30 Model Engineers’ Workshop Rationale Readers, who have followed my articles on aspects of CNC, will be aware that I have a Wabeco CNC mill in my workshop. These readers will now be wondering why I am retrofitting a Chinese mill when I already have a perfectly satisfactory CNC machine. There are a number of reasons why I am embarking on the retrofit. First and foremost I like to keep myself busy. I generally have a project on the go most of the time particularly in the winter when I cannot go fishing. Secondly there are a number of ideas I want to try out with the retrofit. Thirdly model engineers are showing an ever increasing interest in CNC and would like to retrofit an existing mill but are unsure of how to begin what may seem a daunting project. Though this series will be principally concerned with the X3 mill, the underlying principles should be applicable, with modification, to the retrofit of any mill. A CNC mill is only as good as the screws fitted to the machine. For my money, the only satisfactory screws for the long term are ball screws. These may have a certain (very small) amount of backlash but this is not a problem as the backlash with ball screws is constant along the entire length of the screw and should remain constant for a very long time. Acme screws are very satisfactory for hand operation where backlash is not a serious problem. Unfortunately, uniform backlash with new Acme screws rapidly becomes non-uniform, because one tends to use the middle section of the screw more than either end. After this the screw is of little use for accurate CNC work. When I stripped the X3 down I was impressed with the quality of the screws. It seemed a pity to scrap them but that is what happened. Having made this decision, the next question was finding suitable ball screws to replace them. The ball screws that are currently available are either precision ground screws costing an arm and a leg or the more affordable rolled screw. The threaded section of both the X and Z screws on the X3 are 500 mm long. I wanted 12 mm diameter by 2 mm pitch screws as I intended attaching the stepper motors directly to the X and Y screws. An arrangement having 2 mm pitch screws and using Gecko 201 micro- stepping drives, provides 10 micro steps per single motor step and thus gives a resolution of 1000 steps per millimetre. There was a further reason for opting for direct drive. The design of the Y-axis of the X3 meant that if I had opted for a timing belt drive I would have had to replace the front screw bearing. Almost all available ball screws are right hand screws. All milling machines are fitted with left hand Acme screws on the Y- axis, thus a consequence of replacing the Y-axis screw with a ball screw is that this axis becomes opposite handed to the X- axis. For CNC operation this is not a problem, it is simply a matter of changing the sense of rotation of the stepper motor. If the mill will be used as a manual machine this opposite handedness of the X and Y axes is a bit of a bind. To overcome this problem I, together with my friend Tony Jeffree, have designed a “hand operation” unit that drives through the stepper motors. The unit is similar to units found on industrial CNC machines and consists of a hand held box (containing some electronic bits) which connects to the drive unit for the stepper motors. The box is fitted with a single 20 position hand wheel, a range selector switch and an axis selector switch. Rotating the hand wheel clockwise moves the X and Y axes in a positive direction (i.e. with X selected the table moves to the right) and the head down (Z negative). Rotating the hand wheel anti-clockwise reverses the direction. Using the range selector a single rotation of the hand wheel moves the X and Y axes 2 mm or 0.2 mm. Each hand wheel position thus corresponds to a movement of 0.10 mm or 0.01 mm. Full Dick Stephen describes how he added CNC control to his machine. RETROFITTING THE X3 MILLING MACHINE (1) 1. Rags placed under head will avoid damage.
Transcript
30 Model Engineers’ Workshop
RationaleReaders, who have followed my articles onaspects of CNC, will be aware that I have aWabeco CNC mill in my workshop. Thesereaders will now be wondering why I amretrofitting a Chinese mill when I alreadyhave a perfectly satisfactory CNC machine.There are a number of reasons why I amembarking on the retrofit. First andforemost I like to keep myself busy. Igenerally have a project on the go most ofthe time particularly in the winter when Icannot go fishing. Secondly there are anumber of ideas I want to try out with theretrofit. Thirdly model engineers areshowing an ever increasing interest in CNCand would like to retrofit an existing millbut are unsure of how to begin what mayseem a daunting project. Though this
series will be principally concerned withthe X3 mill, the underlying principlesshould be applicable, with modification, tothe retrofit of any mill.
