Warco Major CNC build log

[Robin Hewitt]

When Garry at Zapp finishes machining my ball screws I'm going to have another go at CNC'ing my poor old Warco Major mill.

The plan is to have all the milling done for the X and Y axes so I can strip the machine back then do all the final fitting using a lathe.

The story so far is the pile of bits in the piccy, only a couple of parts left to mill before I'm ready, wonder what I've forgotten.

The theory is simple, combine Garry's 20um backlash screws with a 5um half step resolution and see what it does :beer:

[Robin Hewitt]

Hope Gary gets me my screws soon, I'm running out of bits to make.

Had a bit of a fright when I located the bolt holes that hold the Y handle to the machine. They look like they were drilled at random with a Black and Decker. One is far from straight, had to drill my hole far from straight to match.

Need to buy some grub screws, best selection and price seems to be from J & S Fasteners, they offer a choice of tips so I can get a flat point rather than those horrid cup ends. I've used them before, have a link...
http://shop1.actinicexpress.co.uk/shops/Jandsfasteners/index.php

More when the screws arrive :beer:

[Robin Hewitt]

4 days off, a good time to start :beer:
First the Y screw. On the Warco the Y screw locates in to a 21mm diameter pocket 6mm deep cut in the bottom of the carriage (First pic).

I had to finish the nut holder after the machine was apart so I left a rod projecting out the other side for chucking it in the lathe. Looking to mininmise backlash I cut an interference fit with a groove around the bottom to catch any crut.

Pic 2 shows the Y nut fitted but not screwed down yet.

Pic 3 shows the carriage fitted with the spindle sticking out. The first of 2 angular contact bearings is fitted.

The idea is to hold the shaft between two angular races with a quarter ton of thrust locking it in place. To get a quarter ton I'm using pairs of belleville washers.

I have to crush the washers by 1mm and it would be nice if the nut ended up flush with the end of the shaft so I measure it, Pic 4, at 2.95mm then remove 1.95mm from the collar in the lathe.

The fitting has to do both push and pull on the ball screw so it's one solid lump of T6 alloy with a plate for the stepper screwed to it.

[Robin Hewitt]

Next the X axis. Warco bolt the X nut on from below which is a pain because the Y screw is very much in the way. So I'm bolting it from above.

My X nut holder has a lip to hold it square, luckily I can use the mill to cut a square edge in the carriage for it to run against. Pic 1

Now some hard work. The ball nut fitting is slightly larger than the cutout in the end of the table. I know I could jiggle it over the top but it would be a pain forever more, so now I'm setting out for a jolly afternoon with files and angle grinder opening it out. Pic 2 :rolleyes:

More later

Robin

[BillTodd]

What's the extra spigot poking up/down from the nut mount for?

http://www.mycncuk.com/forums/attachment.php?attachmentid=344&d=1239369960

[Robin Hewitt]

What's the extra spigot poking up/down from the nut mount for?

Something to hold on to when I turned the other end :beer:
After much angle grinding and filing the bed slid comfortably past the nut housing so I put in one bolt to fix it to the carriage. First pic. Not very brave but it's easier to move one misplaced hole than 4. It isn't defying gravity, there's a roller stand supporting it, makes it real easy to put on and take off.

Then the driven end, Pic 2, it sticks out a lot so I can use more of the bed, got this very wrong when I first CNC'd it so this is a slight overkill. This end is in 3 pieces, which is okay because the screw will hold it in tension with the same quarter ton I put on the Y. I plan to Loctite it in place when I drag it all together with the Bellevilles.

Then the handle end...

I haven't cut the bearing seat in the handle end yet, wanted to get my calipers on it first, Pic 3, and find out exactly how far the shaft sticks out.

Next I put that measurement in the drawing and decide how much I trim off and where. I have leeway.

The last pic was supposed to show that my single bolt went in the right place and the shaft is concentric. Would have looked a lot better if I'd put the camera a couple of inches to the left :rolleyes:

The Y axis feels amazingly precise from a backlash point of view. The dial test gauge will reveal all.

More soon...

Robin

[BillTodd]

Looking good Robin :)

I'm Looking forward to the Z axis; it's the one I've yet to figure out for my drill/mill.

[Robin Hewitt]

X handle fitted and screw tensioned, Pic 1

An overall pic 2

I'm going to test this out before I redo the Z, see how Gary's screws perform before I buy any more.

I have to take it all apart tomorrow, add the other 3 bolts to the X nut housing, lube the slides, add some Loctite so things can't twist etc.

2 problems

I'm getting a collision at either end of the X travel so I'm not quite getting quite as much as I was looking for, need to go back to the drawings for that one,

The old pulley on the X motor doesn't want to come off. Easy to shatter the magnet in an Escap Discmagnet motor if you end load the shaft, I'll have to contrive some kind of puller.
Still, not bad for one day :beer:
Incidentally, I have a new 16mm drill chuk on there, 100% better than the Warco supplied one, got it from RDG tools for £22

Robin

[irving2008]

Incidentally, I have a new 16mm drill chuk on there, 100% better than the Warco supplied one, got it from RDG tools for £22

RobinLooking good Robin - I wonder how similar my Minor is? BTW, I have the same chuck, is yours on an MT3 or R8 taper?

[Robin Hewitt]

Hi Irving it's Morse :D

I got the lack of travel nearly sorted today by angle grinder. Tested it with a loose Gibb, lost about 1cm with a tight Gibb. Can only be a few thou left to remove think I'll leave that until I need it an extra cm. Pics attached.

THE BIG BACKLASH TEST (this being the whole point of the exercise)

Put a DTI on the bed and turned the handle, by hand. I could feel the stepper motor cogging and with the 2.5:1 reduction the slide should move .01mm per clunk, 0.0004".

Well it didn't move that far when I sat there rocking it back and forth one clunk, but it did move a little bit so backlash is now <10 microns, not bad for 20 micron nuts :beer:

I totally destroyed one of the stepper motors while trying to remove the old pulley. I was thinking to replace them anyway with some of these new fangled hybrids.

