The PSK-80 ﬁWarblerﬂ - njqrp.clubnjqrp.club/warbler/warbler tech manual - rev c.pdf · 6 PSK-80...

PSK-80 “Warbler” TechnicalManual- rev C 1

New Jersey

The PSK-80 �Warbler�

A Simple PSK31 Transceiver for 80m

Brought to you by the New Jersey QRP Club

Technical Manual - rev C

PSK-80 “Warbler” TechnicalManual- rev C 32 PSK-80 “Warbler” Technical Manual - rev C

PSK-80 �Warbler�A Simple PSK31 Transceiver for 80m

Technical ManualHere’s a little brother to the popular PSK-20 PSK31 transceiver de-signed by Dave Benson, NN1G in 1999. This time, Dave simplifiedhis design and took advantage of inexpensive crystals for a popular80m frequency to create the “PSK-80”. Wishing to proliferate an easyand inexpensive way for hams to get on 80m with PSK, NN1G sharedthe design with the NJQRP Club who is kitting and selling the project world-wide. They named the design “The Warbler” for PSK’s characteristic two-tone sound through the speaker, as well as for the club’s state bird (the mos-quito!). As it turns out, 80m is a great band for local (200 mile radius) commu-nications, and the PSK-80 Warbler is the perfect, narrow band, low power rigfor club round table QSOs and regular fireside skeds with your friends.

XtalFilter

T/R Switch

PowerAmplif.Driver

AF Amp

Audio Out

XtalFilterAudio In

(Q1) (U1)

Ca rrier Osc.

(U1)

Preamp

L PFilter

(U3)

(U2)

Y1/Y2

Y4-Y6

Q9

Q5/Q 6Q3/Q 4

Figure 2: Warbler block diagram

Last winter, while preparing my PSK31talk for Atlanticon, I was casting aboutfor a simple means of putting a PSK31signal on the air on HF. It struck me thatthe standard PSK31 ‘watering hole’ on80 meters is at 3580.15 Khz. This broughta wave of nostalgia for the New EnglandQRP Club’s old ‘colorburst’ evenings,wherein a half-dozen or so of us wouldbe camped right around 3580 Khz. Therigs for this affair were a simple 3-tran-sistor VXO transmitter. It was always athrill hearing multiple QSOs within thespace of just a couple KHz!

I can’t promise to keep the project to threetransistors, but the concept is an intrigu-ing one- a standard 80M frequency andsimple transceiver used as a club ‘inter-com’ over a fairly wide geographical area.The hardware design started with anevaluation of simple filters using color-burst crystals . Figure 1 below shows anexample of a 3-crystal filter and its pass-band response.

The asymmetric skirt response is typicalof a Cohn filter- the upper-frequencyslope is steeper. We take advantage of thisby setting the carrier/BFO on the high

side of the passband. This yields betterrejection of W1AW and results in lower-sideband (LSB) operation. Because thefilter uses series-resonant type crystals,the passband is actually above the markedcrystal frequency. The BFO is pulled tothe high side of the passband using a smallvalue of capacitance in series with theBFO crystal.

So what do we do with a filter which op-erates right at the RF? For starters, con-sider a ‘Neophyte’ DC receiver- it’s noth-ing but a product detector and AF amp.By tacking a crystal filter right on thefront end, it’s still a DC receiver, but itsselectivity and its resistance to out-of-passband IMD are considerably im-proved! On the transmit side, adding sucha filter to the output of a balanced modu-lator creates an SSB signal right at theRF frequency. Take a look at the blockdiagram in Figure 2 ... Pretty simple, eh?Naturally, reducing this simplicity topractice always seems to involve addinga few components.

The complete schematic of this trans-ceiver design is shown on the full-pageinsert to the center of this manual. Pleaserefer to it for the following discussion.

Let’s start with the transmitter. Audiofrom a computer’s soundcard output

(LINE OUT) is connected to J2. Q1 is‘closed’ during transmit and passes au-dio- and DC bias- to Q2. Several hun-dred mV p-p of audio are applied to mixerU1 to generate DSB energy at a (sup-pressed) carrier frequency of 3582 KHz.Crystals Y1 and Y2 and C6-C8 providea measure of filtering to eliminate the bulkof the unwanted sideband energy. Theremaining SSB signal is amplified by Q3and Q4. The output of Q4 includes an L-network (collector choke L1 and C10)which matches the driver impedance tothat of the PA stage.

Things begin to look a little differentaround the PA stage! The push-pull con-figuration used here was done for tworeasons: to improve available outputpower and for better spectral purity. Tothe extent that the two halves of the am-plifier show equal gains on their respec-tive half-cycles of conduction, this bal-ance pays off in rejection of second-har-monic energy. Input trifilar transformerT1 splits the drive signal into two out-of-phase drives to Q5 and Q6. Anothermultifilar winding (T2) combines the col-lector outputs. The third winding on T2is wound separately and has a differentturns count. It’s set for a collector imped-ance of 12 ohms- in theory, up to 6Wshould be available from this stage. In

Figure 1: Receiver passband response

0 d B

-2 0 d B

-4 0 d B

3 .5 79 54 5(m arke d c rys tal fre q.)

3.580 3.581 3.582 3.5833.579

(W 1AW )

3.578

'S ta nd ard' P S K 31 Freq ue ncy

Fc

S ettin g the car rie r o n the u pp er s lop e y ield s b etter reje c tio n o f a lte rn ate-s ide ba nd en erg y.

3 58 0.5

(RF output for 1500 Hz tone from sound card )

F(M H z)

R X F ilte r R esp on se (dB - no rm a lized )


Refer to the full-page Board Layout dia-gram provided on the sheet the the centerof this manual. Components are installedon the silkscreened side of the board.

TOOLSYou’ll need te following tools:- soldering iron ... ~25W- 60/40 solder, small dia- Diagonal cutters- Needle-nose pliers- Small slot screwdriver- Wire-strippers (recommended)- Tweezers (recommended)

TEST EQUIPMENT- Multimeter

LET’S BEGIN!Install only 3-4 components at a time be-fore soldering. Doing so makes it easierto check progress as you go, and reducesthe chances of leaving some component

connections unsoldered.

1) Open the main component bag sup-plied with the kit and verify its contentsagainst the Parts List on the next page.Ensure that you have all parts by puttinga check mark next to each line once veri-fied. Let us know if you are missing any-thing. Components are described by theirreference designators (C1, R3, Y1, etc.,as shown on the Board Layout diagramand Schematic), their values, and when-ever possibleby some distinguishing char-acteristic like size, color, physical mark-ing. There may some discrepancies inthese descriptions due to alternate partssupplied by the vendors, but they shouldbe generally okay and useful in identifi-cation.

2) There are two surface mount ICs forthis project. When installing, be sure tomatch the orientation of the device to thatshown on the pictorial drawing. When the

lettering on the IC reads “upright”, thechip is in the same orientation as shownin the Board Layout diagram. Then, pin1 is in the lower left corner of the device.This part may be held gently in place witha pair of tweezers and tacked down byone lead. Once in place, remaining leadsmay be soldered. A sparing hand with thesolder is recommended, but solder bridgesmay still be difficult to avoid because ofthe close pin spacings. If this occurs, usesolder-wick to clear any solder bridgesbetween adjacent IC pins.

3) Q1and Q7 (2N7000) are static-sensi-tive. Handle this part as little as possible,and if possible, ground yourself wheninstalling this device.

4) Diode D6 (near J4) is bent for “up-right” installation on the board. Installa-tion polarity is as shown below. Be sureto note the orientation of the silkscreenedcircle on the board and install the diodebody over this hole. The cathode(banded) end of the diode is oriented atthe top. The remaining diodes are in-stalled ‘lying down’- match the bandedend to that shown on the silkscreen.

Insta llation- Picto ria l

D iode-Schema tic

banded end is cathode

Com ponen t Silkscreen

Kit Assemblypractice, though, IR drop through R13/13A and saturation effects on Q5 and Q6limit this to something a bit less.

