit from the power supply.The RF output signal is takenfrom TR1 emitter, via C3. Thiscapacitor should have thelowest possible value in order tominimize loading on theoscillator.

Almost any npn small-signaltransistor will work in this circuit

7/23/2019 EPE_07-1999



provided its fT is high enough:RF types are not necessary. Inthis connection, it is worthnoting that most transistors willoscillate close up to their fT if

the associated tuned circuit hasa reasonable Q factor.

Set preset potentiometersVR1 and VR2 to half travel andget the circuit working onmedium waves beforeproceeding to the HF ranges.Adjustment of the presets is notparticularly critical, and it shouldbe possible to set them so thatthe circuit operates reliably from150kHz to above 30MHz, andover the full tuning capacitor

swing on the highest range. Ifdesired, the resistance in-circuitcan be measured after settingup and the potentiometersreplaced by fixed resistors ofthe nearest standard value.

MULTI-BAND USING A FET

An alternative RF oscillatorcircuit using a field effecttransistor (FET) is given in

Fig.7. Biasing arrangements aremuch simpler, but diode D1 hasto be provided to limit oscillationamplitude and prevent forwardconduction of the JFET’s gate.

Limiting the amplitude alsoimproves waveform and makesthe output more constant overthe tuning capacitor (VC1)swing and across the switchedranges (S1). It can also result insome small reduction infrequency drift. Connecting two

or three diodes in series fixesdifferent thresholds for thelimiting, and the oscillatoroutput increases by roughly

0 2V per additional diode in the

chain.

Ω

. . . .

–

Fig.7. Hartley circuit using a field-effect transistor (FET) to produce a RF oscillator, covering 150kHz to 30MHz. See Tables 1 and 2 for suitable coils.

Range No. of Turns S.W.G.

Frequencyat max 'cap'

(MHz)

Frequencyat min 'cap'

(MHz)

1 36 0.15 0.54400 in 4 piles of

100 turns

2 28 0.47 1.80120 in 4 piles of

30 turns

3 28 1.13 4.8570 turns,

close wound

4 24 2.10 9.2030 turns,

close would

5 24 7.00 30.008 turns, spaced

Table 1. Air-cored coils, hand-wound on 20mm OD formers (Frequency coverage with 10pF to 365pF tuning capacitor).

RangeToko coiltype No.

Baseconnection

Frequencyat max 'cap'

(MHz)

Frequencyat min 'cap'

(MHz)

1 B 0.15 0.54CAN1A350EK

2 B 0.45 1.80RWR331208N2

3 A 1.40 6.00154FN8A6438EK

4 A 4.00 17.00154FN8A6439EK

5 A 7.00 30.00KXNK3767EK

Table 2. Toko, screened, ferrite-cored coils (Frequency coverage with 10pF to 365pF variable capacitor).

See Fig.8. for alternative base connection details. Range 1and Range 2 coil windings series connected to increase

tapping ratio.

7/23/2019 EPE_07-1999



TUNING COILS Details of the frequency

coverage obtained with a 10pFto 365pF variable capacitor andhand-wound coils are given in

Table 1. The coverage withinductors manufactured by Tokois set out in Table 2.

Short lengths of 20mmdiameter plastic conduit areused as formers for the hand-wound coils, and the higherinductances have to be pilewound in order to reduce self-capacitance. The piles are best

wound between the cheeks of4mm wide card bobbins afterthey have been slid onto theformer and spaced by 3mm orso.

Glue the bobbin cheeks andsleeves with Durofix or a similarquick-setting adhesive, and cutslots in the cheeks for the entryand exit of the wires. A varietyof tappings and coupling

windings can be provided oncoils produced in this way, andthis greatly facilitatesexperiment with the variouscircuits.

MAKING A POINT The feedback tapping point

for the Hartley oscillator canrange from 10 to 50 percent ofthe total turns, measured upfrom the “earthy” or “cold” endof the coil. Increasing thenumber of turns increases thefeedback and the RF outputvoltage available at the emitter(e) or source (s) of thetransistor. Up to 20 to 25percent is the conventionaltapping point, and it should beadopted, at least for initial trials,with the multi-range circuits.

Not all constructors havethe time or inclination to hand-wind coils, and Table 2 givesdetails of suitable componentsfrom the Toko range. Differentconnections can be made to thewindings and tapping point inorder to optimize results, andthese are detailed in Fig.8.Adjustable ferrite cores, withinthe formers, permit inductancevariation, and the frequencyranges quoted in the table canbe shifted over fairly wide limits.

When several of the hand-wound inductors are mountedclose to one another (within100mm or so), provision must be made to short out the onesnot in use. If this is not done,out-of-circuit coils, brought toresonance by stray and theirown self-capacitance, canabsorb energy from theoscillating circuit and causedead spots. The problem isinvariably caused by the coil ofhigher inductance and next inrange to the one in circuit. Thisproblem does not arise withToko coils, as they usually havea very effective screening “can”.

Experimenters who wish touse the circuits as local

COMPONENTS

Resistors can be 0

25W, 5%

tolerance types. Fixed capaci-

tors used in tuned circuits musthave polystyrene, mica, or, forthe larger values, Mylar film di-electrics. Other fixed capacitorscan be ceramic types, or elec-trolytics for the high values.

Toko coils and IF transform-ers can be obtained from:

o) Bonex, Ltd., 12, ElderWay, Langley BusinessPark, Slough, Berkshire,SL3 6EP, UK.Tel: +44 (0) 753-549502

o) Cirkit Distribution, Ltd.,Park Lane, Broxbourne,Herts, EN10 7NQ, UK.Tel: +44 (0) 1992-441306

o) JAB Electronic Compo-nents, PO Box 5774, GreatBarr, Birmingham, B448PJ, UK.Tel: +44 (0) 121-682-7045

The polythene dielectric“Varicon” tuning capacitor forthe Metal Detector circuit is

available from most componentsuppliers, as are the Jackson365pF air-spaced tuning capaci-tors.

See Shoptalk for furtherdetails.

. .

. .

Fig.8. Toko coil connection details for wide range Hartley oscillators.

7/23/2019 EPE_07-1999



oscillators in general coveragereceivers will, of course, have toswitch appropriate oscillatorcoils and padder capacitors intocircuit to ensure correct trackingwith the set’s RF stages. Fulldetails of these arrangements,for 450kHz and 1600kHz IFs,will be given in a later article.

TUNING CAPACITORS The tabulated frequency

coverage is based on the use of

a Jackson 10pF to 365pF air-spaced variable capacitor, asthese are retailed by a numberof suppliers. Other capacitorscan, of course, be used, anddifferent swings will affect thecoverage achieved.

Polyvaricons, or polythenedielectric variable capacitors,can be substituted, but there willbe some deterioration infrequency stability andperformance. (The measured Q

of a tuned circuit incorporating apolyvaricon is approximately 10per cent lower than one with anair-spaced component.)

LAMPKIN OSCILLATOR Bipolar transistors have a

low base impedance, and abetter match can be obtained,with the tuned circuit, if the baseconnection is tapped down theinductor coil. When the Hartleycircuit is modified in this way itis known as a Lampkinoscillator. Coil switching is morecomplex but, in moredemanding situations, themodification is worthwhile.

The optimum tapping pointfor the base is best determinedby trial and error, but it shouldbe as close as possible to the“earthy” (ground) end of the coilconsistent with reliableoperation.