Robin

[Robin Hewitt]

I have just sent the picture for the Z axis screw machining to Gary and ordered a girt lump of aluminium.

The spindle doesn't like running at 5000 rpm which is top whack, 2500 rpm of the pulleys and 100Hz from the inverter, the bearings run hot. I would actually like twice that so I can run tiny ball end mills.

The plan is a false spindle set inside a 3MT soft end arbor and a lay shaft with two timing belts. That way the false spindle with it's ER11 collet can turn at 9 times the main spindle speed.
Give me something to do while I wait for the Z screw :beer:

[Robin Hewitt]

I tried a micro grinder in a 20mm collet...it works well at 60.000RPM

Sounds tempting, I'm having a slight alignment problems with the double axial race on my tiny collet gizmo. Going to have to bolt a boring bar to the mill bed, cut the seat oversize in situ then fit a sleeve.

To do that I need to make a 1/2" BSW drawbar. Can't make the drawbar because my 1 1/2" die holder has gone walkabout. Can't make the Z axis bottom plate 'til I know if I need to fit the layshaft etc. :rolleyes:

However the Z axis screw has arrived and fits inside it's sleeve okay.

More soon

Robin

[irving2008]

To do that I need to make a 1/2" BSW drawbar.
Do what I did and buy a 36" length of 1/2" BSW threaded rod and 4 matching nuts. Cut rod in half, screw into tool (through quill), place 1/2 washer, screw on 1 nut till tool retained, then a second nut and lock together. Remove bar and nuts together as one assembly, weld up the nuts, repeat and hey presto two draw bars! (one for mill, one for lathe as they share the same MT3 taper and tooling).

[Robin Hewitt]

Do what I did and buy a 36" length of 1/2" BSW threaded rod and 4 matching nuts.


Where did you manage to buy archaic studding in an ISO world? Enquiring minds need to know :beer:

[irving2008]

I got mine from Sowerby Bridge (on recommendation from RDG) Archaic.....are you incinerating it's bent :naughty: :D

RDG got it for me, think it was from the same place Kip.

Or here: http://www.thesitebox.com/Product/126423/bsw-studding-self-colour-12-x-3ft.aspx

or here: http://www.fastfixdirect.co.uk/code/navigation.asp?Diameter=1%2F2*&Thread=BSW&Length=all&Filtered=True&PageID=1&MainCategoryID=4&ProductCategoryID=129 (http://www.fastfixdirect.co.uk/code/navigation.asp?Diameter=1%2F2*&Thread=BSW&Length=all&Filtered=True&PageID=1&MainCategoryID=4&ProductCategoryID=129&fType=Fasteners&search=Show+Items)

[irving2008]

Have you seen these on ebay Robin? (http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&ssPageName=STRK:MEWAX:IT&item=350191694622) Looks like a plan :)
Interesting but I wonder about backlash... aircraft positional control doesnt need high spec in terms of backlash and since it has a 'mechanical detach' I suspect the backlash could be problematic

[Gary]

These are for the First class chair control, i helped design one for a company a few years ago.
Backlash will be a problem.
Also they normally have a long lead so it can be back driven if the power fails.

Interesting but I wonder about backlash... aircraft positional control doesnt need high spec in terms of backlash and since it has a 'mechanical detach' I suspect the backlash could be problematic

[Robin Hewitt]

A honking big lump of Al mixed with a tube and a 20mm NSK


Great minds think alike :beer:

The honking plate holds the screw square to the quill with minimum flex, I was thinking two 8mm pinch bolts at the back.

One vital feature missing in your picture though, probably early days and not forgot but still worth mentioning, you have to be able to release it before you hammer the chuck out of it's taper :eek:

I don't have any convenient flat to bolt a plummer type bearing mount to so I have to extend the sides of the quill housing forwards. This is almost convenient because it ain't easy to get a precise measurement for the separation 'twixt spindle and screw, I thought a bit of slack before I drilled the fixing bolts wouldn't be a bad thing.

[Robin Hewitt]

Back again and cutting metal...

Going for the larger stepper motor this time with half the pulley reduction to put me in the power band. 1 rev = 4mm, half step = 10 microns.

Same pair of angular contact races with the Belleville washers holding it in tension as before.

The front of the mill is to extend forwards with bolt on cheeks. The bearings are set in a 16mm plate which fits into grooves in the cheeks to hold it square and rigid.

The ball nut is enclosed a la Kip, the bottom of the housing is wider to take a pair of 8mm set screws which tie it to the spindle, or not, as the case may be.

Next job, fitting the cheeks so they wedge tight in the ironwork :beer:

[Robin Hewitt]

I don't really do 'excited' anymore :yahoo:

I think I have to fit this top section then fudge to connect it to the quill.

Once it's all bolted together and moving I can measure the fudge and get precise dimensions for the honking plate.

Robin

[irving2008]

Robin,

Been watching this with interest as I want to do the same to mine going forward, although not sure my engineering skills are up to the challenge yet.

I am still struggling to understand how the Z-mechanism will work. Does the 'honking plate' clamp to the quill flange under the head and, if so, where does the remainder bolt to on the head itself. What is the advantage of that over motorising the quill handle itself?

On the X-axis, I can understand why the motorised element sticks out so far to allow more X travel, but again, how is that better than driving the handwheel end? And could it be done without it sticking out so far as this would give me space problems?

I know Kip has done something similar.

I've yet to strip it down and draw it all up to work out how to do it, but just starting to do some planning for later on this year.

[Robin Hewitt]

Hi Irving

Let's not forget that trying to screw this kind of accuracy out of a Warco mill/drill puts us deep in to the realm of silk purse and sow's ear. Silly has already been left way behind :joker:

Motorizing the Z axis handle would be much more in keeping with the machine, but you would never push the boundaries that way.

Would you like to have the z-axis I'm ripping out? It's much the same as the one that's going in but a lot less beefy and made without the benefit of cnc.