As a result of the balance provided by thepush-pull configuration and the improve-ment in 2nd harmonic performance, theoutput harmonic filtering can be consid-erably simpler. In this design, it’s a singlesection! Thanks to the 15-20 dB of 2ndharmonic suppression inherent the push-pull configuration, the minimum har-monic rejection for this design is 33 dB-compliant with current FCC regulations.

Components C13,D3-D4, and L3 are aT-R switch and provide a measure of front

end selectivity for the receiver. Q9 is apreamplifier stage, providing 10 dB ofgain. Aside from the gain feature, its otherpurpose is as a ‘flat’ 1.5K ohm sourcefor the crystal filter. The initial designlacked this stage and the filter passbandwas poorly-behaved. Y4-Y6 and associ-ated capacitors provide the passband se-lectivity charted earlier. Mixer U2 con-verts the filter output to audio, where it’slow-pass filtered and amplified by U3.R23-24 and Q7 provide a muting func-tion to prevent feedback problems dur-ing transmit- this is described in moredetail in the June 2000 issue of QST.

Comparator U4 and related componentsprovide T-R sequencing. This logic turnsPNP switch Q8 on to provide transmitterbias.

A prototype Warbler mounted in an LMB-139 enclosure ... perfect fit!


5) When installing the crystals on theboard, space these components slight(~0.5 mm) above the board to precludeshorts between the crystal cases and boardtraces. This is best accomplished by sol-dering the crystals in place and then re-heating individual connections and pull-ing the crystal gently to pull it slightlyaway from the board.

6) The monolithic capacitors used at C11and C12 are somewhat fragile. Use cau-tion when adjusting the lead spacing tomatch the hole locations on the PC board.Hold each capacitor body firmly betweenthumb and forefinger when spreading theleads apart.

7) T1 is wound using three 4" lengths ofwire (leave insulation on) removed fromthe 4-conductor cable supplied. Four (4)turns of these three wires are wound ‘flat’on a dark grey core, with no wire cross-ing over its neighbors.

Example: Assuming red, yellow andgreen wires are used, if ‘red’ leads in thedirection you’re winding the core, ‘red’will lead all the way around the core.Once these windings are prepared, trimthe excess lead length to 1/2" (1.2 cm)maximum and remove the insulation fromthe protuding leads. (It may be helpful tocut the lead lengths at an angle to easeinstallation, since only one lead-pair at atime then needs to be inserted into theboard. )

Install T1 as shown below.

8) T2 has a primary winding and sepa-rate secondary winding. The secondaryshould first be wound using 8 turns ofmagnet wire distributed uniformly aroundthe circumference of a dark grey core.Important: each time the wire goesthrough the toroid center hole it countsas a turn. Be sure that the turns do notcross over one another..

The primary is wound next using two 4”lengths of distinctly-colored wire fromthe 4-conductor cable. The two wires are“bifilar wound”, or wound together for 4turn s over the magnet wire secondary al-ready on the core. When complete, stripthe leads and install T2 as shown on thepictorial drawing. Note: It’s important toensure that the magnet wire leads arethoroughly stripped- these wire endsshould be scraped gently using a hobbyknife or other sharp blade. Unless ad-equately stripped, this magnet wire willnot accept solder properly, and a coldsolder joint will result. The magnet wirewinding connects to the ouput harmonicfilter and should be installed in therightmost pair of T2 holes. Refer to theT2 diagrams below:

PSK-80 “NJ Warbler” PARTS LISTC3 ........................................... 4-20 pF ......... trim cap - round, red insertC6-8,C15-18 ........................... 33 pF ............. disk cap - “33”C4 ........................................... 47 pF ............. disk cap - “47”C5 ........................................... 68 pF ............. disk cap - “68”C13 ......................................... 100 pF ........... disk cap - “101J”C22 ......................................... 150 pF ........... disk cap - “151”C10 ......................................... 330 pF ........... disk cap - “331”C11,C12 .................................. 1000 pF ......... mono cap - “102J”, blue or yellowC9,C14,C19,C101-106, C108-C110,C112 ................. .01 uF ............ disk cap - “103”C20 ......................................... .022 uF .......... mono.cap - “223” yellowC21,C107 ................................ .1 uF .............. mono. cap - “104” yellowC1,C2,C23.C24 ...................... 1 uF ............... electrolytic cap - blueC111,C113 .............................. 47 uF ............. electrolytic cap - blueD1,D2 ...................................... 1N5236B ....... 7.5V zener diodeD3-D5...................................... 1N4148 ......... diodeD6 ........................................... 1N4001 ......... diodeJ1,J2 ....................................... audio jack...... 3.5mmJ3 ............................................ RS-232 conn . 9 pin D-styleJ4 ............................................ coaxial pwr ... 2.1mm connectorJ5 ............................................ BNC .............. antenna jackL1 ............................................ 6.8uH ............ molded RF Choke: blu-gry-gld-slvrL2 ............................................ T37-2 ............. 23 turns #26 on the red coreL3 ............................................ 22uH ............. molded RF Choke: red-red-blk-gldQ2,Q3,Q4,Q9 ......................... 2N4401 ......... NPN transistorQ5,Q6 ..................................... 2SC2166C .... NPN power transistorQ1,Q7 ..................................... 2N7000 ......... MOSFET transistorQ8 ........................................... 2N3906 ......... PNP transistorR13,R13A ............................... 1 ohm ............ resistor: brown-black-goldR21 ......................................... 10 ohm .......... resistor: brown-black-blackR14 ......................................... 51 ohm .......... resistor: green-brown-blackR4 ........................................... 100 ohm ........ resistor: brown-black-brownR8,R12 .................................... 150 ohm ........ resistor: brown-green-brownR3,R10 .................................... 220 ohm ........ resistor: red-red-brownR5,R17 .................................... 470 ohm ........ resistor: yellow-violet-brownR18 ......................................... 1K .................. resistor: brown-black-redR6,R15 .................................... 1.5K .............. resistor: brown-green-redR2,R11,R23 ............................ 2.2K .............. resistor: red-red-redR9,R19,R30,R32 .................... 4.7K .............. resistor: yellow-violet-redR1,R22,R24,R25, R27,R29............................... 10K................ resistor: brown-black-orangeR26,R28 .................................. 22K................ resistor: red-red-orangeR7,R31 .................................... 47K................ resistor: yellow-violet-orangeR16,R20 .................................. 330K.............. resistor: orange-orange-yellowS1, S2 ..................................... DIP socket .... 8-pinT1,T2 ...................................... FT-50-43 ....... black toroid cores (see text)U1,U2 ...................................... SA612AD ...... SMT mixerU3 ........................................... MC1458 ........ dual amp, DIPU4 ........................................... LM393 ........... comparator, DIPW1........................................... 12" ................. telephone wireW2........................................... 24” ................. power cord (and mating coaxial plug)W3........................................... 24" ................. #26 magnet wireY1-Y6 ...................................... 3.579 MHz .... crystal, series-resonantHS1,HS2 ................................. heatsinks ...... incl. 4-40 nut + screwPCB......................................... Printed-circuit board


Hookup and AlignmentAs illustrated at the bottom of the previ-ous page, the Warbler connects to yourcomputer’s rear panel using the follow-ing connectors:

J1: to Soundcard “Microphone” / 3.5mmstereo

J2: to Soundcard “Line Out” / 3.5mmstereo

J3: to Serial port (9-pin D) / 9-pinmale/female

Setup Download DigiPan if you don’t alreadyhave it. This free software is best accessedthrough <http://psk31.com> and by fol-lowing the links to ‘software’. Afteropening the application, initialize the fre-quency display to ‘3582000’, and select‘LSB’, which places 3582 khz on the rightedge of the display. Connect an antennaand DC power (12-15V) to J5 and J4.Adjust the Soundcard microphone sensi-tivity control (you’ll need to look under‘Options’ after double-clicking on your

Windows Task Bar speaker icon.). Setthe level with this control to yield blue-to-yellow ‘speckles’ on the screen. Thisshould yield a band covering about one-third of the screen width when properlyadjusted.