The windings of Toko coilscan be arranged to give a baseor gate tapping point, and thenecessary connections are

given in Fig.8c. Although theturns ratios are far fromoptimum for bipolar transistors,it is still an improvement worthmaking if the complication of anextra switch bank can betolerated.

Tapping the gate terminaldown the coil will also benefit

FET maintained circuits,especially at higher operatingfrequencies, where thereduction in drift should benoticeable. Tapping the base orgate down the coil reducesdamping on the tuned circuitand enhances its operational Q .

OUTPUT LOADING Extracting the signal from

the oscillators described so farloads the tuned circuit andreduces its Q . Too muchloading can inhibit oscillation,especially at extreme settings ofthe tuning capacitor.It alsobrings with it otherdisadvantages associated with areduced Q factor, namelyuncertain starting, erraticoperation and increased drift.This is why the value of theoutput coupling capacitor has tobe as low as possible and theimpedance of the acceptingcircuit kept high.

Extracting energy via anisolating buffer amplifierovercomes these problems.There is also a Hartley variant

Ω

. . . .

Ω

Ω

–

Fig.9. Dow varient of Hartley circuit using a dual-gate MOSFET to produce an RF oscillator covering 150kHz

to 30MHz. See Tables 1 and 2 for suitable coils.

Lampkin and

Dow Variants

Drift becomes a problem asthe operating frequency in-

creases through the HF spec-trum. How the base or gate of atransistor can be tapped downthe coil to form the drift-reducing Lampkin variant isshown in Fig.8c.

How a dual-gate MOSFETcan be used to simulate Dow’smodification, in which the signaltake-off point is isolated fromthe oscillator, is shown in Fig.9.Signal output is not greatly af-fected by these measures, but

drift is reduced and the wave-form can be improved.

7/23/2019 EPE_07-1999



that provides a measure ofisolation without recourse to abuffer stage.

DOW OSCILLATOR Known as electron coupling,J. B. Dow’s circuit modificationcan be applied to the Hartleyoscillator. Published in 1931,the development involvesextracting the RF voltage fromthe anode of a pentode ortetrode valve, when theoscillating circuit is formed bythe cathode, control grid andscreen grid.

In this way, the output port

(the anode) is isolated from theoscillating circuit, the onlycoupling being via the flow ofelectrons through the valve,hence the name. Thisarrangement eliminates theloading on the tuned circuit

caused by the extraction ofenergy at the feedback tapping,or via a coupling coil.

A dual-gate MOSFET canbe used to simulate this type of

oscillator, and the circuit isgiven in Fig.9. The potentialdivider network formed byresistors R2 and R3 holds gateg2 at half the supply voltage,and the RF output is developedacross drain load resistor, R4.Resistor R6 and capacitor C4decouple the circuit from thepower supply, and C2 groundsgate g2 at RF.

The value of theunbypassed source resistor R5

should be increased if theoscillator behaves erratically.This usually manifests itself asa tendency to frequency double,especially on the higher rangesand at low tuning capacitorsettings. The value quoted

should prove satisfactory withthe Toko coils listed in Table 2.

Whilst the degree ofisolation afforded by thestructure of a solid-state device

is not likely to equal thatprovided by an evacuatedvalve, there is less drift at highfrequencies with thisarrangement. Combining Dow’soutput isolation with Lampkin’sreduced tuned circuit loadingshould produce furtherimprovements in this area.

FREQUENCY DRIFT Drift in LC oscillators

becomes an increasing problemabove 5MHz, and a high degreeof frequency stability is almostimpossible to achieve, withoutrecourse to complex circuitry,above 10MHz. The benefitsafforded by the Dow andLampkin variants will not befully realized unless care istaken with the choice ofcomponents and the actualconstruction of the oscillator.

Drift reducing measures of

this kind will be described innext month’s issue, which willalso cover the construction of abuffer amplifier, and the Colpittsoscillator and its variants.

http://0799n.pdf/

7/23/2019 EPE_07-1999



Robert Penfold looks at the Techniques of Actually Doing it!

Building component panels,whether using a custom printedcircuit board (PCB) or someform of proprietary board suchas stripboard, is not generallytoo difficult for the beginner.You have to be careful to getthe semiconductors andelectrolytic capacitors the rightway round, and with stripboardyou also have the take due careto get the components placed inthe right sets of holes.

Accidental short circuits onthe undersides of boards canalso be a problem, but providedyou follow the published plansprecisely and thoroughly checkthe finished board against thecircuit and wiring diagramsthere will probably be noproblems.

MAKING CONNECTIONS

The wiring to sockets,

controls, etc. can also beperfectly straightforward, butthere is probably more scope forconfusion to creep in with thisaspect of construction. Makingthe connections is quite easy,

and it is just a matter of hookingthe end of the leadout wirethrough and around the hole inthe tag, applying the bit of thesoldering iron, and then feedingin some solder.

As always, make sure youapply the bit to the joint first andthe solder second. This heatsup the tag and wire before thesolder is applied, which helpsthe solder to flow properly overthe joint. Many modern

components have pins ratherthan tags, and it then becomesa matter of coiling a turn of thewire around the pin and thencompleting the joint.

It is a good idea to “tin” tagsand pins with solder prior tomaking connections. It is also agood idea to give the ends ofthe leads the same treatment,but using a thin coating ofsolder. If you use a largeamount of solder it will be

difficult to form the wire into ahook shape or to coil it.

With both surfaces “tinned”with solder there is virtually nochance of producing a “dry”

joint. If one of the surfaces will

not take a coating of solder it iscontaminated with dirt orcorrosion. Gentle scraping withthe blade of a penknife or usinga small f ile should clean thingsup and permit the surface to beproperly “tinned”.

We have been assuminghere that the controls andsockets will be mounted off-board and hard-wired to thecircuit board. Some projects doactually have these components

mounted on the circuit board,and this slightly simplifiesconstruction. However, with thistype of thing you have to makequite sure that you buy the rightcomponents or you may wellfind that they do not fit onto theprinted circuit board.

DOLLY MIXTURES

Sockets and switches areboth potential areas of

confusion, mainly due to thelarge numbers of different typesthat are available. Sockets werecovered in a Techniques articlenot that long ago, so we willconcentrate on switches here.

Fig.1. Contact arrangements used for the four normal types of toggle and slider switches.

7/23/2019 EPE_07-1999



Where a relatively simpleswitch is required the usualchoice is either a toggle orslider type. A toggle switch isoperated via a lever (known as

a “dolly”), and as its namesuggests, a slider switch has asliding control knob.

The full-size toggle switchesare little used in modernprojects due to their large size,and it is the miniature and evensmaller sub-miniature types thatyou are more likely to use.Slider switches are the lesspopular option as they oftenhave rather awkward mountingrequirements, and in my many

years of experience they do notalways operate reliably.

Both types are available infour common forms, and thesimplest is the s.p.s.t. (single-pole single-throw) variety. Thishas just two tags and is a simpleon/off switch. The d.p.s.t. -(double-pole single-throw)switches are basically just twoon/off switches in a single caseand operated in unison.

A s.p.d.t. (single-poledouble-throw) switch has threetags, and is more commonlycalled a changeover switchthese days. The middle tag (the“pole”) connects to either one orother of the other two,depending on the setting of theswitch. A switch of this typecould, for example, be used toselect either the square orsinewave output of a signalgenerator and connect it to theoutput socket.

A d.p.d.t. (double-poledouble-throw) switch iseffectively just two s.p.d.t.switches in a single case andoperating in unison. Fig.1 showsthe tag arrangements normallyused for all four types of switch,together with their circuitsymbols, which should help to

clarify the differences betweenthem.