One more set of pics and the operation should become clear.

Robin

[irving2008]

Hi Irving

Let's not forget that trying to screw this kind of accuracy out of a Warco mill/drill puts us deep in to the realm of silk purse and sow's ear. Silly has already been left way behind :joker:

Motorizing the Z axis handle would be much more in keeping with the machine, but you would never push the boundaries that way.

Would you like to have the z-axis I'm ripping out? It's much the same as the one that's going in but a lot less beefy and made without the benefit of cnc.

One more set of pics and the operation should become clear.

Robin
Robin,

I'd be happy to get a reasonable degree of accuracy, not interested at pusing the boundaries at first. Cost is still an issue, but if i am going to invest a few £100 in ballscrews etc. I need to be comfortable I know what I am doing and I have a reasonable chance of success.

I may well be interested in your old Z-axis... its still likely to be an order of magnitude better than anything I can produce (yet).

[Robin Hewitt]

Robin have you given much thought to the "hammering out ceremony" once the ballsrew is attached?

I'm doing it the same way I did last time, the screw housing bolts to the honking plate. Only difference is I'm using 2 screws this time instead of one. I think it needs a bit extra to lock the screw parallel to the quill.


I may well be interested in your old Z-axis... its still likely to be an order of magnitude better than anything I can produce (yet).

Not exactly pretty but you can have the X and Y screws if it helps They are cluttering up the workshop, probably going in the bin next time I tidy the shelves. If I find a box I could send the whole she-bang standard parcels.

I bought zero :joker: backlash nuts from Marchant Dice for the X and Y, odd things that you had to crush sideways to set. I never got the backlash below 3 thou. Also I didn't get the ends machined, just glued on extension sleeves with Loctite. With the benefit of hindsight, a false economy.

The Z axis was a disappointment, nice screw off ebay but I didn't eliminate flex and it always fell slightly short of the mark. I'm hoping the honking plate will fix that.

Robin

[irving2008]

I'm doing it the same way I did last time, the screw housing bolts to the honking plate. Only difference is I'm using 2 screws this time instead of one. I think it needs a bit extra to lock the screw parallel to the quill.



Not exactly pretty but you can have the X and Y screws if it helps They are cluttering up the workshop, probably going in the bin next time I tidy the shelves. If I find a box I could send the whole she-bang standard parcels.

I bought zero :joker: backlash nuts from Marchant Dice for the X and Y, odd things that you had to crush sideways to set. I never got the backlash below 3 thou. Also I didn't get the ends machined, just glued on extension sleeves with Loctite. With the benefit of hindsight, a false economy.

The Z axis was a disappointment, nice screw off ebay but I didn't eliminate flex and it always fell slightly short of the mark. I'm hoping the honking plate will fix that.

Robin
Robin,
That sounds very generous of you, if it would get me up to speed quicker and allow me to create updated parts later then that would be great. How likely are your Major parts to fit the Minor ?

[Robin Hewitt]

Pic 1: Cheeks put in with a couple of bolts. Needed a bit of extra removing from the left cheek because the casting was far from straight. Heavy application of the angle grinder at the bottom before the ball screw bearing plate would slide home, had to cut out an extra 7.5mm

Pic 2: Ball screw fitted, view from the side. This is as close as you can get the ballscrew to the quill. Hence requirement for honking plate.

Pic 3: Pile of bits removed, Irvings' if he wants 'em :joker:

Should have taken a front view with the ballscrew installed. Maybe next time.

Next I have to relieve the cheeks a tadge so I can fit the motor plate, bit of a poor measurement there.

Then make a false honking plate, cut the pulleys and wire it up so so I can make a real honking plate :beer:

Robin

[irving2008]

Honking plate is the word indeed....I'm going to aim at 63mm thick for my honking plate.

You had 3 thou backlash with the Isel ballnuts? Did you use oil or grease?

Snatch his arm off Irving :DI'm snatching... Robin, I want please... I'll even come and collect if you want me to... (although may be a while before I can as I have broken the rear suspension on the car)

[Robin Hewitt]

Honking plate is the word indeed....I'm going to aim at 63mm thick for my honking plate.

Uh? :naughty:

If your thickness goes beyond the height of the lump you are clenching it on to, surely it will be in the way, one way or t'other.

I'd better go try and find an Irving box.

Robin

[Robin Hewitt]

Does your quill twist....peg in slot or other type of guide?

Peg in slot, slack enough to convince me that I can't hang a spindle motor off the side.

Only got an hour on it last night, wife dragged me down the pub, forced me to drink beer then drove me home. I was in no condition to continue.

I did cut the extra relief in the cheeks so the motor plate can slide home then bored and faced a 40t XL pulley to fit the ball screw shaft.

Took the camera but forgot to put the memory card in it so no pic.

Robin

[Robin Hewitt]

Me again

Put the pulleys on, belt was too short.

Checked all the measurements, I don't usually get a belt wrong.

Eventually found I'd mounted a 44t pulley where a 40t was required :whistle:

RS seem to be taking orders at 7pm for next day. Wonder if I'll have it tomorrow. Perhaps it will it be a 44t pulley in a 40t bag, don't remember buying a 44t, fairly sure it said 40t on the bag :naughty:

Robin

[Robin Hewitt]

40t pulley just arrived. RS must be slacking, I ordered it at least 11 hrs 30 minutes ago :naughty:


You must feel awful so I'll not mention that I intend using a pair of 30mm SC30UU linear bearings to prevent the trig equation needing to be worked out :D

Okay, curiosity aroused, what trig?

Robin

[BillTodd]

It could be pythagsrus not trig....
It's trig whether you use pythag or not ;)

BTW Was Michael Jackson's last hit called 'the floor'?

[John S]

BTW Was Michael Jackson's last hit called 'the floor'?

Dunno but the O2 comeback tour is sure going to be special................

[Robin Hewitt]

11.5 hours....useless!