AdjustmentThere’s only one adjustment on the trans-ceiver board itself- trimmer cap C3! TheARRL was kind enough to furnish us acalibration marker to adjust these rigs.Between the hours of 4-6 and 7-9 PMEastern (local time), W1AW is transmit-ting on 3581.5 Khz. You should havelittle trouble spotting their CW transmis-sions on-screen, and C3 is simply ad-justed with a jeweler’s screwdriver (Ra-dio Shack- $1.99) until W1AW is linedup under the 3581.5 tickmark on theDigiPan frequency display.

Note: If you’re close enough to W1AWthat their signal causes spurious traces onthe display, and reducing the microphonesensitivity to cure this effect causes

Figure x: W1AW provides convenient calibration!

9) Before installing Q5 and Q6, mountthe two heat sinks to them using the sup-plied 4-40 screws and nuts. Install thetransistors into the recess in the heat sinkso that the heatsink fins ‘cradle’ the tran-sistor.

10) Install final amplifier transistors Q5and Q6 last. Observe the installation ori-entation as shown on the pictorial draw-ing. These parts are quite large and ifinstalled early, will make it more diffi-cult to install nearby components. Thetransistor leads also prone to fatigue ifbent too many times during construction.

11) Make output filter inductor L2 bywinding 23 turns of the #26 magnet wirearound the red T37-2 toroid core. Use theillustration above for guidance. Clip ex-cess length from the ends, strip the enamelfrom the ends with a sharp hobby knife,and solder in place on the board.

12) A length of power cord is suppliedin the kit, along with a coaxial plug tomate with the power connector on theboard. Noting polarity (center pin is posi-tive), wire this cable and plug to connectwith your favorite 12-15V power source.15V is preferred for maximum poweroutput.

C 20 28

CO MPUTER

MICR. IN

LINE or SPKR OUT

'AF IN '

J1

SERIAL PORT

J2

'AF OUT'

'SERIAL PORT'J3

'PWR'

'ANT.'

S ou ndc ardIn terfa ce

'Warb ler' Board

1

1

2

CABLE DESCRIPTION:

1 3.5mm stereo (3-cond) to 3.5mm stereo (3-cond)

2 DB-9 (Serial po rt) Female to DB-9 male Cable Radio Shack # 26-117 or Jameco (800-831-4242) #25700

Radio Shack # 42-2387 (2 required)

J4

J5

Cableconnections

between the

Warblerand

computer


GETTING FULL POWER FROMYOUR WARBLERSome folks have reported difficulty ingetting full power out from their War-blers. To that end the designer Dave,NN1G, has suggested changing C10 from220 pf to 330 pf. This has helped someget 3-4 watts PEP on a 13.8 V power sup-ply as opposed to 2 W before the change.Here’s the word from Dave: “The NJQRPclub has already incorporated this changeinto the upcoming run of Warbler kits. Ifyou already have a Warbler, there’s noreason to make this change unless you’regetting inadequate power output from therig. If you’re already getting rated powerout, this change will not yield more out-put from the rig.” For what it’s worth I’vebuilt two of them, a beta and a produc-tion version and both have produced fullpower with the proper power supply volt-age. Individual mileage may vary depend-ing on component tolerances. ... de JoeE., N2CX, email: [email protected]

BEING AN SWL WHILE WAITINGFOR THE WARBLER TO ARRIVE

I can’t wait till the 2nd shipment of War-blers departs NJ! I’ve sent my money —but I have the fever now!! I was so des-perate after reading all the posts fromDoug, Bill, et. al., that I actually went toRadio Shack and purchased a stereo-mini-to stereo-mini patchcord so I could listento PSK! The DigiPan freeware is fantas-tic!!! Even if you don’t have a Warbler,or any other way to transmit PSK, YOUOWE IT TO YOURSELF TO LISTENIN ON THE PSK QSO’S!! The softwareis free (look for the links on http://www.psk31.com/ ) and if you have acomputer with soundcard in the shack, itis as simple as RCVR Audio output(headphone jack) - patch to - soundcard

LINE input! ... de Alan Kaul, W6RCL,LaCanada, CA [email protected] , http://home.att.net/~alan.kaul/index.html

IMPROVE WARBLER “IMD” WITHPOT ON AUDIO INPUTOne evening during a round table I askedseveral stations to check my IMD. I wasdismayed to find it was only -15 dB. Thishad to be fixed. I discovered that theWindows 98 “volume” control settingchanged in rather coarse steps, not lin-early. I was already down to the lowestpossible step that produced output butwas still overdriving the sound card. Thecure was to break the PC board trace be-tween J2 and the junction of capacitorsC1 and C101 and insert a 100k ohm pot.Audio from the sound card is fed to theoutside pins of the pot (from J2) and thecenter (slider) connects to C1 and C101.Now with the Win98 volume control setto mid scale I can adjust the audio drivefrom zero to four watts output withDigiPan in Tune mode. I adjust the potto the point where the rf output just stopsincreasing and then back down slightly.I am happy to report that my IMD hasgone from - 15 dB to -24 dB. ... de BillJones, KD7S, [email protected]

MORE MOUNTING DETAILS FORPSK-80 IN LMB-139 ENCLOSURE

If you use the 1/2" datum shown on thediagram for the audio jacks to the insidebottom of the enclosure “floor”, and mea-sure your Warbler, you’ll note that a 5/16" standoff would be needed. That is,put the 1/2" mark of a scale at the centerof one of the audio jacks and you’ll seethat the length of the scale protrudingbelow the board is the length of the stand-off you would use with the panel dimen-sions shown. I think I used a 1" long 4-

Tips & Techniques forBuilding & Operating your Warbler

PSK31 signals to disappear into the mud,a different approach is needed. In thiscase, set the DigiPan start frequency to‘3581500’ and adjust C3 so that W1AWis zero-beat at the right extreme of thedisplay. This takes advantage of the low-frequency rolloff characteristics of thereceiver to knock their signal down tomanageable levels. I live about 2 milesfrom W1AW and they’re very stronghere. Once this adjustment is performed,though, I can copy PSK31 signals with-out interference

Transmit adjustment:Click on ‘Tune’ on the DigiPan menu-this sets the transceive into transmit modeat 100% duty cycle. Click on the speakericon and advance the volume slider untiloutput power (viewed on a watt meter)no longer increases. Back it off slightlyuntil power output begins to decrease- thissetting assures linear operation. (If youdon’t have a watt meter, a 50-ohm resis-tive load and peak detector circuit willaccomplish the same job effectively. )

Operation:Clicking your computer’s mouse cursorover the typical “railroad track” PSK31should cause text to begin appearing inDigiPan’s upper text window. Clickingon ‘TX’ in the DigiPan menu puts you in‘transmit’ mode, and your typed text inthe lower window is streamed out on theair.

Results:Success with 80M PSK31 is considerablyinfluenced by antenna performance.NN1G’s present 80M antenna is a hast-ily-erected dipole up only 12-13 feet atthe center, and he’s routinely able to work

DigiPan’s creator Skip, KH6TY/4. He’sin a condo in SC and his antenna is uponly 25’ maximum -- one side of it runsthrough the attic, the other end slopesdown to the ground at the far end. An-other SC station reports a considerablybetter copy with a better antenna setup.For QSOs over shorter paths the situa-tion improves considerably. A ragchewwith a VE2 about 250-300 miles awayyielded 95% or better copy at both ends.

FURTHER INFORMATIONIf you have any questions concerningyour PSK-80 Warbler Kit, please checkthe NJQRP website page maintained forthe Warbler. It’s located at http://www.njqrp.org/warbler/. Useful tips &techniques are also collected on the KitNotes web page at http://www.njqrp.org/warbler/kitnotes.html.