A CNC mill is only as good as thescrews fitted to the machine. For mymoney, the only satisfactory screws for thelong term are ball screws. These may havea certain (very small) amount of backlashbut this is not a problem as the backlashwith ball screws is constant along theentire length of the screw and shouldremain constant for a very long time.Acme screws are very satisfactory for handoperation where backlash is not a seriousproblem. Unfortunately, uniform backlashwith new Acme screws rapidly becomesnon-uniform, because one tends to use themiddle section of the screw more thaneither end. After this the screw is of little
use for accurate CNC work. When Istripped the X3 down I was impressedwith the quality of the screws. It seemed apity to scrap them but that is whathappened. Having made this decision, thenext question was finding suitable ballscrews to replace them. The ball screwsthat are currently available are eitherprecision ground screws costing an armand a leg or the more affordable rolledscrew. The threaded section of both the Xand Z screws on the X3 are 500 mm long.I wanted 12 mm diameter by 2 mm pitchscrews as I intended attaching thestepper motors directly to the X and Yscrews. An arrangement having 2 mmpitch screws and using Gecko 201 micro-stepping drives, provides 10 micro stepsper single motor step and thus gives aresolution of 1000 steps per millimetre.There was a further reason for opting fordirect drive. The design of the Y-axis ofthe X3 meant that if I had opted for atiming belt drive I would have had toreplace the front screw bearing.
Almost all available ball screws are righthand screws. All milling machines arefitted with left hand Acme screws on the Y-axis, thus a consequence of replacing theY-axis screw with a ball screw is that thisaxis becomes opposite handed to the X-axis. For CNC operation this is not aproblem, it is simply a matter of changingthe sense of rotation of the stepper motor.If the mill will be used as a manualmachine this opposite handedness of theX and Y axes is a bit of a bind. Toovercome this problem I, together with myfriend Tony Jeffree, have designed a “handoperation” unit that drives through thestepper motors. The unit is similar to unitsfound on industrial CNC machines andconsists of a hand held box (containingsome electronic bits) which connects tothe drive unit for the stepper motors. Thebox is fitted with a single 20 position handwheel, a range selector switch and an axisselector switch. Rotating the hand wheelclockwise moves the X and Y axes in apositive direction (i.e. with X selected thetable moves to the right) and the headdown (Z negative). Rotating the handwheel anti-clockwise reverses thedirection. Using the range selector a singlerotation of the hand wheel moves the Xand Y axes 2 mm or 0.2 mm. Each handwheel position thus corresponds to amovement of 0.10 mm or 0.01 mm. Full
Dick Stephen describes how headded CNC control to hismachine.
instructions for constructing this unit willbe given at the end of this series. Readersnot wishing to construct the unit can stilloperate the machine by hand by fittingdouble ended shaft stepper motors.
Selecting the ballscrews The ball screws are the first item thatneeds to be selected. The pitch of thescrews will determine the size of steppermotor required. I searched the Web forsuitable screws with little luck. There werelots of screws available; unfortunately theones available were either prohibitivelyexpensive or the wrong length. The onlymanufacturer that was able to supply 12mm diameter 2 mm pitch screw was THK(MBF 1202 nut 500LC7 rolled screw). The
maximum length of screw rod they wereable to supply was fortunately 500 mm,exactly what I needed. When I fitted ballscrews to my Wabeco I used THK screwsand found them absolutely excellent. Thestandard C7 grade rolled 12 mm 2 mmpitch ball screw has a maximum error of50 microns for a 300 mm travel. THK havea new design of ball nut for this screw,which features replaceable nylon wipersconforming to the profile of the screwthread to prevent the ingress of all butthe very finest dust. The nuts are alsoself- lubricating. The only lubricantrequired is a light oil, applied to thescrew. If any reader is intending topurchase these or any equivalent ballscrews he should not pay more than£145.00 for a single screw and a nut. Thismay seem expensive, in fact it isn’t asthese screws will last a lifetime.
Selecting thestepper motorsHaving decided on the screws for themachine the stepper motors can now beselected. Many model engineers I havespoken to on the subject of stepper motorsfor CNC are of the opinion that one shouldfit motors as powerful as possible. Theconsequence of this may be that themotors, and hence the associated driversand power supply are all larger (and moreexpensive) than necessary. A completeapproach to selecting motors, includingaccounting for resonance effects is reallyoutwith the scope of this article, but I willdescribe in detail the steps I followed inselecting the motors for the X3 retrofit. Ishould note that as my work generallyuses small cutters and fine feeds, I havesimplified things somewhat and calculatedthe torque required firstly to move theinertial mass of the table and secondly toovercome the frictional forces generatedat the slideways. As my work invariablyuses small cutters, I chose to ignorecutting forces.
To understand how to carry out thiscalculation consider what happens whenthe stepper motor is initially turned on.The torque generated is used to acceleratethe mass of the table from rest to the finalvelocity. The amount of torque needed todo this will depend on the mass of thetable, the magnitude of the final velocityand the time taken to reach this finalvelocity (determined by the acceleration ofthe table). In addition torque is alsorequired to overcome friction in theslideways. The initial friction (“stiction”) issignificantly larger than the friction oncethe table has begun to move. It is thisinitial friction that is important whenconsidering the size of motor required.