Strangely true, when I looked in the box I found an empty 44t packet and a brand spaking new 40t. Nice to have a spare I suppose :whistle:



My thought on the swivelly bit at the bottom..if it twists it shortens the effective length of the ballscrew=bad?

I don't think it can twist without bending the screw, I didn't stint on the metal :nope:

I now have the 40t pulley and belt in place with a faux honking plate fudged out of a bit of 2" x 3/4" iron.

It feels good, any shift on the motor moves the handle, moving the handle spins the motor. I would have wired it up but when I looked for a bit of chocolate block connector all I could find was 80A :rolleyes:

Time for some G-code to make the real honking plate :beer:

Robin

[Robin Hewitt]

BTW Was Michael Jackson's last hit called 'the floor'?

I had heard he was a bit off colour but it was still a surprise.

[irving2008]

Strangely true, when I looked in the box I found an empty 44t packet and a brand spaking new 40t. Nice to have a spare I suppose :whistle:




I don't think it can twist without bending the screw, I didn't stint on the metal :nope:

I now have the 40t pulley and belt in place with a faux honking plate fudged out of a bit of 2" x 3/4" iron.

It feels good, any shift on the motor moves the handle, moving the handle spins the motor. I would have wired it up but when I looked for a bit of chocolate block connector all I could find was 80A :rolleyes:

Time for some G-code to make the real honking plate :beer:

RobinIs 'honking plate' a real term or one you made up to describe a 'big lump o' metal'?

I am looking at the pics but still not seeing quite how the quill gets moved... and still moves the motor when you turn the quill handle...

[John S]

It's actually British Standard Honking Plate.

[Robin Hewitt]

I am looking at the pics but still not seeing quite how the quill gets moved... and still moves the motor when you turn the quill handle...

The frictional losses on a ball screw are low. The screw can turn the nut if you have the means to push hard enough.

Spin the motor by pulling on the handle and you can feel if something is binding. Should be smooth :beer:

[irving2008]

The frictional losses on a ball screw are low. The screw can turn the nut if you have the means to push hard enough.

Spin the motor by pulling on the handle and you can feel if something is binding. Should be smooth :beer:Ah, what is the lead on the screw... I hadn't realised it was that big a lead. But what stops the quill dropping under its own weight then when the motor is powered off? Or is there sufficient stickiness in the cogging to stop that?

[Robin Hewitt]

It's 5mm lead, the 32:40 pulley reduction gives 4mm/motor rev

The quill return spring is there but it wouldn't drop under it's own weight anyway. You gottta pull to spin it.

[Robin Hewitt]

Tentatively wired the hybrid Z axis motor up today, a bit of light relief after doing the VAT return.

Only had 2 Amps spare for it to play with but it doesn't seem to mind one jot, even though it is moving the quill at twice my normal speed :beer:

Also rough cut the honking plate on a bandsaw, see pic, don't know if I'll get anything done tomorrow.

Robin

[Robin Hewitt]

Milled out the honking plate, still needs a slit and a couple of holes opening out but nearly there.

Not sure I actually need it, I sank the detail 20mm in and it cut it 20.04mm with the bodged plate.Wish I'd tried harder setting the Z in retrospect :rolleyes:

Pic shows honking plate next to the Warco depth stop plate it replaces :beer:

Robin

[Robin Hewitt]

Oops! :whistle:

[BillTodd]

Oops! :whistle:
Bugger!

Is that cast AL? (doesn't look like cast)

[John S]

Looks cast to me.

Do you want it welding Robin ? or are you going to make a new one.

I can stitch you that back for the price of the postage.

[Robin Hewitt]

Do you want it welding Robin ? or are you going to make a new one.

I can stitch you that back for the price of the postage.


I weakened it by putting in steel insert nuts then went a bit wild tightening the bolts :naughty:

All part of life's rich tapestry but still a silly mistake, looks like I'm further down the learning curve than I thought :rolleyes:

I was going to run a bolt down through, maybe fillet the cracks with Al solder but this here welding sounds interesting, I could use a smaller bolt. See what you are up against before volunteering though :beer:

[John S]

No problem with that, wouldn't put the broken bit back anyway, just build the whole end up with filler weld, 15 minute job.

.

[Robin Hewitt]

No problem with that, wouldn't put the broken bit back anyway, just build the whole end up with filler weld, 15 minute job.

A generous offer, OTOH I am more than a bit impetuous and an M8x40 looks rather inviting. If I can close the gap I could be up and running today :naughty:

I have a spell in the dentist's chair at half past ten, a good opportunity to cogitate :beer:

[Robin Hewitt]

M8x40 did the trick, gap closed, put an arrow on the pic in case you couldn't see where it was :beer:

[irving2008]

M8x40 did the trick, gap closed, put an arrow on the pic in case you couldn't see where it was :beer:Did you drop some loctite in the gap before closing?

[Robin Hewitt]

Did you drop some loctite in the gap before closing?

Because if it gets the faintest jiggle it will never close that sweetly again :nope:

Good idea, but then I'd have had to clean off the RTD and I was rather enjoying putting it together :naughty:

If I take it off I promise to Loctite it :beer:
Kip, I don't know how it compares to driving the handwheel because I've never tried that. If I had to guess I'd probably plump somewhere between amazing and freakin' miraculous :whistle:

[Robin Hewitt]

Curious.

I wound up the feed rate on that hybrid stepper and it would do 18mm/s but clacked out at 20 mm/s

I disconnected the Y axis to get some spare Amps then switched the hybrid peak current from 1.9A to 4.2A peak.

Wouldn't do 18 mm/s anymore :nope:

Possibilities...

The cable on the Z is altogether too long and it's having problems controlling the current.

The stepper driver can't actually cope when you increase the current.

I imagined it :heehee:

So, any ideas what driver/PSU combination I need to shop for?

However, 18mm/s is a lot better than I had before so I made new plates to fit them on the X and Y. Stuck a fanciful strain relief bracket on the X so I can bolt the flexy trunking to it.