You may also contact:George Heron, N2APB2419 Feather Mae CourtForest Hill, MD 21050email: [email protected]

The NJQRP wishes to thank DaveBenson, NN1G, for his gracious supplyof the PSK-80 Warbler design to theNJQRP. With it we’re able to proliferatethis simple-yet-effective means of gettingonto PSK throughout a wide AmateurRadio community. Simultaneously, theNJQRP is able to leverage the small prof-its from its kitting efforts back into QRPby hosting its annual QRP forum calledAtlanticon.

NN1G may be contacted at:Dave Benson, NN1GSmall Wonder Labs80 East Robbins AveNewington, CT 06111email: [email protected]://www.smallwonderlabs.com


“original” style it is on the flat side.

STANDARD FOR RS-232 CON-NECTOR ON PSK-xx BOARDSI’m happy (so far) to report that effectivewith all new PSK-series shipments, thetransceiver will use the female DB-9 con-nector to allow use of the ubiquitous se-rial-port extender (male-to-female) cable.This is *not* merely a matter of supply-ing the other connector flavor- the twogenders are physically ‘mirror-imaged’and they are not interchangeable on theprinted-circuit board. I’m awaiting a ship-ment of boards which incorporates achange in this regard.Along the way, I’d heard from severalpeople who threw away the connector Iprovided and specified and installed onewhich matched the picture in QST. Badmove- that picture made a QST deadlinewith only hours to spare, and once I ac-tually applied power, quickly discoveredthat there was little Transmit Joy!Anyhow, there’s nothing to ‘vote’ on.Both the PSK-series and the Warbler willnow use the female board-mounted con-nector. If you’ve already got a PSK-20and want to switch back and forth to aWarbler or a PSK-40 <g>, a ‘gender-bender’ plug (straight- not null modem)is available from Radio Shack, Jameco,etc .... de Dave, NN1G

FINDING THE LMB-139 ENCLO-SUREIn the Tech Manual, and during my PSKtalk at Pacificon, we showed a nice andconveniently-sized Warbler enclosurecalled the LMB-139. There’s really noth-ing special about this aluminum clamshelldesign from LMB — one could use justabout any handy enclosure in the junk box(or even make a NJ Homebrew PCB en-closure!!) — but it just turned out thatthe LMB-139 fits snugly around the PSK-80 Warbler circuit board. I’ve used theseenclosures for several years for my QRPprojects in the shack, and you can see

some of the results at http://www.njqrp.org/mbrproj/enclosures.htmlTom Jennings, KV2X tells us that youcan order the LMB enclosure directlyfrom LMB (http://www.lmbheeger.com/products.htm). John Sielke, W2AGN tellsus the LMB 139 box is Mouser Part #537-139-P @ $7.36. It is 5.5 x 3 x 1.25.LMB has another box, just the same size,but fancier and with dark blue top andlight blue panel (their “Crown Royal” se-ries), which is PN 537-CR-531 for $8.91.NN1G reminds us that many other en-closures will also work, so check out theJameco (http://www.jameco.com),Mouser (http://www.mouser.com) andDigi-Key (http://www.digi-key.com)sites! ... de George, N2APB

RUNNING THE PSK-xx BOARDSWITH A MAC?A number of QRPers have written to askif there’s a software program for theMacintosh equivalent to Digipan on thePC that controls the PSK-xx boards. Wellthere is! Check out Multimode at http://www.kender.es/~edu/psk31.html Thereis also a neat new one for Linux by aGerman, but the link eludes me at themoment. ... de Skip KH6TY/4

WD9EYB RUNNING A WARBLERGROUP BUILDING PROJECTI have slapped on some of the things Ihave said about a group building projectof PSK-80’s for West Central Indiana andEast Central Illinois at http://bu t le r .q rp .com/~wd9eyb/psk80/rantings.html ... Jim, WD9EYB,[email protected]

KD7S STARTS “WESTERNWARBLIERS” NET - SUNDAYS7PM PSTI’ve always wanted to coin a new wordor phrase, so here goes: Warblier(warbel-ear), n 1. One who operates PSK-31, esp. with a NJ QRP Club PSK-80Warbler transceiver. 2. Anyone runningPSK-31 (BPSK or QPSK) with any kind

40 screw in each corner of the board, in-serted from the bottom of the chassis andsecured with a 4-40 nut to the chassis.Then I put another 4-40 nut onto eachscrew and ran it done tight against thefirst. I then carefully angled the boardconnectors into the panel cutouts andslipped the board onto the 1" screws. Fi-nally I secured the board with another 4-40 nut on each screw. So as it turns out,two 4-40 nuts were about 5/16" high. ...de George N2APB, [email protected]

FERRITE CHOKE ON AUDIOCABLES CURES STRAY RFINTERFERENCEIf I moved the cables going from thesound card to the Warbler, I found thatthe output power would change slightly.I put a ferrite choke on the cables and theproblem disappeared. Note that my an-tenna is a roof mounted vertical and oneof the 80 meter radials passes directly overthe ceiling of my shack. It is no more thanthree or four feet away. And the Warbleris sitting naked (no cabinet) on top of thedesk. ... de Bill Jones, KD7S,[email protected]

CHECK OUT eHAM FOR PSK-80BUILDER REPORTSPost your own reviews, or see those ofothers PSK-80 builders on eHAM at http://www.eham.net/reviews/detail/1137 ... deKen N9VV, [email protected] http://h o m e . u c h i c a g o . e d u / ~ k h o p p e r /psk80project.html

LOWERING WARBLER AUDIOOUTPUT ON COMPUTERSHAVING ONLY “MIC” INPUTBecause of the fact that I am using aToshiba 2595CDT Laptop for PSK-31with the “Warbler”, and the laptop doesnot have an audio “Line” input, only a“Mic” input, the receiver audio from theWarbler was much too high for the laptop,making the entire received spectrumbright yellow. Even with the laptop’s re-corder mic volume turned all the way

down, the screen was still mostly yellowin the received spectrum area. So, aftertaking a look at the schematic, I changedR23 to a much higher value. After acouple of experimental tries, I settled on56K as a new value. With this value, theaudio level to the mic input now givesthe proper amount of display adjustment,from almost totally blue to pretty heavilyyellow speckled. So, if your waterfall dis-play is mostly or all yellow in the receivedspectrum and your volume control won’ttake it to the level you want, try chang-ing R23 to a higher value. ... de WayneNB6M, [email protected]

Q1 & Q7 PACKAGING ANDMOUNTINGWhen the Warbler manual was done, theFETs were packaged in the familiar TO92style package which has a pronounced“D” cross-section. The manual andprinted circuit board component markingwas for that device. However mid-waythrough the kitting we had to make aquick turnaround order for 2N7000 de-vices and we got them from a differentdistributor, what we got was in a differ-ent case type. So, if your Q1 and Q7(2N7000 transistors) have the familiarpronounced “D” cross-section packagesfollow the manual and pc board mark-ings. (This package style is the same asthe 2N4401’s used at Q2, Q3, Q4 andQ9.) That is, the flat side of the transistorwith printed identification matches theflat side outline on the board and therounded side matches the rounded out-line on the silk screening. But if the2N7000’s you got have a flatter packagewith slightly beveled edges on one side,you mount it basically the same way sothat the flatter side faces the flat side ofthe board marking, and the side with thebeveled edges faces the rounded edge ofthe board marking. On this “alternate”style transistor package, the printing ison the more rounded side while on the


1. Power cord (NJQRP supplies the cordand the plug, NICE touch)2. BNC cord to connect to your antenna.If your antenna coax ends in a PL259,then you will need an adapter to connectto the BNC connector on the Warbler.3. DB9 Male to DB9 Female cable toconnect between the Warbler and the se-rial port on your computer.4. Two shielded audio cables with stereo1/8" plugs on both ends. One connects tothe Mike input of your sound card andthe audio output of the Warbler, and theother connects to the audio ouput of thecomputer and the audio input of the war-bler. (Hope I got that right, grin). 5.Digipan or similar software installed onyour computer. (Available from a link onDave Benson’s Small Wonder Labs page,www.smallwonderlabs.com look underPSK31)The antenna that I am using is a SkeltonCone up about 30 feet at the apex and 6to 10 feet at the end of the radiators. Askelton cone is basically a double dipole.Think of having 2 radiators on each sideof the center insulator spread apart, in-stead of just 1. It is fed with ladder lineand tuned with a tuner. Works on allbands. The legs are 51 feet long, and thefeedline is about 50 feet.With my setup, I am able to work bothclose in and dx stations, having solid copyon the locals, and being able to workNW7DX from Washington and KI0ROfrom Colorado. Hope this helps. ... deDoug KI6DS