The mass of the table needs to beestimated first. You can either put the tableon the bathroom scales or work it out fromthe approximate dimensions of the tableassembly. The estimated mass should erron the high side.
Table dimensions 55 cm x 16 cm x 4 cm
Dovetail assembly 20 cm x 20 cm x 3 cm
Volume of the table = 3520 cm3
Materials required for the complete retrofit
1. Ball screws. (3), Ref. Nut MBF1202; Rolled screw rod 12 mm 2 mm pitch 500 mmlong. THK UK Ltd Tel 01908 222159
2. Stepper Motors. (2), 23HSX-206 for X and Y axes. (1) 23HSX-306 for the Z axis.McLennan Servo Supplies Ltd. Tel. 08707 700700
3. Oldham Couplers (4), 25 mm 8 mm bore (2 for each screw) for attaching the steppermotors to the X and Y screws. I got mine from RS-Components stock number 359-6434 A packet of Nylon torque discs is also required stock number 359-7579.
4. Timing pulleys for the Z-axis drive 15 and 30 tooth 5 mm pitch 12 mm wide and a295 mm long 5 mm pitch 12 mm wide timing belt. HPC Gears Tel 01246 268080.
5. Enclosure for re-housing the electronic motor control unit. I used the following onefrom RS-Components, Stock number 129-628.
6. EN1A mild steel round rod. 300 mm of 20 mm; 1 m of 15 mm; 500 mm of 12 mm;300 mm of 10 mm.
7. Brass bar. 1 piece each 70 mm x 38 mm x 22 mm, 100 mm x 32 mm diameter.8. Aluminium. 1 piece 300 mm x 80 mm x 12 mm. 3 pieces 200 mm x 80 mm x 8 mm.
2 pieces 250 mmx 200 mm x 1.6 mm. 1 piece 300 mm x 25 mm x 12 mm.9. Loctite 326 Structural adhesive. One 50ml bottle. Somewhat expensive (£21) but
very useful around the workshop. Produces a very strong joint nearly as strong assoft solder but a lot more convenient. A useful tip. NEVER apply any Loctiteadhesive (particularly 603 or 326) straight from the bottle to the work piece. Alwaystransfer the Loctite from the bottle on to a piece of plastic sheet (I use a scrap ofPerspex) then use a new toothpick to apply the Loctite to the work.
10. Ball races 3 off 6 mm i.d. 11. Assorted Hex head screws. 3 mm, 4 mm, 5 mm and 6 mm12. Screened cable 6 metres 5amp 4core for connecting the stepper motors to the drive
unit.13. 5 Amp screw terminal block. Several sections.14. 3 mm solder tags.15. Gas Spring (replacement). The standard gas spring fitted to the Z-axis to support the
weight of the head stock is non adjustable and does actually over compensate forthe weight. The actual weight of the head stock is about 30Kgm. The compliance ofthe correct gas spring should be just enough to support the weight. I replaced minewith an adjustable one. These are widely available. RS-Components supply suitablesprings (stock number 686991), unfortunately only in packs of 2 units. I got minefrom RS, fortunately I was able to pass the extra one on to a friend. For a suppliercontact the manufacturer Arvin Motion Control Tel. 0116 274 3600.
2. Scraper, flat, and engineer’sblue used for slidewayimprovement.
Mass of dovetail assembly = 1200 x 7.8 = 9.36 Kilogram
Total mass M of the table assembly = 36.8 Kilogram
The effective inertial load, J, seen by themotor is given by the following formula: -
JLoad = MTable p2/(2π)2
Where p is the pitch of the lead screw incm.
The motor for the Y-axis is required tomove both the table and the dovetailassembly. If a suitable motor is selectedfor the Y-axis the same motor will be morethan adequate for the X-axis as this motoris only required to move the table. Thescrew chosen for the X3 has a 0.20 cmpitch (2 mm). Substituting the value of thetable assembly the effective inertial load Jis
J = 36.8 x (.2)2/(2x π)2
= 0.037 Kgm cm2
To proceed further with the calculationthe maximum velocity of the table needsto be considered. For most practical modelengineering CNC work a maximum tablespeed of 15 mm per second is adequate.For industrial production work where timeis money far higher speeds are frequentlyused. To calculate the required torque themaximum velocity needs to be expressedin the number of rotations of the screw persecond (or radians per sec). To have alinear table speed of 15 mm per sec the 2mm pitch screw need to rotate 7.5 timesper sec. The maximum angular velocity, ωis then equal to
ω = 7.5 x 2π radians/sec= 47.12 radians/sec
Finally the acceleration time to achievethis velocity has to be found. Readers willbe aware that I use and very much like theCNC software DeskCNC. One facility of themachine set up procedure is that it enablestable velocity and acceleration to becorrectly linked.