MIND OF THEIR OWN

Most component cataloguesinclude a few “biased” switches.Probably the most commonexample of a biased switch isthe simple pushbutton variety.These are s.p.s.t. switches, butthe pair of contacts is onlyclosed while you press thebutton. As soon as you releasethe button it springs out and the

switch returns to the off state.There is actually an

alternative type, which isnormally switched on, andswitches to the off state when itis operated. Both types ofswitch are biased to oneposition, and will always take upthat position unless the user -intervenes.

Some toggle switches areavailable in biased versions, butthis type should only be usedwhen it is specifically called forin a components list. Sometoggle switches are alsoavailable in three-positionversions, possibly with a versionthat is biased to the middle -position. In the middle position,the center tag does not connectto either of the other two. Thisgives two modes of operation

plus a sort of central “off”position where neither mode is

selected. Again, switches of thistype are only needed for a fewspecialized applications, andshould only be used whenspecifically requested in acomponents list.

When using toggleswitches, do not make theclassic mistake of getting thetwo positions of the switchconfused. With an on/offswitch there is little risk of thishappening, but with a switch

that is used to control someother function there is agreater risk of confusion.

Many years ago, a numberof readers had problems with atransistor tester design of mine.The testers actually worked fine,but an ambiguous drawingresulted in some readers gettingthe pnp and npn modesmuddled up.

The upper drawing in Fig.2

shows the on and off positionsfor single-throw toggle switches.With double-throw switches themiddle and lower tags areconnected together when thedolly is in the “up” position, as inthe lower drawing of Fig.2.

Note that slider switcheshave simpler mechanisms thatoperate the other way round.

Fig.2. Try not to make the classic mistake of confusing switch settings.

7/23/2019 EPE_07-1999



The middle and upper tags areconnected together when thecontrol knob is in the “up”position.

ROTARY SWITCHES

Rotary on/off switches aremainly intended for use withmains powered equipment, andthey are d.p.s.t. types that canswitch both sides of the mainssupply. They can be and aresometimes used in batterypowered equipment, where boththe positive and negativebattery leads could becontrolled by the switch.

However, this is generallyconsidered to be a bit “over thetop”. One pole of the switch isleft unused and the othernormally controls the positivebattery lead.

The tag arrangement ofrotary on/off switches gives littleclue to the correct method ofconnection, and the same isalso true of the switches fitted to“switched” potentiometers. Fig.3shows the correct method of

connection for both single anddouble-pole operation. Forsingle pole operation the upperpair of tags are used in Fig.3,but the correct switching actionwill obviously be obtained if thelower set are used instead.

MULTI-WAY SWITCHES

The switches described sofar only provide two-way

operation, or a sort of 2

5-way

operation in the case ofswitches with a central “off”position. Some applicationsrequire switches having three ormore positions, and there aretwo general approaches to thistype of switching.

One is to have a bank oflinked push button switches,and banks of pushbuttonswitches are common in ready-made equipment. This methodis not popular amongst home

constructors though, as it tendsto be extremely expensive toimplement.

Unless some really complexswitching is required, a standardrotary switch is the normalchoice. These are available in12-way 1-pole, 6-way 2-pole, 4-way 3-pole, and 3-way 4-poleversions, and all four types lookbasically the same.

Getting this type of switchconnected correctly can be a bit

tricky and you have to proceedwith due care. Modern rotaryswitches are marked with lettersand numbers that aid thecorrect identification of the tags.The pole tags are marked withletters from A to D, and the

other tags are numbered from 1to 12, as in the two examples ofFig.4.

With the 6-way switch, poletags A and B respectively

connect to tags 1 and 7 with theswitch at position one (set fullycounter-clockwise). Moving theswitch to position two connectstags A and B to tags 2 and 8respectively, then tags 3 and 9at position 3, and so on.

With the aid of a wiringdiagram and the markings onthe switch itself it should not betoo difficult to get everythingconnected correctly. In practicematters are made more

awkward by the fact that thesmall letters and numbersmolded into the case of theswitch are difficult to read onceit is mounted in the case.

It is often helpful to markthe position one tags (e.g. tags1, 4, 7, and 10 of a four-poleswitch) with a small blob ofpaint on the body of the switch.This makes it easy to navigateyour way around the switch andhelps to avoid errors.

Modern rotary switcheshave adjustable end-stops sothat they can be used with lessthan the maximum number ofsettings. Also, in many projectsone or more poles of a rotaryswitch are left unused. If (say) a

Fig.4. The tags of rotary switches are normally labeled, as in these two examples.

Rotary switches are normally equipped with an adjustable end-stop.

7/23/2019 EPE_07-1999



5-way 1-pole switch is required,you would actually use a 6-way2-pole type with the end-stopset for 5-way operation and onepole left unused.

If you remove the fixing nutand washer from one of theseswitches the metal end-stop canbe dislodged using the blade ofa small screwdriver or apenknife. With the end-stoprelocated in the appropriate slotthe switch is ready to bemounted on the project’s frontpanel.

MAKE OR BREAK

Multi-way rotary switchesare normally offered in twoversions that are called “break-before-make” and “make-before-break” switches. Thedifference between the tworevolves around what happensas the switch is moved from one

position to the next.

With a break-before-maketype, the pole is disconnectedfrom one tag before it isconnected to the next. This

leaves the pole tag momentarilyconnected to nothing. With amake-before-break switch thepole is still connected to one tagwhen it makes contact with thenext. This means that the twonon-pole tags are brieflyconnected together as theswitch is adjusted from oneposition to the next.

If a components list does notspecify one type or the otherthere should be no problem using

either kind of switch. If a make-before-break switch is specifiedand you use the other type it isunlikely that there will be any direconsequences, but the projectmay glitch in some way each timethe switch is operated. Forinstance, there may be a loud

“click” from a loudspeaker, or thepointer of a meter might “twang”across the scale and hit the end-stop.

Using a make-before-break

switch instead of a specifiedbreak-before-make type almostguarantees dire consequences.Each time it is operated therewill be a momentary shortcircuit on the supply lines, twooutputs will be briefly connectedtogether, or something of thisnature. Apart from greatlyshortening the life of the switchthis could also result inexpensive damage to othercomponents in the project.

http://0799o.pdf/

7/23/2019 EPE_07-1999



IT’S A HOAX!

In previous months Ihighlighted the differencebetween computer virii , TrojanHorses, and worms. Thesepotentially damaging files canbe transmitted via the Internet,and unsuspecting Internet userscan find themselves on thereceiving end of an unprovokedattack from some malicious orirresponsible elements.

Before I develop this topicany further this month, here’s ariddle. What do the threeBudweiser frogs and YellowTeletubbies have in common?They are actually examples of afourth class of virus, namely thehoax virus. These cause a greatdeal of unnecessary concernamongst inexperienced users.Typically, a hoax virus arrives inthe form of an importunateemail message warning of

possible dangers that mightarise, for example, if acompletely unrelated computer

file say a screensaver is

opened. Panic spreads and thewarning is flashed to otherregular contacts. Thus the virusescalates.

These virii tend to comeand go. Whilst some types ofhoaxes are obviously harmlessfun, there is no doubt that othertypes really do have a worrying

effect on their recipients. Thehoax panics people intospreading the word about apossible “new” virus which isalleged to be doing the rounds.It is the actual propagating ofthese well-meaning messagesby the inexperienced which is

the “damage” caused by thehoax virus; I have recentlyreceived several fromconcerned and well-intentionedreaders.