WHAT KIND OF DISTANCE AREYOU 80m PSK OPS GETTING?I’m using a low dipole at my Connecti-cut QTH - it’s up only 12-13' at the cen-ter. Copy is solid out to 200+ miles, al-though it varies some with conditions. My2-way with Joe, N2CX, near Philly wasabout 80% both ways with our Warblers,and I’ve had a rock-solid QSO with a VE2in Sherbrooke-about 250 miles. Best DXto date has been the Atlanta area but it

was a struggle. Distances seem compa-rable to what you’d work using CW.Just since I first put the Warbler proto-type on the air last spring, I’ve watchedthe PSK31 activity growing on 80M.Here in the Northeast, you’ll usually seeone or more traces on the display at night,especially during W1AW’s quiet period(9-10 PM Eastern). I’ve seen as many as4 simultaneous traces in the Warbler’s 1Khz passband. That doesn’t even beginto fill that slice of spectrum with PSK31!We’ve got a unique opportunity here to“make it happen”. Don’t let the fact thatit’s 80M, as opposed to a ‘DX’ band, deteryou in the slightest. Many of us are run-ning short (or low) antennas and still get-ting out. Joe Everhart’s ‘Squirt’ antennais half-sized and gets out quite well- lookfor it in the next issue of the NJ QRPclub’s ‘Homebrewer’ journal. From thesound of Doug’s recent post, none of theantennas in their wonderful 3-way QSOwas a ‘killer’ skywire. This is the seasonto begin thinking about the low bands,and enthusiasm for 80M yields creativ-ity in the antenna area. This is somethingwe all enjoy hearing about, whether we’reWarbleHeads or not, and this enthusiasmgets us off our computers (maybe) and*on the air*. ... 73- Dave, NN1G

CUTTING THE REAR PANEL OFYOUR ENCLOSUREA convenient diagram on the next pageshows hole dimensions for the “rearpanel” of any enclosure you might hap-pen to put the Warbler into. As illustrated,the enclosure happens to be of the LMB-139 aluminum clamshell, but the dimen-sions are really only specific to the edgeconnectors of the PSK-80 board. Hence,you can use this diagram as a template(when reproduced at actual size) for mak-ing the cutouts in your own panel, or justuse the dimensions to map out the holesneeded on any panel you might want touse. ... de George N2APB

of transceiver, QRP or QRO. GeorgeHeron sold a hundred Warbler kits atPacificon [and another 100 via mail]. Iwould guess that 50% of those kits wentto west coast QRPers who live within acouple hundred miles of the bay area. Ifyou’re one of those, how about joiningDoug (KI6DS), Dave (AB5PC) and my-self (KD7S) in a round table chat on 80meters. Set up DigiPan on 80 meters at3582 kHz, BPSK, LSB and a 1,500 Hztone. Activate AFC, NET and SNAP. Ifyou chose to use Squelch, set the thresh-old fairly low (no more than 2-3 notchesfrom the bottom). I’d like to suggest wecreate a “Western Warbliers” round tableQSO party on Sunday evenings beginningat 7:00 p.m. PST. No net control, no for-mal check-ins, just a round table discus-sion using PSK-31 running 5 watts or lessfrom any rig (but especially the Warbler).Who’s in? ... de Bill Jones - KD7S,Sanger, California http://www.psnw.com/~kd7s/

MONITOR PSK31 ACTIVITY ONPSK-xx BOARDS WITHOUT ACOMPUTERI monitor the receive audio output with asmall amplified speaker and can tell ifthere is activity on the frequency withoutusing the computer. ... Dave Epps,AB5PC, Fresno, CA

EASY WAY TO MOUNT CRYS-TALS ELEVATED FROM PCBTo allow for spacing between the crys-tals and the board I cut a strip of heavypaper (or whatever) the thickness of thecrystal. I then cut a slot in the center ofthis strip and slide it between the crystaland the board. Solder and pull out thestrip. ... Dave Epps, AB5PC, Fresno, CA

ALIGN YOUR WARBLER WITH-OUT W1AWI’ve heard talk about how some PSK-80Warblers that don’t hear well or have lowpower output. One possible explanationis that they are not aligned properly. The

manual touches briefly on alignmentusingW1AW as a frequency marker.Unfortunately, some of us on the westcoast can’t hear W1AW on 80 meters. Iused my own transceiver to align myWarbler. Here’s how ...1. In DigiPan, click on the “Configure”menu.2. Click on “Band.”3. Click the “Activate” button for 80munder “Band.”4. Under Spectrum Start” type 3582.5. Under “Spectrum Options” click onLSB.6. Click “Okay.”7. Fire up your Yaecomwood, set it upfor QRP and CW operation.8. Attach a dummy load to theYaecomwood.9. Set the dial to exactly 3581.00 kHz.10. Key the Yaecomwood. You shouldbe able to see its signal in the DigiPanwaterfall.11. Adjust trimmer C3 on the Warblerboard until the signal is directly under3581 on the DigiPan display. Your War-bler will now be calibrated to the accu-racy of the frequency display on the bigrig.Click the waterfall at about 3580.5 kHz.From the DigiPan menu, click “Mode”and “Tune. With a 13.8 volt power sup-ply you should see about3 or 4 watts out-put. Your transmitter output powershould drop off below 3580 andabove3581 kHz. That’s normal. On myWarbler, the “sweet spot” is right at3580.5 kHz and is fairly flat 400 Hz ei-ther side. ... de Bill KD7S

HOW DO YOU HOOK IT UP??One of the questions that I had about mywarbler once I got it built was how tohook it up to the computer and the out-side world. Bill Jones, KD7S and PaulMaciel, AK1P have been my PSK gurus.You will need the following to connectand get on the air:


MAKING A PANEL TEMPLATE &USING NUT/BOLT STANDOFFSCut a piece of heavy paper (file folderstock) to the exact dimensions of the frontof your enclosure. Then using an inex-pensive (plastic, $1.00) calipers, or evena ruler, measure the positions of yourcontrols, etc. with reference to some land-mark, such as the bottom right corner ofthe PC board. Lay these dimensions outon your paper replica and make the holeswith a paper punch. Then test it out, mov-ing the holes as necessary. When every-thing is right, cut a new replica and trans-fer your accurate dimensions to it to makea template. Then proceed as you wouldwith a “real” template. With regard to theusing nuts as a standoff. You can makethis adjustable if you run a screw (4-40size, for example) up through the bottomof the enclosure and secure it with a nut(and possibly a lockwasher). Then threadanother nut down the screw to achievethe desired height. Leave some spacebetween the nuts, and you can make veryfine adjustments of the standoff height.Then mount the board on the screws andsecure it with a third nut. ... de RichardMeiss, WB9LPU

FIRST WARBLER-WARBLER-WARBLER QSO: KD7S-AB5PC-KI6DSAlthough Chuck Adams, K7QO had thefirst production-version Warbler contactlast week right after Pacificon, Bill JonesKD7S and Dave Epps AB5PC had thefirst production-version Warbler-to-War-bler 80m PSK31 contact on October 28.And to top it off, Doug Hendricks, KI6DShopped in for the first Warbler round tableQSO with Bill and Dave! Read moreabout it below:

Oct 29, 2000 ... Wow, Wow, Wow!!What a great weekend. I built my NJ QRPClub PSK80 Warbler kit yesterday, andtook it to Sanger to get a crash course inoperating PSK from Bill Jones, KD7S.He had me up and running in no time,helped me install the software, showedme the cable connections, and I was offfor home so we could make the first War-bler to Warbler production board qso. Igot home after an hour and a half drive,fired up the rig, and there was a qso go-ing. It was Dave, AB5PC and Bill, KD7S,who were both using their warblers. Darn,I missed out on a chance to make history.Then I grinned, typed, break de KI6DS

and waited for a break. I got my chanceand jumped in. Bingo, now it was athreeway between KD7S, AB5PC andKI6DS, all using home built PSK80 War-bler transceivers and we had almost solidcopy for about a 45 minute round table.