[John S]

Speed is a factor of voltage, not amperage, try it at 440 volts and see how fast it goes.

[Robin Hewitt]

Hi Kip

The cable goes from the front of the cabinet, out the back, up by the column and then forward again.

I'm using the MSD542 driver at 40 Volts.

18mm/s is top whack, would never dare run it at top whack, 10-12mm/s is probably credible for the G00 though.

It's a bit cluttered at the top, but I could probably mount a driver and PSU within 6-10" of the motor if I had to :naughty:

Robin

[irving2008]

Hi Kip

The cable goes from the front of the cabinet, out the back, up by the column and then forward again.

I'm using the MSD542 driver at 40 Volts.

18mm/s is top whack, would never dare run it at top whack, 10-12mm/s is probably credible for the G00 though.

It's a bit cluttered at the top, but I could probably mount a driver and PSU within 6-10" of the motor if I had to :naughty:

RobinAs John said, speed is a factor of voltage and, I'll add, inductance of the motor. Higher inductance, lower voltage = lower speed. Also remember torque, speed and power are related. So it may well be that these motors running at that voltage simply can't produce enough torque to operate reliably at a higher speed. Can you get higher revs (step rate) off load under the same conditions?

Cable length is going to make little difference unless its so thin as to be dropping considerable volts.

What size/spec of motor are you running?

[Robin Hewitt]

As John said, speed is a factor of voltage and, I'll add, inductance of the motor... What size/spec of motor are you running?

Yes, but the question is...

Why did the max speed decrease when I doubled the Amperes? :whistle:

[BillTodd]

Did you saturate the motor?

[Robin Hewitt]

Did you saturate the motor?


Possibly right :beer:

I'm sitting looking at the motor spec sheet and it shows parallel as 4.2A/phase, but series as only 3A per pair. Sounds illogical :joker:

If I knew what "hybrid" meant it could explain it...

...perhaps

[Gary]

That motor is 4.2A in unipolar and 3A in series and 6A in parallel.
If you are connecting it is parallel and only giving it 4.2A, you are under powering it.



Possibly right :beer:

I'm sitting looking at the motor spec sheet and it shows parallel as 4.2A/phase, but series as only 3A per pair. Sounds illogical :joker:

If I knew what "hybrid" meant it could explain it...

...perhaps

[BillTodd]

If I knew what "hybrid" meant it could explain it...

As I understand it, Hybrid motors have a magnetised rotor as opposed to the soft iron cores used in earlier types.


If you are connecting it is parallel and only giving it 4.2A, you are under powering it.
That makes it very unlikely to be motor saturation :(

I wonder if it's a resonance (force/mass) issue - extra current = extra force. Might be worth trying a mass damper on the motor (i.e. I would stick a large/heavy washer to the motor pulley with a bit of double sided tape and re-test)

[edit] ISTR you have a spare pulley, try sticking that on top of the existing one (use DS tape or blu-tack something with a bit of 'give')

[Robin Hewitt]

Today I reworked the tapered Gibb strip fixings.

The original adjusters are totally naff allowing the the Gibbs to move, so they went from tight to loose as you reversed direction on the slide :nope:

To fix it I fitted 35tpi x 3/8" studs, screwed the adjusters on to them, (still sloppy) then bolted on end caps to pinch it all together... TIGHT :beer:

[BillTodd]

That's a good idea Robin, consider it stolen ;)

I have a small snippet of steel wedged between the adjuster and gib ATM doing the same job.

[Robin Hewitt]

Next, tramming it 'cause it ain't square :thumbdown:

I chucked up a dti on a 9cm radius and measured to the bed at 4 points of the compass. Plus is a dip, minus is a bulge.

North 0.0"
East -0.004"
South +0.0035"
West -0.0087"

As I measured each point I drew a circle around the DTI tip so I could measure all four points moving the bed rather than the dti...

North 0.0"
East +0.002"
South +0.001"
West +.0025"

Yes, my column is out of square and the bed is a ski jump :nope:

RS do pre-cut shims with a MOQ of 10, so I ordered the 0.05mm (0.002")

Never tried this before, think I need 8 of them and a bed skim :whistling:

[BillTodd]

Did you use a parallel or similar between the bed and the DTI? It helps to smooth out bed irregularities.

My base was shimmed from new. The shims are brass coloured and less than 1 thou" (fitted near the bolts).

BTW Don't get too carried away - the head on these things can be flexed a few thou" with not too much trouble.


Next, tramming it 'cause it ain't square :thumbdown:

I chucked up a dti on a 9cm radius and measured to the bed at 4 points of the compass. Plus is a dip, minus is a bulge.

North 0.0"
East -0.004"
South +0.0035"
West -0.0087"

As I measured each point I drew a circle around the DTI tip so I could measure all four points moving the bed rather than the dti...

North 0.0"
East +0.002"
South +0.001"
West +.0025"

Yes, my column is out of square and the bed is a ski jump :nope:

RS do pre-cut shims with a MOQ of 10, so I ordered the 0.05mm (0.002")

Never tried this before, think I need 8 of them and a bed skim :whistling:

[Robin Hewitt]

Did you use a parallel or similar between the bed and the DTI? It helps to smooth out bed irregularities.

I measured it two ways to compensate for the lumpy bed, but seem to have cocked up :nope:

I reckon my -4 mils East measurement was actually a +4 mils, giving a +-6 mil East West tilt. I mean the tilt has to show even either side, doesn't it? :whistling:

6 mils over a 9cm radius translates to nigh on 0.3mm of packing at the column base, it's around 0.1 degrees out. I could see it was a bit bent when facing with larger endmills but didn't think it was that far out.

New plan. I have some 0.47mm brass sheet, I'll put a strip of that down the West side of the column base then shim the East side until it comes true(ish).

Not much of a plan, but hey, it's a plan. Now where did I put my piercing saw? :naughty:

[BillTodd]

6 mils over a 9cm radius translates to nigh on 0.3mm of packing at the column base, it's around 0.1 degrees out. I could see it was a bit bent when facing with larger endmills but didn't think it was that far out.