In view of the current TVadvertising campaign in the UKby Budweiser beers (featuringthe three frogs), it seems timelyto remind readers of the hoaxbudsaver.exe virus. This hoax(also known as the Bud Frogs)relates to a mythicalscreensaver calledbudsaver.exe, which is said tocrash a hard drive. It is acomplete sham, and there is noneed to be worried by any suchmessage. (I suppose thissituation is fine until someonereally does release a virusunder that name.)

A similar hoax calledBuddylst.zip (or Budweiser)has been on the go for some

time using several forms ofmessage which can equally beignored, as can one relating toYellow Teletubbies (!). There isa whole raft of hoaxesdescribed on the Symantec website (www.symantec.com/avcenter), and readers shouldmake a habit of checking itoccasionally to stay abreast ofboth hoax and malevolent virii.

FALL OUT

On the other hand, a highlymalicious virus which hasrecently been causing muchconcern is called W95.CIH, alsoknown as Chernobyl, NetworkNuke, and other names. Ittriggers on the 26th of a month(the anniversary of Chernobyl)

and infects Windowsexecutables. Worse, it may alsotry to modify certain Flash BIOSchips, which can cripple somesystems. Symantec has a freetool available from their website called kill_cih, whichenables users to check for thepresence of Chernobyl. It runsfrom the DOS prompt, andalthough it will not cleanse asystem, says Symantec, it will -disable the virus and alert the

user.

Don’t overlook the fact thatapart from introducing virii ontoyour system via email or FTP’dfiles, there are other ways ofinflicting damage, perhaps of atemporary and easily-recoverable nature. Some websites are deliberately set up inorder to demonstrate a possibly

harmful feature

software bug

e.g. a

or to make a

particular point. Hopefully youwill be warned of any possibledangers in advance. Theseevents are sometimes calledexploits .

If you have a Pentium IIIprocessor, a good example ofan exploit that will test yourcourage will be found atwww.zeroknowledge.com/p3/.This informative web sitepurports to blow the whistle onthe serial number embedded in

P3 processors. Even with theserial number disabled, thenumber can still be accessed,says the web page. It clearlywarns that after running thedemo, your system may lock upafterwards and it also says thatany anti-virus warning that yoursoftware may generatebeforehand, should be ignored.

By Alan Winstanley

SURFING THE INTERNET

http://www.symantec.com/avcenter






http://www.zeroknowledge.com/p3/






7/23/2019 EPE_07-1999



This latter warning isunderstandably enough to detersome hardened computer users,but you can access the site andsee for yourself!

STUCK ON THE ON-

RAMP

Having spent many weeksagonizing about upgrading toBT Home Highway to obtainISDN access, I have decided toshelve the idea, at least fornow. I found there are severalpotential “gotchas” that needthinking about. One is that BTwon’t guarantee the speed of

access using ISDN2e anyway,which might defeat the point ofinstalling it to begin with.

Problems with modemspeed may also arise in caseyou find that you havepreviously been “DACSed” byBT, which involves using aconverter to multiplex twophone lines down a singletwisted pair copper wire. Itsaves BT the cost of having toinstall extra wires. Using

ordinary dial-up modem access,some users who have been“DACSed” complained that theycould barely scrape much morethan 30Kbps, even though theywere supposed to be paying fora second “clean” line. Aftersome persistent complainingfrom users, BT have in somecases relented and installed aseparate wire.

Another problem was that,due to my current DACS

connection and the way in whichtelephone numbers are used in“blocks” by BT, they advisedthat a change of number willprobably be required, even

though BT’s literature statesthat for the majority of users thiswon’t be the case.

BT Home Highway doesoffer several advantagesthough. A primary one is thatInternet access is almost

instantaneous a second or two

rather than 45-60 seconds. Formyself, this would probably bethe biggest boon. Faster accesstimes may or may not beenjoyed by Highway users, and

by “bonding” two channelstogether, they could experiencea theoretical speed of up to128K. (Users of earlier Windows95 versions require laterversions of Microsoft Dial-Up

Networking (DUN) for this

check the Microsoft web site.)

What readers may not beaware of is that, in usingchannel bonding, two calls arebeing made at twice the cost,and two ISP connections areparalleled at the same time. Notevery ISP permits this with anordinary dial-up account, sotheir terms of service should beconsulted.

Other benefits include thefact that one Home Highwayline will offer two analog andtwo digital ports. You can useany two ports at the same time,i.e. an analog phone or faxmachine alongside either aterminal adapter (an ISDN

external “modem”) or anotherphone, or two PCs each on 64K,or two digital ports bonded -together to produce atheoretical 128K on one PC.

The economics arereasonably simple to calculate:the cost of calls is the same asan ordinary analog BT line, anddiscount number schemes stillapply. Presently it costs 40 UKPounds per month excludinginstallation, but includes a 15Pound call allowance. You maybe able to recover more of thecost by making an existing BTline redundant, as HH gives youtwo lines. A BT survey is

necessary before ISDN can beinstalled and more advice fromBT is available onwww.homehighway.bt.com. Inthe meantime we can onlydream of the advent of highbandwidth ADSL (asymmetricdigital subscriber line) and9Mbps download speed, whichis likely to be several yearsaway for most UK users.

LINKS

As always, I welcomereaders’ suggestions forinteresting electronics-relatedaddresses. My email address [email protected].

http://www.homehighway.bt.com/




http://0799p.pdf/


7/23/2019 EPE_07-1999



British TV viewers will soonbe spoilt for choice if they want togo on-line without struggling touse a PC. Although Internet TVhas been tried in the US,consumer uptake has been lessthan expected. Prompted by aunique marketing strategy fromRupert Murdoch, the major

service providers are now liningup for a showdown in the UK.

WebTV set-top boxes, whichconnect between a phone lineand TV to display Internetimages, first went on sale in theUS in 1996. A year later Microsoftbought the company, and re-launched as WebTV Plus. Boxesfrom Philips, Sony, andMitsubishi cost $200 and havebuilt-in Browser software.Because WebTV gives open

access to the Internet, and Websites are intended for display on aPC monitor, proprietary circuitrytries to modify the image so thatit looks acceptable on a lowerresolution TV screen.

Microsoft predicted that itwould have one mil lionsubscribers by the end of 1998,but achieved less than 700,000.Now the company has tied upwith the BBC, NatWest bank, andthe Carlton and Granada TV

groups to experiment with 100homes in London and Liverpool.The BBC adds flag signals tosome digital TV programs, whichput icons on the screen of areceiver made by Pace. Theviewer uses a remote control toclick on the icon. The box thengoes on line and searches theInternet for extra information.

NatWest says the system couldbe used for “buying a house and going on holiday’’ .

NatWest’s rival Midland is a

partner in the Open service,which British InteractiveBroadcasting are due to trial inreadiness for a full-scale launch

A roundup of the latest Everyday News

from the world of electronics

INTERNET TV SHOWDOWNGetting access to the Internet is becoming increasingly easy.

Barry Fox reports.

CLAMP IT !There are some things that even solder cannot secure in the

electronics workshop from time to time, adhesives of various typeshave to be used. So what better way to ensure the surfaces stay to-gether while the adhesive hardens than to use a mini-vice, such asthe Quick-Grip & Spreader from Minicraft.

The new clamp is a top quality tool made from the f inest materi-als. It incorporates a pumping action that adjusts the clamping areaand can also be used in reverse to spread pressure so you havecomplete control at all times. The bar which the pumping action runs

on is made from hardened chrome steel, and the detachable padsare made from a special soft material that will not scratch or markdelicate surfaces.