So, Dave and Bill get the honors for thefirst ever Warbler to Warbler qso, but Iwas part of the first ever roundtable withproduction warbler kits. Guys, this isFUN FUN FUN. Hey, it might even beillegal it is so much fun. The round tablepart is great. I live in Dos Palos, Davelives in Fresno 60 miles away, and Billlives another 20 miles or so farther southin Sanger. Bill was using a vertical, I wasusing a skelton cone up 30 feet, and Davesaid his antenna was terrible, and he didnot elaborate, but did say he was doingsome antenna work tomorrow.

Man, $35 is all that it costs to get on thisnew mode!! Thanks to Dave Benson thedesigner and the NJ QRP Club for kittinga fabulous kit. The board is first class,NorCal quality, as are all the parts. What

a great kit this is. All the parts are there,and it is easy and fun to build. And youknow what? There are 2 surface mountSOIC 8 pin ICs and I didn’t even thinkabout it until after it was over. Hey, I justinstalled a surfacemount part, and didn’teven give it a second thought while I wasdoing it!! The SMK-1 has done its job.

George Heron sold 100 of these kits lastweekend at Pacificon, so I know the yare out there. Get them built guys, andlet’s set 8:00 PM Pacific time as a time tocheck the frequency for West Coast PSKactivity. I’ll be on tomorrow night and Ibet Bill and Dave will be too!!

Note, you will need to buy a db9 male todb9 female cable and 2 stereo 1/8" plugaudio cables, but that is all you need. Thesoft ware is free. Check the NJ QRP CLubweb site for full details. Ordering infor-mation for the kits is there too. Remem-ber $35!!!!!!!!!!!!!! A great, great kit.

One more comment. Dave Benson is agenius! ...72, Doug, KI6DS


terfall’ display at the bottom of the im-age. This display is a panoramic slice ofspectrum 4 KHz wide in the 20 meterband. New signals appear at the top ofthe waterfall just below the frequency tickmarks and proceed downward as they getolder.

In this image, there are 4 distinct ama-teur signals- this snapshot was taken earlyin the day under poor conditions. Thebroad trace on the right side of the water-fall is a Pactor station present through theentire time history onscreen. A secondPactor station may also be seen startingup (top center of waterfall.) The faint lineon the left of the waterfall is a steady car-rier, and although not evident in the re-production copy, is accompanied by anumber of fainter 60 Hz artifacts fromunintentional modulation from thatstation’s power supply. The signal we’reinterested in, though, is the distinct traceto the right of screen center; this is aPSK31 station calling CQ. The upper-most dialog window on the display showsthe decoded text from this station’s trans-mission- it appears onscreen as it’s beingdecoded.

You’ll note a small black diamond cen-tered over this signal’s waterfall trace.This symbol is placed by ‘pointing andclicking’ with the computer mouse. Ifyou tire of reading the mail with one sig-nal, click on the next one and ‘you’rethere’. There’s no need to touch any tun-ing adjustments, as the frequency selec-tion is done in the software. This point-and-click feature obviously represents asignificant advantage over narrow-bandtuning indicators, and is now being usedfor other narrowband digital modes aswell3.

As you can see from the upper dialogwindow, the early copy for this sessionis ‘gibberish’ resulting from thesoftware’s best efforts to make sense ofthe noise it’s receiving. Peter Martinez

touched on the rationale for (and largelyagainst ) the use of active error-correc-tion schemes for this live QSO ‘keyboard-to-keyboard’ mode. Since there’s no suchcorrection applied, the software providesan output whenever it decodes a ‘valid’character in the noise . As soon as thePSK31 station came on the air, I clickedon its trace and the text in the upper win-dow immediately changed to 100% copyfor this fairly strong signal.

The middle window (blank in this in-stance) is used to type in messages beingtransmitted. These can be either typedahead and then streamed out by clickingthe ‘TX’ button, or once you’re transmit-ting, the software sends the contents ofthis window as you type. If you’re asskilled a typist as I am <g> , only the plen-tiful use of backspaces and correctionskeeps me ahead of the outgoing data. Asyou transmit, your own typing is shownthe upper window- for better or worse.

The buttons across the top provide ac-cess to a number of menus used to setdisplay options and also to access vari-ous Windows functions such as “Save”,“ Edit” and so on. The “Call” and“Name” boxes are used to store personalinformation on the station you’re talkingto, and when you click on the “BTU”(‘Back to you’ button, it incorporates thatinformation into your transmission as youturn it over.

How does the mode work?Transmitted signals use a pair of toneswhich are separated by 31.25 Hz. Be-cause of the modulation method and re-quirements for the hardware, nearly allof the energy in a PSK31 transmission isin only these two tones. As a result, thismode is very spectrum-efficient. Intheory, anyway, you could fit about 80PSK31 QSOs into the space occupied bya single SSB voice transmission! A spec-trum display for this mode is shown be-low, this is a “clean” signal. In actual

Figure 8

by Dave Benson, NN1G

One of the most exciting new develop-ments to come down the road in amateurradio recently has been the explosion ingrowth with the newer digital modes. Ofthese, the best known to the QRP com-munity is the PSK31 mode.

What is PSK31?It’s a narrow band digital communicationsmode. It takes advantage of the consid-erable processing power available in yourhome computer, and uses Digital SignalProcessing (DSP) techniques and someother sophisticated filtering. As a result,it does a good job of recovering weak sig-nals from the noise, and that attribute iswhat makes it attractive to our commu-nity. The ‘price of admission’ is a com-puter with sound card and a stable SSBtransceiver. During the course of this dis-cussion, I’ll be referring to the Digipansoftware, this is free software availablefor downloading from the Internet1.

PSK31 & DigiPan ...A Short Primer

PSK31 was developed by Peter Martinez,G3PLX, and originally published inRadCom in Great Britain as a 2-part se-ries in 1995/6. It was subsequently re-printed in QEX2. As originally imple-mented, the PSK31 signal used a fixedtone-pair frequency and the accompany-ing receiver software featured anarrowband tuning indicator. This put apremium on tuning a receiver quite care-fully to put the tone-pair within the rangeof the tuning indicator software. Themost recent entry into this mode has beenthe DigiPan software, and within about amonth of its release has proven to domi-nate the mode. The reasons why will be-come obvious in a moment as we reviewthe screen-captured image from the com-puter.

The most dramatic feature in this image[at least in the color rendition] is the ‘wa-

Dave Benson, NN1G(as excerpted from the Atlanticon 2000 Proceedings)


variety of levels of sophistication. Someof the more recent mult-mode controllersinclude PSK31 modem functions in DSP,and this is a ‘Cadillac’ approach. At theother end of the spectrum, I know there’sat least one gentleman using a PIC to de-modulate PSK31 data. Without know-ing how robust his algorithm is, that is,how sophisticated the correlation processis, it’d be tough to predict how well thisapproach would work for all but the stron-gest incoming signals.

Operating adjustments: “Look, ma-no hands”

At present and for the foreseeable near-future, PSK31 activity is clustered withina narrow band of frequencies on eachband. This means that aside from dialingin a particular transceiver frequency,there’s nothing to be done with the trans-ceiver hardware after that. All operatingadjustments from there on out are handledby the computer sound card interface.The figure below illustrates the Windowssound card dialog box used to set Micro-phone Gain. This actually refers to theMicrophone input to the sound card, soit’s functionally the transceiver’s receivergain control. The sound card’s master

volume slider (in the Windows lowertray) sets the output level to adjust trans-mit output power.