Seems a long way out even for a Chinese tool. Have you checked the column and base for paint or rust intrusion? Is it bolted down onto something level and flat?

[Robin Hewitt]

You think I might have done a Hubble? :naughty:

It's a cheap mill bolted down to a cheap stand, I don't think level and flat were ever really part of the equasion.

Don't know about paint and rust either, I've never separated it. I will certainly have a look see before I put the shims in. Engine hoist job, wasn't actually planning to remove any of the bolts beyond loose. Don't want to risk it getting away from me :whistling:

[Robin Hewitt]

BTW Don't get too carried away - the head on these things can be flexed a few thou" with not too much trouble.

How anal is too anal? :naughty:

It is now 0.005 degrees out east/west and 0.002 degrees out north/south. Trying to resist going for better but whaddya do.

Started on attempt #2 at a faster spindle because the bearings are unhappy if I try to run the main spindle up to credible small tool speeds.

The fast spindle is an ER11 chuck fitted to the honking plate and driven using a 1:3 timing belt off the main spindle.

Same axial bearings and Belleville washers that have served me well so far, but less than the 1/4 ton preload this time.

I've been looking at those splendid tool location doo-dads based on an electrical contact between tool and workpiece. As featured on this very site :beer:

To that end I have added Tufnol insulators to the shaft housing. No electrical contact between tool and machine body until it hits the workpiece.

[Robin Hewitt]

Spindle housing now bolted together and impossible to assemble :nope:

Have to add a bit of internal clearance so the pulley can get past the belt :whistling:

Measured the resistance top to bottom, slightly worried when I found 8 million Ohms then realised I was in the circuit. Moved fingers and it went away :rolleyes:

[Robin Hewitt]

Stuck waiting for bits :rolleyes:

Did add an air driven spindle lock, now working on the suds nozzles.

Bill says it's a z axis lock and he's right :beer:

[BillTodd]

Stuck waiting for bits :rolleyes:

Did add an air driven spindle lock, now working on the suds nozzles.

Is that a spindle lock or a Z axis lock?

Have you a picture of the whole 'honking plate' z axis? (looks good :))

[Robin Hewitt]

Is that a spindle lock or a Z axis lock?

Have you a picture of the whole 'honking plate' z axis? (looks good :))

By Jingo you're right, I used the wrong word :heehee:

You want a picture from a different angle?

Have to fix a front gate this afternoon. Strict instructions, on pain of nagging :naughty:

[BillTodd]

You want a picture from a different angle?

Just a wider angle showing the whole of the axis would be good :)

[Robin Hewitt]

Testing out my new quill lock then the quill jammed :eek:

Two hours finding the problem. Thought I'd got cast iron dust in the guide when I drilled the holes for the pneumatics. Hadn't. Though the clamp was binding. Wasn't. Thought I'd deformed the quill with excessive locking force. Didn't. Thought it had to be gunge, fetched the quill out and got it clean enough to eat off. Did no good.

If your Warco quill seizes up for no apparent reason and defies all efforts to free it I have the solution. 2 minute job, miracle cure, no disassembly required, will cost you a bottle of Hobgoblin posted to Sussex :beer: :naughty:
Quill now tracks up and down easy peasy, the Z depth on the computer screen agrees with the linear scale on the mill +-0.01mm moving in either direction. However, cutting forces can put them up to .05mm out. Hence pneumatic quill lock.

Yes, I know 0.05 is only 2 thou :joker:

[Robin Hewitt]

Just a wider angle showing the whole of the axis would be good :)


Here 'tis, and here's a link to the 1.7Mb original image http://www.robinhewitt.net/big.jpg

I left the Z screw detached from the honking plate, it bolts on from below sort of thing :beer:
Next I have to fix the ball screws, I didn't look after them and 10u backlash has become 75u, at least I think it's the screws. Have decided to spring the nuts and go for true zero backlash :naughty:

Assuming 1Nm on the screw is enough to move the table under all circumstances, that's 1256N force on the nuts, call it 280 lbf because I don't think in Newtons :heehee:

I have the design, appropriate Belleville washers have been purchased, waiting for the oil pipe fittings, blooming postal strikes, grumble moan complain.

[BillTodd]

You've made a nice job of that Robin :)


Assuming 1Nm on the screw is enough to move the table under all circumstances, that's 1256N force on the nuts, call it 280 lbf because I don't think in Newtons

I pre-loaded my roller nuts to about 400N per roller. There isn't any actual axial load on my lead screw; the pre load stops the roller moving axially within the nut. That seems OK for manual milling with my machine.


It's actually hard to tell exactly what pre-load you've set with a belleville you have to guess using a displacement and trust the manufacturer's table for the spring.

[Robin Hewitt]

It's actually hard to tell exactly what pre-load you've set with a belleville you have to guess using a displacement and trust the manufacturer's table for the spring.

I know, but whaddya do? :whistling:

The extended ball nut housings are now cut, see pic. I plan to rip it all apart tomorrow and try for zero backlash by preloading the nuts :joker:

Ran the fast milling collet today at a tentative 4150rpm and looking good. I'm driving it with 44:13 pulleys off the main spindle.

[Robin Hewitt]

ZERO BACKLASH ON X AND Y :yahoo:

Well, none that I can feel anyway :heehee:

First pic, horrible things done to Gary's ballnuts :naughty:

Pic 2, Belleville washers just about to get bolted onto nuts.

Pic 3, My extra speed ER11 collet chuck thingy :beer:

[Robin Hewitt]

Added a set of Gary's 240 VAC driven stepper drivers and a splash guard.

Don't know how fast they go yet because my computer can't crank them up past top whack and refresh the screen at the same time.

I suddenly feel a need for belt guards on the steppers.

YouTube- Noise

Would you describe this as "quiet"? :heehee:

[Robin Hewitt]

I added the belt covers on the X, Y and Z motors.