The pads measure 112mm and the spreader capacity is around207mm. At a price of only 9.99 UK Pounds, the clamp seems idealfor anyone in the model making, hobby and electronics fields.

For more information, contact Minicraft, Dept. EPE, 1 & 2 Enter-prise City, Meadowfield Avenue, Spennymoor, Co. Durham DL166JF, UK. Tel: +44 (0) 1388-420535. Fax: +44 (0) 1388-817182.

7/23/2019 EPE_07-1999



in the autumn. The otherpartners in BIB are RupertMurdoch’s Sky, BT andMatsushita (Panasonic). Theygive a subsidy of over 150 UK

Pounds to anyone who buys aDigibox and dish to receiveSky’s digital satellite service,and agrees to connect to aphone line for at least a year.

The Digibox contains a28.8k modem and all thecircuitry needed to access anonline service. So when BIBlaunches Open service it willalready have a captive market.

Walled Garden

Until recently BIB plannedto provide Open users with a“walled garden” of vetted safesites, like the CampusWorldservice which BT provides forPC users in educationalestablishments. BIB alsointended to reformat the vettedsites, enlarging text fonts, forclearer display on TV.

BIB has now taken thesurprise decision to bar all

access to the Internet. ExplainsBIB’s Chief Engineer AlecLivingstone “There is just too much work involved in monitoring and reformatting thousands of sites” . BIB’smarket research also showedthat talk of Web Browsers andthe Internet scares offcustomers who just want towatch TV.

BIB will work with anexisting Internet provider

(believed to be BT) which willgive each subscriber an emailaddress. Incoming messageswill be flagged on the TVscreen. The viewer can thenreply by using an infraredkeyboard or a remote controlwhich sends pre-packagedreplies, like “Yes” or ”No”.

Two transponders on the

Astra satellite continually deliver a68 Megabit/second stream of dataand video, such as a multimediashopping catalog, direct to thehome dish. The receiver dials out

when it needs to exchange data,such as encrypted credit carddetails. Woolworths has alreadysigned up to provide a service.

BIB says it can re-jig Open foruse on other TV services, such asSky’s terrestrial rival OnDigital.

NTL is offering a third option,after buying the commerciallyunsuccessful NetChannel servicelast year and re-vamping it as TV-Internet. After trials with NTL’s staffand friends, the service is now

being advertised in the Newcastlearea. TV-Internet offers full emailand Internet access, except forbarred pornography sites.

Hedging bets, Microsoft hasbought a f ive per cent stake inNTL.

ics Manual , Electronics Service Manual , and UCANDO Videos .You can, of course, obtain allthese products direct from EPEin the UK, and the videos con-

tinue to be available from Far-nell Components Ltd.

Note that Maplin SouthAfrica are not affected by thedecision made by the UK HQ ofMaplin Electronics, and theycontinue to supply our products.

If you previously purchasedyour copies of EPE from aMaplin store in the UK, pleasenow place an order with yourlocal newsagent or take out asubscription (you save 44 pencean issue with a subscription

UK prices).

GREENWELDON HOLD

We are sorry to report thatGreenweld Electronics in the UKhave ceased to trade at the presenttime. They have voluntarily ap-pointed a liquidator who is currentlytrying to sell the stock, databaseand trading name.

If you have sent orders re-cently, they will not be fulf illed, butchecks etc. will not be cashed, orcredit cards charged. We may beable to report more informationnext month.

MAPLINWITHDRAWS

Maplin Electronics, in a regret-table rationalization of their productsuppliers, have decided to deletefrom their UK stores the followingproducts supplied by WimbornePublishing: EPE , Modern Electron-

HERBERT HOWARDWe regret to have to record

the passing of Herbert Howard,who died on May 13th after ashort illness, borne with greatcourage and dignity. Many read-ers knew Herbert through theBritish Amateur ElectronicsClub.

Herbert had been an elec-tronics hobbyist for seventyyears, a member of the BAECfor twenty years and its chair-man since 1990. Always willingto help anyone, he championedthe interests of hobbyiststhrough his unstinting work withthe club.

Our sympathies go to hisfamily and many friends, he willbe sadly missed.

BT NETS SEGABritish Telecom has an-

nounced that it has been cho-sen by Sega to provide a pan-European subscription-free ac-cess service for users of SagaDreamcast. The Dreamcastgames console, to be launched

7/23/2019 EPE_07-1999



later this year, will provide userswith Internet access throughtheir TVs on a pay-as-you-gobasis via BT’s network of Euro-pean partners. This is the firsttime that subscription-free Inter-

net access has been madeavailable across Europe in thisway.

Dreamcast will include a33.6k modem and a browser. Inaddition to Internet access, itwill enable users to accessemail, chat and on-line gamingservices. Future plans includethe offer of on-line shopping fa-cilities. It is expected to retail ataround 199 UK Pounds.

For more information, con-

tact your local Sega retailer.

RECYCLEDPRINTER

CARTRIDGESAlan, our Online Editor, has

filed this interesting item:

Whilst many consumers resortto the messy job of refil ling spentinkjet and laser printer cartridges,commercial and educational usersoften discard empties that containperfectly good components, whichthey would gladly recycle if only aconvenient scheme existed. Theempty cartridges merely go to land-fill waste, where some types could

ALAN DOUGLAS1899-1999

It is with sadness that wereceived the news that AlanDouglas has died, just short ofhis 100th birthday. Alan was

strongly associated with theElectronic Organ ConstructorsSociety for many decades, andwas its President over a periodof 19 years.

Older readers may well re-call Alan’s Electronic Organpublished in Practical Electron-

ics in about 1969 or 1970 itwas a very significant design forits era.

A brief biography of Alan’s

life is included in the memorialbooklet produced by the EOCS.For more information, contactTrevor Hawkins, Hon. Sec.EOCS, 23 Blenhein Road, StAlbans, Herts AL1 4NS, UK.

take up to 80,000 years tobiodegrade, it is claimed.

The British Institute forBrain Injured Children (BIBIC)now participates in a collectionand recycling scheme, whichenables many popular brands of

empty inkjet and toner car-tridges to be recycled. TheSomerset-based charity helps inthe rehabilitation of brain-injured children and relies totallyon voluntary donations for itswork. It receives a payment foreach suitable cartridge col-lected. “Last year BIBIC raised enough funds from its ‘toner

FASTER CHARGINGAn innovative new battery charger which reduces the charging

time of 9V block PP3 batteries from 14 hours to just two has beenlaunched by leading European power supply specialists, Friwo.

Using a microcontroller, the Speedy 9V automatically switchesbetween three charging modes to maximize the life of the battery,whilst maintaining high performance and durability. Once the batteryhas received its full charge in fast-charge mode, the chargerswitches to trickle-mode to maintain peak efficiency. The third modeenables deeply discharged batteries to be recharged, switching tofast-charge mode once the battery has been reactivated.

For more information, contact Haredata Standard Products Ltd.,Dept. EPE, Hyde House, Victoria Avenue, Harrogate, N. Yorks HG11DX, UK. Tel: +44 (0) 1423-530347. Fax: +44 (0) 1423-524645.

7/23/2019 EPE_07-1999



donor’ scheme to fully support twelve youngsters for a whole year”, explained Erica Wheelerof BIBIC’s External Affairs.