Needless to say, with no adjustments onthe transceiver itself, this lends itself tosome pretty neat ‘remote’ applicationslike telephone or cross band HT control.There are now a number of Internet-ac-cessible amateur transceivers- if they’renot doing anything else at the momentyou can steer them to the PSK ‘wateringholes’ and monitor the goings-on fromanywhere. Just the thing for those busi-ness trips with a laptop but no rig!

Hardware requirementsComputer: The requirements for thecomputer portion of the system are prettysimply stated: Windows 95, 75 MHzPentium (or equivalent) or faster, soundcard and Serial port.

Transceiver: This is worthy of a littlemore attention, since many of us will beinterested in the homebrewing or kit-com-patibility aspects of this mode. PSK31uses SSB rather than the simpler (Class-C) designs used for CW for the transmit-ter design.

First and foremost, if you have a synthe-

practice, non-linearities in the transmit-ter chain will degrade this signal qualitysomewhat.

In the time domain, the signals closelyresemble the classic two-tone SSB testwaveforms, as you might expect from thespectrum plot above,

The above picture shows a stream of logic‘zeroes’. If you’ll observe closely, eachwaveform envelope has an opposite‘sense’- starting positive or negative-go-ing in polarity. A logic ‘0’ is defined bya phase-change at the bit period.

The figure below shows a logic ‘1’. Thisis mechanized by the absence of a phasechange, and as you see from the figure,the envelope is briefly a steady carrier foran extra bit period.

The character set devised by PeterMartinez is called Varicode, and likeMorse code, uses shorter code sequencesfor the most commonly-used letters. Thespace between characters is formed by

two or more sequential zeroes. Whenthere’s no data being sent but the trans-mitter is on, it’s sending zeros continu-ously so the waveforms resemble that ofthe first picture above. Each characterin the alphabet contains primarily ‘ones’-never is a ‘zero’ seen for more than onebit time within a character. The charac-ter set encompasses all 128 ASCII char-acters- alphanumerics, punctuation andcontrol codes- that last one a crucial ad-vantage for the average typist. I’ve grownto love my ‘backspace’ key!

Receiving PSK31 SignalsThe recovery software employs a ratherclever scheme to ensure good freedomfrom data corruption by noise, and this isone of the features of PSK31 whichmakes it attractive to our QRP commu-nity. In addition to the narrowband DSPfiltering which is centered over the tone-frequency pair, there’s a special-purposealgorithm at work as well. This is a so-called Viterbi decoder, and it consists of32 parallel functions. Each of these func-tions is examining the DSP output overthe last 5 bit times and ‘scoring’ the cor-relation between its own code (00000through 11111). In a sliding-windowfashion - at each bit time- the sixteencodes with the best correlation scores arekept and the next 32 parallel functionswill include the 16 ‘winners’ plus the two‘children’, i.e., a ‘1’ or a ‘0’ at the new-est bit time. This processing means thatthere’s a 5 bit-time delay or ‘pipeline’before received data starts trickling outof the decoder.

This correlation process may be extendedin time up to the point where the evalua-tion delay becomes objectionable to theuser- somewhere around one second orso.

This sounds like a fair amount of process-ing is involved, and indeed it is. There’snothing to prevent the homebrewer fromtackling this rather interesting topic at a


approaches are practical. For an approachwith the minimum number of cables be-tween the rig and computer, a VOX (au-dio-derived) switching scheme seems at-tractive. On the ‘plus’ side, the outgoingaudio to the transmitter is pretty well de-fined in terms of its wave shape. Severalpractical considerations soon intrude,however. Due to the wraparound / feed-back situation mentioned earlier, the au-dio out of the sound card is never reallyshut off. Further, with some of the Win-dows functions liable to make noiseswhen the operator performs a mouse-click, there’s a potential for triggering thetransmitter at unintended times. Users ofseveral different PSK31 software pack-ages have reported this behavior and havelargely shelved the VOX approach.There’s one other factor in this discus-sion- remember the ‘CD volume” adjust-ment? The computer soundcard probablyfinds use for other applications. Remem-bering to turn off the rig power beforeexiting the software application is neces-sary in this case. This ensures that you’renot treating other band users to your fa-vorite music or video game sound effects!

A more secure method in involves the useof one of the RS-232 signals present onyour computer’s Serial port connector.While this increases the number of con-nections between rig and computer, itadds some ‘piece of mind’. The currentlyavailable software packages use eitherDTR (Data Terminal Ready) or RTS (Re-quest to Send) to signal the transceiver todo the T/R switching. When connectedto the PTT line of most transceivers, thesesignals require modest extra circuitry toconvert the (bipolar) RS-232 signal to aPTT-compatible level.

In search of the simple PSK31 rigI was approached by KH6TY regardingmodifying my existing WM-20 trans-ceiver for PSK31 use. With a sample inhand, Howard came up with modifica-

tions to that design to put it on the 20MPSK31 frequencies using crystal controlfor high stability. Best of all, the crystalsare all standard microprocessor types!

When this became a ‘real project’, as op-posed to a marathon e-mail correspon-dence, I re-laid out the WM-series with anumber of changes, largely for reasonsof improved ease of construction. Thisalso provided the opportunity to incor-porate the necessary I/O connectors ontothe PC board itself. These connectorsmate with standard computer interfacecables (3.5mm stereo and serial port/DB-9), available through electronics retailoutlets. Interested readers may requestcopies of that project’s schematic fromme (business-sized SASE for mail inquir-ies, please).

“Onward and, er, Upward ” - A still-simpler PSK31 rig

Last year, I had some remarks about the‘rock-bottom’ approach to getting on theair, in that case with some relativelysimple ideas for 50 MHz. I’d like to con-tinue in that vein, and in the case ofPSK31, there appears to be a verystraightforward way to put a low-partscount transceiver on the air successfully.

The result of this simplified PSK31 trans-ceiver is my design of the PSK-80 trans-ceiver ... or the “Warbler” as the NJQRPguys like to call it!

As I hope you can sense from the tone ofthis material, I’ve really enjoyed gettinginto PSK31. I hope also that some of myenthusiasm is contagious! The PSK31mode really combines the best aspects ofamateur radio and computer technology,and it’s a natural for our QRP specialty.

Further Reading:For more information on PSK31, checkout the following sites:

http://psk31.com

and http://aintel.bi.ehu.es/psk31.html

sized SSB rig - a ‘Yaecomwood’ , high-end Ten-Tec, Index Labs QRP+, etc.,you’re all set. The phase-locked commer-cial rigs provide sufficient frequency sta-bility for PSK31. In the kit domain, theK2 offered by Elecraft (Wayne Burdickand Eric Swartz), with its SSB adapter,fits the bill nicely. I’ll also touch on mywork in this area later in this discussion.

Technical Considerations:

Frequency stability:

As a narrowband mode, PSK31 is prettyintolerant of drift. With the ‘AFC’ fea-ture in PSK31 decoding software en-gaged, it’ll retain good copy on strongsignals for drifts of up to 0.5 Hz/ second.This means, though, that for successfulQRP operation you want good stability.This requirement may be filled either bya synthesized rig or by the use of crystalcontrol. The full band -coverage analog-VFO rigs are likely problematic in thisregard.

I had considered the ‘Huff-and-Puff’ fre-quency-stabilization as an adjunct to my‘stock’ White Mountain SSB series. Al-though this type of scheme is capable ofthe necessary long-term frequency stabil-ity for PSK31, there’s a catch- the ‘Huff-n-Puff’ schemes are prone to ‘limit-cycle’(frequency cycling) behavor as theydither between several pseudo-lockpoints. Although this behavior maysound fine to the ear for SSB and evenCW applications, it exceeds the drift-ratelimit I mentioned earlier. It’s possibleto filter the error voltage very heavily (ineffect making the loop response as slug-gish as cold molasses) but that in turnaffects cold-start lock time. This approachis not for the faint-of-heart homebrewer.The Ten-Tec Scout uses this scheme, andalthough I know it’s being used to receivePSK31, I haven’t heard yet how it stacksup with the more demanding ‘transmit’application.