The factory fitted screws holding the z axis motor together stopped 3mm short of the base. Not much but enough to hold a flexi conduit fixing so I could tidy the wiring up.

If I'm doing pretties I must be nearly there. The ultimate aim is to cut injection tooling, so thought I'd try the fast spindle and instantly found I needed yet another belt cover :heehee:

I cut a 3mm slot 0.5mm deep across a piece of aluminium at 6100rpm. A fairly good mirror finish resulted except the centre of the cut where the tool proved a mere approximation to centre cutting :rolleyes:

Extreme close up of the cut below. I measured the width at 3.00 to 3.01mm so runout is around 5um, not bad considering the ER11 collet was only finger tight.

Next the electronics...

I currrently run DOS because it leaves the PC timer interrupt free which is nice. OTOH, new computers do better graphics, are much faster and prettier, even if serial and paralel ports have now both been consigned to the dustbin of history.

Obviously I need a processor on t' mill end of the wire. I am opting for a Freescale 8 bit micro running C. It comes in a dinky little 48MHz, 64 pin, square package with buckets of I/O pins, 32k of flash, 2k of RAM, timers, serial port for the phase converter and a full speed USB 2 connection for the PC link. (Real reason for this choice is that our tech bod is looking at it for another product so we already have all the dev kit, programmer and free samples) :naughty:

[Robin Hewitt]

I do like making things difficult for myself, but when it comes to the firmware I become a control freak. If it doesn't do what I want it to do then I want to reprogram it down the USB rather than fudge it. Nothing quite so reassuring as having all the source code :whistling:

Stepper drivers, phase controller and pendant connect with RJ45 patch cords to keep everything neat and tidy, lots of buttons and well buffered I/O for everything I could think of.

The limit switch and tool touch inputs are designed so I can short them to ground via an LED and get a visual check at the far end of the wire. Three FET's to drive the suds pump, quill lock and something I haven't thought of yet. A socket for one of Gary's jog encoder do-berries, a main motor kill switch to keep me safe while setting up, fuses on board and an input to tell me if 12V goes away. If the 5 volt fuse goes then it's goodnight Vienna so no way to monitor that.

[Robin Hewitt]

There's my Christmas entertainment arrived in the nick of time, not bad for 35 Euro's :yahoo:

I've only found one blooper so far, a 6 pin opto isolator with the wrong pinout. Luckily RS had a 4 pin part that fixed it :rolleyes:

Think I'll put everything down before I mount the processor, that way I can check it's connections, make sure I'm not overloading it anywhere.

The three little boards on the right convert Kinco stepper drivers to RJ45 with all the wiring neatly out of sight, should look great if I got the measurements right :whistling:

[Tom]

Nice!

I hadn't thought of using RJ45s for signals - neat idea, and shielding already built in... What current and voltage are they good up to?

I was planning to use Dsub connectors, but that would look a bit 1980s... RJ45's are much more "naugties"! I suppose the only tricky bit (if you're panel mounting rather than PCB mounting) is the crimping....

Anyway, good idea.... Food for thought... :)

[ptjw7uk]

I hope the rj45 are not being used for current carrying as they are not propper connections just insulation displacement connection for signal etc.
Even the punch down connectors are not that good for current much better to get solder or screw types.

peter

[Robin Hewitt]

I hadn't thought of using RJ45s for signals - neat idea, and shielding already built in... What current and voltage are they good up to?

Not a lot, I do occasionally take liberties and they haven't let me down yet :naughty:

With this board I do feed 5 volts out on an RJ45 to drive the pendant. A bunch of pull ups and a GAL22V10. It was either a logic array or only 7 buttons. I want a lot of buttons.

One component is worth a mention, a lucky find, RS part number 249-050

The board was getting cluttered with clamp diodes, the DALC208 has 8 low capacitance diodes in a dinky 6 pin smt pack. You feed it power, ground and 4 inputs that may need clamping. A real space saver, I used 6 of them :beer:

[bambuko]

...they are not propper connections just insulation displacement connection...
As an ex(retired) AMP guy :rolleyes: ...
Perhaps you would like to re-phrase this statement :nope:
They are perfectly "proper" for the job they are designed for :clap:

[ptjw7uk]

Possible but I have spent hours trying to find the one that wasn't done proper, even had trouble with the puch down ones.
There again in the past I have known screw down terminals to cut through the wire.
Best ones I remember were the old round plessy ones as long as you had all the tools they were easy!

peter

[Robin Hewitt]

Everything tested okay barring a couple of missing components so I fitted the cpu.

I couldn't find a 180R resistor to drive the opto isolator, one resistor pack and RS came up no stock on the USB socket, had to fudge a right angle component in to a vertical. :whistling:

Nothing there that will stop it stepping so I can get on with the firmware.

It protrudes through a cut out in the mill stand door, so on the back it has 5 push buttons, a speed control pot and a socket for the pendant. Everything else comes out the front inside the plinth.

The switches will be three press on/press off for main motor, suds and quill lock, a pause button and and an 'everyone off'.

[Robin Hewitt]

A quick progress report..

Slight mis-wiring on the USB and I didn't notice I needed a 1 Meg resistor across the crystal before it would spark up.

Whatever, the computer now makes the dee-dum USB noise when I plug/unplug the controller and a test program can read inputs and write outputs under XP.

Probably doesn't sound like much but from here on it's only typing. Fortunately I've got my big brother doing that bit for me :naughty:

Robin

[Robin Hewitt]

I cut some metal and it came out a bit naff, see before and after pic, arrow points to the problem bit.

Two changes, I wrote my own dxf to G-code software taking control of the tool path. Bubba, over on CNC zone, told me the Warco column pinch bolts were naff.

I've been measuring the new part and dimensionally it is slightly undersize, somewhere between -0.00 and -0.05mm

Theoretically with a bit of compensation I could get the tolerance down to .025 mm if I always used this tool and this collet :dance:

Bubba suggested I changed the bolts. I looked at them and found they had worn badly. When I thought I was doing them up tight I was actually trying to swage a new thread. I fitted overlength M16's with spacers and anti-seize compound. I can now use a ring spanner rather than the awkward Warco offering and it seems to like it.