Every toner or inkjet car-

tridge on their approved list iswelcomed, which includes manyHP, Lexmark, Canon, Appleand Epson types. BIBIC ar-ranges free collection of lasercartridges or provides a Freep-

ost label for inkjets. The system’sviability is only maximized whenten or more laser cartridges be-come available for collection, oth-erwise a Freepost address can be

used to send up to four laser car-tridges per parcel. Up to six inkjetcartridges can also be returned inspecial pouches supplied by BIBIC,but they stress that only approvedtypes of cartridge can be used.

To find out how to supportBIBIC’s work in treating brain-injured children as well as

helping the environment con-tact Erica Wheeler on +44 (0)

1278-684060 or [email protected]. Now youcan recycle those empties witha clear conscience.

http://0799r.pdf/

7/23/2019 EPE_07-1999


EPE Online, June 1999 - www.epemag.com - 722

EPE Mood PICker

Reading between the lines,the conclusion to be drawn re-garding the use of the chokeversion of the inductor for the

EPE Mood PICker is don't!

Apart from being an extremelymessy operation to produce anair-cored coil from it, the cost isabout three times greater thanthe alternative coil.

Constructors are, therefore,recommended to stick with the“Eagle” LT44 driver transformerand modify it as outlined. TheLT44 is sometimes listed by ad-vertisers as a “transistor drivertransformer” and should bewidely available. If any difficul-

ties are experienced in locatingit try a Maplin store, quotingcode HX82D. The same sourcewas identified as stocking theAD8532 dual opamp, codeOA16S.

A ready-programmedPIC16F84 microcontroller isavailable (mail order only) fromthe author for the sum of 8 UK

Pounds inclusive. Overseasreaders should add 1 UK Poundfor postage and packing. Ordersand payments should be madeout to Mr. A. Flind and sent to:Mr. A. Flind, 22 Holway Hill,Taunton, Somerset, TA1 2HB,

UK.For those readers who wish

to program their own PICs, the -software can be downloadedFree from the EPE Online Li- brary at www.epemag.com

12V Battery Tester

Most of the parts requiredfor the 12V Battery Tester arenearly all standard items andmost of our components adver-

tisers should be able to comeup with the goods, or a suitablealternative. The extra flexiblelead is usually sold as meterlead, but multistrand cablecould prove to be just as reli-able.

Some readers may experi-ence difficulty in locating a suit-able case, with probe, and the

REF50Z 5V voltage referencechip. The probe case came

from Maplin (code JX57M), andthe REF50Z voltage referencefrom Electromail (code 283-851).

LED Stroboscope

We do not expect any buy-ing problems to arise when se-lecting components for the LED Stroboscope , this month's low-cost Starter Project. Both thespecified TIP122 and TIP121power Darlington transistorsshould be stocked by the major-ity of component suppliers.

Practically any of the ultra-bright LEDs should operate wellin this circuit, the higher theirefficiency the brighter thepulses of light produced. Thelatest 5mm types appear to givethe highest light levels. It isworth experimenting with sev-eral LEDs to find ones that givea nice evenness of illumination.

Forget the 3mm types, they lackthe light output for this applica-tion!

The style of case is left toindividual choice. Almost anymedium sized plastic or metalhousing can be used, but bearin mind the bulk of the 8-cellbattery pack when making yourfinal choice.

If you intend to mount thestrobe light LEDs remote fromthe main unit, you could try

adapting a probe type box or asmall torch to give a handheldunit. A probe-case will costabout 2 UK Pounds.

Intruder Deterrent

Checking out componentsfor the Intruder Deterrent pro-

ject, we came up against a mi-

with DAVID BARRINGTON

Web: www.farnell.com

Magenta Electronics (UK)Tel: +44 (0) 1283-565435Web:www.magenta2000.co.uk

MicrochipWeb: www.microchip.com

Rapid Electronics (UK)

Tel: +44 (0) 1206-751166RF Solutions (UK)

Tel: +44 (0) 1273-488880Web: www.rfsolution.co.uk

Some Component Suppliers for EPE Online Con- structional Articles

AntexWeb: www.antex.co.uk

EPE Online Store and LibraryWeb: www.epemag.com

Electromail (UK)Tel: +44 (0) 1536-204555

ESR (UK)Tel: +44 (0) 191-2514363

Fax: +44 (0) 191-2522296Email: [email protected]: www.esr.co.uk

Farnell (UK)Tel: +44 (0) 113-263-6311







http://www.farnell.com/



http://www.magenta2000.co.uk/



http://www.microchip.com/




http://www.rfsolutions.co.uk/





http://www.antex.co.uk/






http://www.esr.co.uk/












7/23/2019 EPE_07-1999


EPE Online, June 1999 - www.epemag.com - 723

nor problem in finding a sourcefor the neat “plug-box”, incorpo-rating a mains outlet socket. Wefinally tracked it down as onebeing distributed by BCL Distri-

bution and coded as PS13AS.Having contacted BCL, they putus onto Harrogate ElectronicServices who, they suggested,might be willing to supply read-ers on a mail order basis.

Richard Page of Harrogateinforms us that they are happyto supply the specified PS13AScase for the sum of 5.95 UKPounds inclusive. (Note that the

pin arrangement is the UKthree-pin standard.) Ordersshould be made out to: Harro-gate Electronic Services, 25 Re-gent Parade, Harrogate, N.

Yorks, HG1 5AZ, UK. (Tel +44(0) 1423-564353). Quote codePS13AS.

The BP103B phototransistoralso took some finding and wasfound listed by Rapid Electron-ics, code 58-490. It is most im-portant that only a new Class-Xcapacitor rated for direct con-nection across the mains supplybe used for C4. One can be pur-

chased from Electromail, code115-196 or 210-500.

Practical Oscillator De-

signs Investigating the “practical”

circuits presented in the Practi-cal Oscillator Designs feature,we came across a Toko coilnumber that was alien to us.The one in question was theRHCS45328AC2 i.f. transformerand, referring back to the au-thor, we were informed that it isa replacement for an earlierToko IFT. It is currently listed byBonex Ltd (Tel +44 (0) 1753-

549502), code 355328. As faras we are aware, it is not listedor carried by Cirkit.

http://0799c.pdf/

7/23/2019 EPE_07-1999



WIN A DIGITALMULTIMETER

The DMT-1010 is a 3 1/2 digitpocket-sized LCD multi-meterwhich measures a.c. and d.c.voltage, d.c. current, andresistance. It can also testdiodes and bipolar transistors.

Every month we will give aDMT-1010 Digital Multimeterto the author of the bestReadout letter.

LETTER OF THE MONTH

WHAT WONDERS!

Dear EPE,

First I would like tocongratulate you on yourexcellent magazine. I am 16years of age and have beenenjoying the wonderful world ofelectronics for quite a few yearsnow. There are three main pointsthat I would like to mention.

The first is concerningelectronics and the media. Wheremagazines are concerned, I thinkthey go a long way to help peoplelearn about electronics, EPE

catering for that nicely, but whatabout “electronics through TV”? Ihave never seen a TV programthat is completely aboutelectronics, and given the numberof channels available, there’s noexcuse for the lack of them. Imay be wrong (I hope so), so ifanyone can tell me of any thathave been, or will be shown on

TV, then I would be mostgrateful.

My second point isregarding the cost ofcomponents. If I see a companyselling reasonably pricedcomponents then I will buythem, but there is another placecomponents can be bought

cheaply boot sales. My family

enjoys going to them and I havefound many electroniccomponent bargains. I knowthat it’s true you run the risk ofbuying damaged/faulty goods,but that’s a risk I am prepared totake.