Linearity:

It’s possible to transmit- and decode- sig-nals which have been amplified and dis-torted to the point of being 100% duty-cycle transmissions. Besides beingtougher on your finals, it’s not in accor-dance with good amateur practice. In thisspecial case, the signal becomes FSK andsprouts considerable extra sideband con-tent. One of the very nice features of thepanoramic display environment is that alltransmitted signals are on-screen for allto see. Set your SSB drive level too highand you may rest assured that you’ll hearabout it!- overmodulation is easily seenon the waterfall display. As a rule ofthumb, the 3rd-order intermod productsshould be down at least 27 dB, this isroughly comparable to a decent SSBsignal’s distortion figure.

If you’re building a rig for PSK31 fromscratch, several other considerations comeinto play-

At least in the software implementationsI’ve seen, the sound card settings are notaccessible and re-configurable betweenthe ‘transmit’ and ‘receive’ intervals.This means that the receiver output, forinstance, is always being applied as asound card input (along with the desiredtwo-tone output) during transmit. Thisraises the specter of unwanted feedbackpaths, and indeed, without the proper caretaken, the transmitted waveform ends updistorted by the unwanted receiver-wrap-around path. Similarly, unless the trans-mitter chain is effectively disabled dur-ing receive, the unwanted modulation willbe clearly audible in the receiver, givingrise to the same sort of distortions. Inboth cases, effective muting of the trans-mitter and receiver paths eliminates thesefeedback paths.

Transmit-Receive SwitchingA means of switching a rig between trans-mit and receive is needed, and several

9VZ

�1��9�

�

%

��

��

��.

��

��

$17�

7KH :DUEOHU

��0 36.�� 7UDQVFHLYHU'� %HQVRQ� 11�* ��

�

�

��

�

9�

��9�.

��

�� 0K]

��

%

&��9

$) 2XW

$

��9

�

��

�

�

576 ��

*1' ��

9�

��.

� X)

$

��.

9�

9VZ

��

��

��

9�

�

'

�/0��

4�

4�

/�7�7�

8�6$�� 8�

4��1��

8��$/0�� 8��%

�

�

�9�

�� X)

4��1��

/�� X+

��

�6&��&

1RWHV�

�� $OO GLRGHV �1�� XQOHVV RWKHUZLVH QRWHG�� $OO FU\VWDOV �� +&��8� VHULHV�UHVRQDQFH

4�

��

$

��9

� �

�

9�

��9

�1��

� X) �1��

�� 0+]

��.

�

�

��

��.

��.

� X)

��

��9

� �

&

��

��

��

��

4�

4�

8�6$��

4�

��

��

��.

��.

�� X+

�� S)

��1��

��.

�6&��&

��

��.

��.

��.

��.

��.

9VZ

��

��

��

� X)

�� X)

��.

��.

��.

�� S)

��

�

�

��

��

��

��

��.

9�

��.

��

��

� �

9LQ

��

��&��

&� &�5�

5�

5�

5�

5�

5� 5�

5�

5�

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��-�

-�

&��

��

&��

&�

&�

&�

&�

&�

&�&�

&��

&��

&��

/�

&��

&�� &��

-�

<� <�

'�

<�

&��

'� '�

&��

&��

&�� &��

&��

&��

&��

&��

&�� &��

&��

&��

&��

&��

&��

'�

<� <� <�

'��1��

�-��

-�

&��&��

&��

'�

5��

5��

5��

5��

5��

5��

5��

5��

�1��4� ��.

��.

5��$ �

9VZ

�1��9�

�

%

��

��

��.

��

��

$17�

7KH :DUEOHU

��0 36.�� 7UDQVFHLYHU'� %HQVRQ� 11�* ��

�

�

��

�

9�

��9�.

��

�� 0K]

��

%

&��9

$) 2XW

$

��9

�

��

�

�

576 ��

*1' ��

9�

��.

� X)

$

��.

9�

9VZ

��

��

��

9�

�

'

�/0��

4�

4�

/�7�7�

8�6$�� 8�

4��1��

8��$/0�� 8��%

�

�

�9�

�� X)

4��1��

/�� X+

��

�6&��&

1RWHV�

�� $OO GLRGHV �1�� XQOHVV RWKHUZLVH QRWHG�� $OO FU\VWDOV �� +&��8� VHULHV�UHVRQDQFH

4�

��

$

��9

� �

�

9�

��9

�1��

� X) �1��

�� 0+]

��.

�

�

��

��.

��.

� X)

��

��9

� �

&

��

��

��

��

4�

4�

8�6$��

4�

��

��

��.

��.

�� X+

�� S)

��

�1��

��.

�6&��&

��

��.

��.

��.

��.

��.

9VZ

��

��

��

� X)

�� X)

��.

��.

��.

�� S)

��

�

�

��

��

��

��

��.

9�

��.

��

��

� �

9LQ

��

��&��

&� &�5�

5�

5�

5�

5�

5� 5�

5�

5�

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��

5��-�

-�

&��

��

&��

&�

&�

&�

&�

&�

&�&�

&��

&��

&��

/�

&��

&�� &��

-�

<� <�

'�

<�

&��

'� '�

&��

&��

&�� &��

&��

&��

&��

&��

&�� &��

&��

&��

&��

&��

&��

'�

<� <� <�

'��1��

�-��

-�

&��&��

&��

'�

5��

5��

5��

5��

5��

5��

5��

5��

�1��4� ��.

��.

5��$ �

��9

��9

��9

��9

��9

��9

��9

��9��9

��9

��9

��9

��9

��P9

$& '00

��P9

$& '00

7<3,&$/ 92/7$*(6 $5( 6+2:1 ,1 %2/'

$6 0($685(' :,7+ 5) 352%( 86,1*

08/7,0(7(5 ,1 '& 92/76 326,7,21�

�(;&(37 )25 0,//,92/7 5($',1*6:+,&+ $5( 0($685(' ,1 $& 92/76326,7,21 :,7+287 7+( 5) 352%(��

��9 �5;�

��9 �7;�

�$FWXDO UHDGLQJV

PD\ YDU\ VOLJKWO\�

�9 �5;�

��9 �7;�

�$FWXDO UHDGLQJV


DSSUR[ ��9 �5;�

DSSUR[ ��9 �7;�

�$FWXDO UHDGLQJV


�9 �5;�

��9 �7;�

-�

%1&0,&5�

/,1(

287

-�

3:5�

-�

� �

-�-�

7�

&�

7�

/�

&��

&��

8�

5��

8�

'�'�

��1��

5��

/�

5��

4�

5��

5��

&��

�1��

<�<�<�

&�� &�� &��

��

��

��

��

��.

��.

&��

��

��

&��

��X+

��

��

&��

�� X)

4�

4�

�

5��$

�

5��

4�

4�

5�

5��

&��

��

��

5��.

� �

'�

&��

6$��$

6$��$

��

��

��

&��

&��

8�

&��

5��5��

��.

��.

<� <�

��

��

��

&� &� &�

8�

��

��

&��

&��'�

5��.

��

5��

��

&��

&��

��

��

4�

5��

5��

��.��.

&��

5��

&�� X)

��

��.

�1

�� X)&��

��

&�� X)4�

�1

��

��.

��

� X)&�

&��

4��1

��

'� ��9

5�

5�

&�

��.

��

�� &�

<�5�

&��

��

&��

5��

5��

��.

��.��

��.

/0��

��

5��

5��

��.

��.

� �

��.

��

��.��.

&��

� X)

5�

&�

5�� 5�

��

&��

11�* ��

4�

5��

5�� '�

��

��

��.

��. ��.

/�

&��

��

��

X+

5��

5��

&��

Date post:	19-Mar-2020
Category:	Documents
Upload:	others
View:	5 times
Download:	0 times

The PSK-80 ﬁWarblerﬂ - njqrp.clubnjqrp.club/warbler/warbler tech manual - rev c.pdf · 6 PSK-80...

Documents