Earlier I said that the stepper motorswould drive the lead screws directly. Witha 2 mm pitch and a x 10 micro step drivethe calibration factor for the X and Yscrews is 1000 steps/mm. A 20 mm/secvelocity then equates to a step rate of20,000 steps/sec. The Desk CNC softwareincludes a machine set up table whichallows the velocity and acceleration to becorrectly linked with an appropriateacceleration time, t, in this case 0.017 sec(17 milliseconds).
The motor torque required to this tablemovement is given by the formula:-
T = J.ω/10tJ = 0.037 K gm cm2
ω = 47.12 radians/sect = 17 milliseconds
Inserting these values the requiredtorque is
T = 0.01 Newton metre= 1 N cm
This result shows that the torquerequired to accelerate the table mass isvery small and essentially can be ignored.There is a rather simpler way of assessingwhether a stepper motor is suitable that isto compare the effective inertia calculatedabove with the value of the rotor inertia ofthe proposed motor. Suitable motorsshould satisfy the condition:-
JTable< Rotor inertia
Table 1 lists the mechanicalspecifications of a range of 23 frame sizeMAE stepper motors supplied byMcLennans. The last column in the tablelists the rotor inertia of the motors. Thecondition is satisfied for all the motorslisted.
Slideway FrictionIn addition to accelerating the mass of thetable assembly the stepper motors have toovercome the slideway friction. There is nomethod of reliably calculating themagnitude of the slideway friction. Theonly simple way to obtain a value for thefriction is to actually measure it. I used aspring balance to measure the slidewayfriction on the X3. There are twocomponents to the friction, the initialstarting value as the table accelerates fromrest and the value during motion atconstant velocity. The initial friction issignificantly larger than the friction atconstant velocity. For the table assembly(both X and Y) of the X3 the initial frictionforce was approximately 10 K gmdropping to about 5 K gm once the tablecommenced motion. When selecting thestepper motor it is the initial friction that isimportant. The frictional load seen by themotor is reduced by the mechanicaladvantage of the lead screws. A furtherfactor that determines the frictional loadseen by the motor is the efficiency of thescrew. For Acme screws the efficiency is ofthe order of 35%, while for ball screws theefficiency is in excess of 85%. This isanother good reason for fitting ball screws.The motor torque required to overcomethe friction is given by the formula : -
Torque = (Friction force) x p/2πe
Where p is the lead screw pitch and ethe screw efficiency. For the X3 tableassembly the required torque is equal to : -
T = 10 x 0.2/0.85 x 2π= 0.37 Newton cm
Referring to Table 1 all of the motorswould be adequate for the mill. In makingthe final choice you should remember thatthe torque estimated above does not takeinto account the forces generated duringmachining or the additional weight of avice. Anyone wishing to evaluate cuttingforce effects might perhaps fit a slavepulley in place of a handwheel and use aspring balance and string method. I settledfor motors with a bi-polar holding torqueof 163 Newton cm for all three axes. Thisis possibly a bit over the top as I probablycould have got away with motors with alesser holding torque. I have to admit I likea very good margin of safety in matterssuch as this.
Dismantling themachineIf you have purchased the cabinet standundo the bolts that attach the machine tothe stand. If possible place the mill in themiddle of your workshop. You will need tobe able to work all around the machine. Inaddition you will need to get access to thetwo hex head bolts that secure the Y-axisnut, from underneath the base. This accessis possible if the base is allowed tooverhang the cabinet front by about 20cm.There is no danger of the mill and standtipping over as the centre of gravity of themill is close to the column when themachine is stripped down.