My final point is regardingwhat could be called“Electronics vs Software”. A lotof companies and individualsnow make all sorts of things thatare software driven. I love

making all kinds of circuits andrepairing things which don’twork, and wonder which will

become the most favored a

circuit that is (usually) reducedin size plus software, or a circuitthat is (usually) larger in sizewithout software. The formerseems to be proving promisingand gaining acceptance (PICs

are helping a lot

amateur).

especially the

Either way, there will always

be something to repair(although malfunctions arebecoming rarer). This maymean that the electronicsengineer may not only berequired to have a knowledge ofelectronics circuitry, but alsosoftware.

Matthew StuartRomford, Essex, UK

It’s good to hear from you Matthew, and I am delighted to choose your letter as this month’s leading letter. I trust the meter we are sending you will last for years and not need your skill at repairs!

On your first point, the only fully electronics programs that I have ever seen on TV are those broadcast by the Open University on BBC2, after midnight. Extremely well- presented they have been too.They are not too advanced, and basically take the form of early introductions to different aspects of the subject (at least,those that I’ve seen have had this nature).

Details of contacting the OU are often given at the end of

their programs which are also

broadcast on some mornings (although I’ve never watched them at that time). If you find out when such broadcasts are going to take place (as the OU should be able to tell you), you could video them.

Regarding boot sales, I’ve not used them, but in my early days of electronics I used to be able to buy old TVs and radios from street markets; they were an excellent source of

components for a young (nearly pennyless!) experimenter.

Electronics vs software?

There’s surely no competition

it’s not that cut and dried. The whole scene is too intermingled and can’t really be fully separated. Software is not just that which you load into your

John Becker addresses some of the general points readers have raised. Have

you anything interesting to say? Email us at [email protected]!

7/23/2019 EPE_07-1999



PC, its the code that is written for such things as PICs and other controllers. Then it gets more complicated in that the functions performed by the

software often become turned into hardware chips that have

the controlling programs actually designed into them as silicon (or other) semiconductor structures.

Incidentally, PICs are not the only microcontrollers that offer powerful solutions for

designers there are many

other types available from many manufacturers.

Periodically we are criticized

for the emphasis we put on software projects. In a nutshell,it is essential that we should.Whilst many simple designs can be constructed using “pure” electronics, once you start getting into slightly more sophisticated or complex requirements, it has become undesirable to attempt to do them without software (and I use the term in a very general sense) in some form or other,

whether its loadable code or set in silicon.

All Matthew’s points are worthy of further discussion,let’s hear from the rest of you!

APPRECIATING PLUGS

Dear EPE,

Thank you very much for

taking the time to review our catalog and your kind words said about it in your June ’99 issue. We worked hard on the catalogue and I am pleased that our hard work was worth it.

I am glad you are impressed by our products, it feels good to be praised by the experts!

Phil Goodman,Marketing Coordinator,

Pico Technology, Cambs

Thanks, Phil. On a personallevel I know the amount of workthat goes into designingproducts such as yours (myVirtual Scope of Jan-Feb ’98took months of my time).

On a wider level, we arealways pleased to publicize newproducts introduced by ouradvertisers, all of whom areinvited to send us a briefdescription, and a photo as wellif appropriate. We have always

offered this option through ournews pages. The wonder is thatso few advertisers actually takeadvantage of it, it is, after all,FREE publicity.

PC PORTS AND DELPHI

Dear EPE,

I was greatly enthusiastic upon reading your June ’99 Interface to discover you are finally thinking about dumping

the truly fantastic (for its t ime)language of QBasic, to go into the world of visual programming. However, your Delphi programming code did not recognize PORT as a command, although it is in the Help file. As with you, I too have had problems getting it to work with Delphi 4. I have an answer,though, the following standalone .EXE files work on my ’98 machine.

To output in QBasic, you would normally write:

OUT PortAddress,Value

To do the same in Delphi 4,in between the begin/end,write:

ASM

PUSH DX

MOV DX,$(Port

address, in hex) MOV AL,Value

OUT DX,AL

POP DX

END;

As for the read-from-port variant, before the begin statement write:

var

ByteValue;byte;

Then between the begin/ end write:

ASM

PUSH DX

MOV DX,$(Port

address, in hex)

IN AL,DX

MOV ByteValue,AL

POP DX

END;

Michael Ure

via the Net

We passed your letter toRobert Penfold who says thathe too had found a similarsolution and that he will discussit further in a future Interface.

It’s interesting to note thatmuch of your code is almostidentical to that which I haveused for many years whenwriting in 8086 assembler.

There are many examples inmy code for such projects asthe Virtual Scope (Jan-Feb ‘98)and PIC Toolkit Mk1 (July ’98).

The assembly software Iuse is the excellent sharewareA86/D86 Assembler/ Disassembler, in conjunctionwith Intel’s 8086 User Guide.The codes are seemingly

7/23/2019 EPE_07-1999



compatible with processors of the8086-upwards variety, includingPentiums. A further example ofsuch code will be shown inconnection with my forthcoming8-Channel Analog Data Logger,which uses the new PIC16F877microcontroller. It can beinterfaced to a PC via a mixtureof QBasic/QuickBASIC and ’86machine code.

The A86/D86 software isavailable from the Public DomainShareware Library (PDSL), DeptEPE, PO Box 131, Crowborough,East Sussex, TN6 1WS, UK. Tel:+44 (0) 1892-663298: Fax: +44(0) 1892-667473. It is well worthwhile asking for their catalog.

PhizzyBLY IMPRESSED

Dear EPE,

My compliments to all concerned on such an excellent product as the PhizzyB. The idea behind the concept is just brilliant! I would like to pass on a few ideas that I think all PhizzyB users would benefit from.

When people purchase a

product such as the PhizzyB, you can’t beat having a hard copy of the user manuals. If the books were in print people would not mind paying extra for that printed hardback/softback for referring to when experimenting and (playing)

with the PhizzyB a home made

printout version is not quite the same.

Will you be setting up a web- based user group, where people could submit their designs,circuits and code? Here people could share ideas, and there could be a community of PhizzyB users helping each other and learning together.

Is it possible to connect a QWERTY keyboard to a PhizzyB.If so, where can I find out more details? As a future venture I am

going to add process control,from which I can switch on lights, control temperature and set up cascade loops and much more. The list of ideas is endless. With a QWERTY keyboard connected it will be the ultimate project.

Thanks again for providing me and many other PhizzyB users with such an excellent and versatile educational piece of equipment. I am beginning to realize that the back-up behind this board is second to none and that you are available via email to help with any problems.I have found that questions are

always answered! I love the PhizzyB and I

can’t stop using it. My mind just works overtime thinking up new ideas and projects which the PhizzyB can be used to control.

Stephen (otherwise ANON)via the Net

We all love to be praisedand take the bows, but it’s“Max’’ and Alvin that Stephen isreally lauding, so Lord Max,what do you say?

In fact we are thinking aboutcreating a PhizzyB book, whichwould include all of the articlesfrom EPE and EPE Online , plussome of PhizzyB’s circuitdiagrams which were neverpublished thus far (anddiscussions on them), plus the

manuals

book.

everything in one

We think the web-baseduser group is a great idea if asufficient number of people areinterested. The first stage of thisis in the final part of the PhizzyB series (June ’99), where wesuggest that readers might wantto submit some PhizzyB -related

circuits to the Ingenuity Unlimited column. Also, if you(or anyone else) did design aninteresting circuit (or program),we’d be delighted to add it tothe EPE Online Library .