Begin dismantling the mill. Do this in avery systematic way and lay each part as
you remove it in an orderly waysomewhere that they can be leftundisturbed for however long it takes youto complete the project. Remove first theentire table assembly. The head stock canbe removed next. Completely assembledthe head stock is too heavy for a singleperson to easily lift. To make the removalmanageable, the head needs to be partiallystripped down. Remove the motor, accessto the screws that secure the motor areunder the cover over the motor pulley. Themotor will still be attached to the machineby its power cable. To completely removethe motor, the entire electrical assemblyfor the machine located behind the columnwill also have to be removed. To do thisundo the four screws that attach thecolumn cover to the machine. The entireelectrical assembly will need to beremoved later and fitted into a separateenclosure. Remove the cast iron cover ontop of the head. This will allow you toremove the motor mounting bracket.Finally remove the handle for lowering thespindle. Don’t attempt to remove thespindle; re-tensioning the return spring isdifficult! Lower the head as far as it willgo. Pack the space between the head andthe base of the machine with old rags sothat when the head is loosened and comesoff it cannot drop and cause damage - SeePhoto 1. The dowels that locate the headare tapped for 4 mm screws. You may beable to insert a screw and pull them out. Ifnot you will have to make an extractor.This is no more than a 20 mm length of 12mm mild steel rod drilled 8 mm to a depthof 15 mm and then with a 4 mm drill forthe full length. Thread a nut on a 30 mmlong 4 mm screw. Place the extractor overthe dowel and screw the 4 mm screw intothe dowel end. Tightening the nut will easeout the dowel.
With the motor and the electroniccontrol unit removed the Z-axis assemblycan now be completely removed. The gasspring must be removed first. To do thisthe wind the Z-axis nut up as far as it willgo. In this position the gas spring exertsno force. The spring is held in position bytwo 8 mm pegs one screwed into the nuthousing the other at the bottom of thebase of the mill. With the spring removedthe remainder of the assembly comesapart easily.
The Y-axis nut will still be attached tothe base. Slide the base forward tooverhang the stand by about 20 cm. Thiswill give access to the two 5 mm hexscrews that hold the nut in place.
Correcting minorimperfectionsIt is worth spending some time correctingsome of the minor imperfections you willdoubtless have noticed as you dismantledthe machine. It is really worth getting theslides as good as possible as this reducesslideway friction. The X-axis slideways onmy mill were very good, but I felt that theZ-axis slide was not quite perfect. Theseating for the Y-axis nut was adequate foran Acme nut but not for the ball nut I used.I decided that the rear surface of the headcasting where it is bolted onto the Z-axisslide could be improved. The gib stripsseemed to be fine but nevertheless Iscraped then perfectly flat and got the
surface really smooth. The difference wasamazing, the slides felt so much better.
For this rework, my chosen method isto use a scraper, a test flat and a tin ofengineers blue. I was able to borrow acarbide tipped slideway scraper illustratedin Photo 2. A very adequate alternativecan be made out of an old 20 mm wideflat file. Grind off all the teeth for about 30mm from the end. Make sure that oneface is absolutely flat. Grind a curve (seePhoto 3) on the end (radius ~ 50 mm)square to the flat face. Heat the end to acherry red and plunge into cold waterwith a layer of oil on top. Re- grind theend to sharpen and stone off any burrs ona fine oil stone. Next you will need tomake a flat or borrow one. The flat needsto be about 300 mm long 25 mm wideand about 10 mm thick. The easiestmaterial to make the flat out of is fine castiron, mild steel will also do. Photo 2shows the one I was able to borrow.Begin by cutting a 45 deg. bevel one edgeand milling one face flat. You can arguethe relative merits of fly cutters v endmills; I used a fly cutter using a very slowfeed and to as good a finish as possible.You may want to do this before youdismantle the mill! The face of the flatneeds to be very smooth and perfectlyflat. Remove all machining marks usingwet and dry paper on a FLAT surface. Thesurface of the flat needs to be checked forirregularities. To do this a reference flat isneeded. I used as a reference flat, for
truing up the gib strips (see Photo 4), thesurface of the mill table.
All the slideway surfaces need to bechecked for high spots. To do this smeara small amount of the engineers blueonto the surface of the flat. Place the flaton the surface being tested and rub theflat against the surface. Any high spotwill be marked with blue. Carefullyscrape the blue away, removing a smallamount of the underlying metal at thesame time. New cast iron scrapes veryeasily. If you are unsure about scrapingand have a piece of cast iron handyspend a few minutes practising scrapinguntil you have mastered the art. Re-applythe flat and scrape again until the surfaceis quite flat.
The surfaces that require very carefulchecking are the seatings for the X and Ynuts. As we will be using ballscrews, theseseatings need to be both flat and parallelto the slideways. The subsequent lining upof the screws depends to a large extent onthese seating surfaces. On my machinethe X seating was perfect, the Y not quiteso. Photo 5 illustrates how I checked the Yseating. I attached my dial indicator to alength of 16 mm square ground bar.Moving the dial indicator over the surfaceof the seating located the high spots.These were marked with blue a felt tippedpen and then scraped.
The next section will cover thereassembly process, fitting the ball screwsand stepper motors.
Motor type length Shaft No. of Mass Uni-polar Bi-polar Rotordia. leads holding Holding inertia