A bigger idea is tosynthesize the Beboputer intoan FPGA or an ASIC to create areal CPU that executes nativeBeboputer machine code (thiswould be much faster than thecurrent PhizzyB, whichinterprets it).

The real problem is thatwe’re up to our ears in other

projects at the moment one

such project is Bebop

Calculates Cunningly which willbe another Beboputer -basedbook/project. In this case thefront end interface to thesimulator will be a full-function

calculator the task of the book

is to teach integer and floatingpoint concepts, and to createthe assembly code that “goesbehind” each of the calculator’sbuttons. The great thing aboutthis is that we’re designing it tobe PhizzyB -compatible.

As to adding a QWERTYkeyboard to the PhizzyB, thisshouldn’t be too complicated atall. What you want is akeyboard that when you press akey presents the ASCII code onan 8-bit port that you canconnect to one of the PhizzyB’s input ports. Then you have towrite a program that loopsaround waiting for you to press

a key the program would then

have to decide what to do with

each key. Our book Bebop BYTES Back discusses suchprograms.

Thanks very much for your

enthusiastic support tell all of

your friends! Let us know howyou get on.

Clive “Max” Maxfield

7/23/2019 EPE_07-1999



ABSOLUTELY, KELVIN

Dear EPE,

I read with interest your letter of the month (June ‘99). I must agree with all that has been

commented on, however just to

throw some salt onto the wound:

When we are using the Kelvin scale for temperature, the word “degree” is always dropped. For example, water freezes at 273 Kelvin, water boils at 373 Kelvin,Absolute is 0 Kelvin.

Now it’s my turn to be super critical! I thought this might be of interest to some of your readers.

Maurice Clarke BABanbridge

Co Down, Ireland

Yes, I know so, but forgive achap when he’s hastily keying in(yes, I do my own keyboardthing!) two pages of Readout tomeet a press deadline, if heoccasionally overlooks a slip ofthe tongue!

In fact, as any dunderheadprobably knows, the name“degree kelvin” was discontinuedby international agreement in1967 (though why they shouldseemingly waste their time evenbothering to discuss it is beyondme). The unit is named after LordKelvin (1824-1907).

Just to be pedantic, a

temperature expressed in C is

equal to the temperature in

kelvins less 273·15 C (note the

small k for kelvins, although thesymbol itself is capital K).

Your criticism is taken in

good humor we’re quite happy

to be criticized as long as it’scourteous and constructive!

BAR QUOTES

Dear EPE,

If you must delve into

definitions, you should try to quote correctly! Referring to the letter Hot Bars in the June ’99 issue, a Bar is defined as 10

5

Pascal and hence a millibar is 100 Pascal, not a hundredth of a Pascal as you quote in Mr Phillips’ letter. The hecto prefix means 100x and so a hectoPascal correctly equates to a millibar.

T.B. OwenAberystwyth, Wales

But I did correctly quoteexactly what John Phillips saidin his E-mail. What you havespotted is that he “inverted” his reference (excusable, perhaps,as he lives “down-under”!).

DAC VAGARIES

Dear EPE,

Having read the very

interesting and practical Voice Record/Playback Module article by Robert Penfold in your April ’99 issue, his comments on the “vague description” of the “digital-to-analog converters” in the ISD chip are misleading.

In fact, this chip uses directly sampled analog voltage storage on what would otherwise be an EEPROM.Normally, such storage devices have their stored voltage quantised as high or low, but in this chip, a voltage is written and read out directly. This allows a whole “sample” to be stored on one cell, and does not require, for example, eight or twelve bits to store a digitized version of the sample.

Keith Lambell

via the Net

Robert replies:

On taking another look atthe ISD14xx data sheet, I seewhat Keith means. The

“multilevel storage array’’ doesseem to store voltage samplesrather than digitised versions ofthem. The data sheet for theearlier ISD devices is a bit moreforthcoming, but does not gointo any great detail about thisaspect of things.

These chips are not actuallydigital storage devices at all,and are more akin to “bucketbrigade’’ delay line chips than adigital storage medium. This

would account for the excellentaudio quality. Although onemight expect the samplevoltages to decay over a periodof a few years, the data sheetquotes a typical retention periodof 100 years!

Robert Penfold

CACHE CATCH

Dear EPE,

Ernest Flint’s article on PC microprocessors (PC Engines)in the May ’99 issue contains a few errors. In particular:

1. Celeron processors which have the 128KB L1 cache (the ‘A’ types) run the cache at the full CPU clock speed, not half speed as stated. So, although the cache is smaller than in the P II, it’s twice as fast. This information can be found at

www.tomshardware.com,the link given at the end of the article!

2. The original Pentium range started at 60MHz

3. The PIII range starts at 450MHz

As for why IBM used an Intel processor, I believe it was

http://www.tomshardware.com/





7/23/2019 EPE_07-1999


because they wanted to use off- the-shelf components, and all the required support chips were available for the Intel processor,unlike the other 16-bit processors of the time.

I got this information from Accidental Empires, by Robert X.Cringely (ISBN 0-14-025826-4).For anyone interested in the history of Microsoft, IBM, Apple and others, and the people behind it all, this book is great.

Anyway, thanks for an interesting article and magazine.And I like the new history section on the web site. The first issue of EE I bought was the combined

June/July 1980 one. I built my first project

amplifier

a 1W audio

from that.

Alan Edmondsvia the Net

Alan’s comments were sentto our Dave (Deputy Editor) whofelt the reply should come fromErnest Flint’s “mighty quill’’, whichprompted the following:

I’m not sure that I’m too keenon people talking about my“mighty quill” in public like this, Idon’t know what my motherwould think:-) Anyway, myresponse is as follows:

Arrggghhh! Alan E. isperfectly correct and I bow myhead in shame to the master (it’sa pity he didn’t write this articlehimself and save us all a lot oftrouble). The original Pentiumrange did indeed start at 60MHz,th P ti III d t t t

type processors do indeed runtheir L2 cache at the full clockspeed.

In my own defense, I wouldpoint out that we are all subject to

the occasional finger slip. EvenAlan in his point (1) refers to theCeleron’s “128KB L1 cache”,when in fact he meant to say “L2cache”.

On the brighter side, I’mdelighted that Alan found myarticle to be of interest and Ivery much appreciate his takingthe time to inform me aboutthese errors. I will now retire tomy study to chastise myselfsoundly and enter a disciplinary

note into my permanent record.

Ernest Flintvia the Net

We were slightly takenaback by Alan E’s comments onthe “double EE issue”. Checkingback on our bound copies forthat year, we were remindedthat indeed the June and Julyissues had been combined intoone. Editor Mike recalls that in

those days, problems betweenIPC (the then publishers) andUnions occasionally came to thefore. On this particularoccasion, there was a strikeover some matter or other(memory fails on what it was)and once it was over, there wasno option but to combine twoissues.

These days, as anindependent magazine underMike’s ownership, we pride

ourselves about being ontime,every time! (Hope I get extra

ENBRIGHTENING!

Dear EPE,

Would you know how I can contact Kingbright, the

manufacturers of some large 7- segment displays that I bought at a boot sale? I want to obtain the data sheets.

W. AltonBroadstone, Dorset, UK

Our Online Editor Alanlooked it up on the web: theirsite is at www.kingbright.comand Rapid Electronics are one

of their distributors, Tel: +44 (0)1206-751166.

http://www.kingbright.com/





Date post:	18-Feb-2018
Category:	Documents
Upload:	ipse-lute
View:	218 times
Download:	0 times

EPE_07-1999

Documents