[Jonny]

Any further news?
I have the smaller Machine Mart one made redundant 6 years ago, very accurate.
Was rather intrigued by the lead screws, i broke mine 9 years ago and close on £200 and 6 months delivery from MM.

I always left a ratchet and socket at the back for the two nuts.

[Robin Hewitt]

I did inset 18" of 10mm square gauge plate along the back of the bed to set things square against and it is proving incredibly useful.

It was a very tight fit, had to tap it in with a hammer, half expectied the back of the table to split off but I seem to have got away with it.

Haven't done the splashguards and pneumatic spindle lock yet.

Next I want to wire up the 1 micron tool locator. Hopefully this will let me set the tool position exactly where I left it the day before and autoset the height following a tool change.

[swarf-boy]

Hi Robin,

I have read this thread (more than once) on the work you carried out converting your Warco Major to CNC.

I too have the same machine and I am contemplating doing the same thing. (but I am new to the CNC field, but do have some local backup on seting up using Mach3 etc)

I do have a couple of questions for you that are specifc to this conversion, if you would be so kind.


Was the overall expendature and effort worth it, in terms of the usefullness of the machine when finished?

I am perticularly interested in the Z axis arangement you came up with. How is that working out for you now?

Which motors and controllers did you finally go with and on what voltage do you run them. Do still you consider them to be the correct choice?

With hindsight, is their any aspect of the project you would change and why?

[Robin Hewitt]

I do have a couple of questions for you that are specifc to this conversion, if you would be so kind.


Overall expendature and effort worth it? For me, yes. I had already CNC'd it without the benefit of CNC and as a budget job I got what I paid for. If you want to get the tolerances down, and keep them down, I think preloading nuts and screws is a good idea. The mill has 2 problems when you try to cut better than .001". The column flexes, the quill is sloppy. Grease the quill, try to keep the head wound right down when milling. Value for money? I did it because I became obsessive about seeing a couple of zero's after the decimal point on my calipers, value for money depends on how freaky you are.

The 'Z axis arangement' works well. Setting up a mill is a time consuming business, so putting in two 8mm screws to lock the Z is not exactly onerous.

Motors and controllers? I used 3.5Nm NEMA34's and Garys 240 volt drivers that come straight off the mains with no seperate PSU. The motors were rated 3A wired in series, the lowest setting on the drivers was 3.18A RMS, 4.5A peak. They seem unstallable, snapped a 12mm end mill like it wasn't there when I had a whoopsy. Not sure how fast they go, my computer can't keep up. I'm quarter stepping at 200 steps to the mm. They run a bit hot but it doesn't seem to worry them.

Anything I would change? Yes, I would mount the Y axis nut 12mm towards the front so I could have gotten the tool centre in line with the front of the bed when I cut it flat. The bolt heads collide at the back. Also a few annoying clearance errors. The Z axis digi scale top interferes with one of the bolts holding the cheeks in place, had to miss the bolt out. The Z axis handle comes very close to the Z stepper conduit clamp. I was slightly concerned that the heat from the X axis motor might be getting on to the lead screw but it doesn't seem to get above blood heat so probably okay.

[swarf-boy]

Hi Robin,

Thanks for the swift reply.

That is all very usfull info.

May I just stretch you a little more on a few points.

Do you have any suggestions on how to "stiffen" up the column

Which motors exactly did you use ( part numbers would be extremely usefull) and were did you get them from?

What is the voltage at the motors?

Thanks

[Robin Hewitt]

Gary was no stock on the motors at the time so I got them from ArcEuro. I think they are the same motors wherever you go.

I never measured the actual voltage at the motors, I soldered and shrink wrapped everything before I plugged it in. Not sure I want to measure it at the driver end either, I am a bit nervous about exploding voltmeters. I used fat wires with fat insulation, had to cut part of the wire core away before it would go in the driver connector. Overkill perhaps, but I don't expect much in the way of losses even on the long run to the Z so probably around 200VDC.

The drivers aren't cheap but as they plug straight in to the mains you save on the PSU, also you can be fairly sure you won't have to replace everything later due to a lack of volts at the motors like I did.

Not sure you can stiffen the column. I thought about concrete but the movements are so small, thermite? Also I can't be sure the flex is actually in the column, may be the whole thing bending, may be where the column joins the base where I shimmed it to cut square. It is practically impossible to measure.

[M250cnc]

I have read the whole build and it brought back memory's as i have a similar mill, i hope Robin does not mind me adding some info of how i tackled the problems.

My biggest issue was the Z axis namely the play of the quill in the headstock. I also made three failed attempts at cnc'ing the Z

My own quill had just over .1mm slack which is abysmal, i had a quote to hard chrome the quill and grind it to size of £300 and thought there must be a better way.

I cut through the front of the headstock with an angle grinder also at the top of the headstock to allow the quill bore to become flexible in diameter.

So now the quill was too tight, so i have adjusting screws that allow me to open the bore for the quill this allow me to get a beautiful fit of the quill in the headstock tight enough that there is no play but free enough that there are no issues with lost steps.

To CNC the Z i based my design on a Bridgeport mill (Getting the Nut As Close To The Quill As Possible) which i have also, by machining a slot in the quill i have mounted a custom housing that takes the ball nut it is made in steel and by using two strips of steel in the headstock i can take all the wobble out of the quill.

In fact i have the quill rotated 180 degrees in the bore as that was where it seemed to slide as easy as possible so now the the teeth for the gear are at the front.

It has two steel cheeks attached to the headstock and looks very similar to Robin's in that respect.

I do not suffer from flex in the column and my belief it that the flex in Robin's machine in fact comes from the poor fit of the quill.

Although clamping the quill will minimise problems it will not eliminate it completely.

Phil

[Robin Hewitt]

I cut through the front of the headstock with an angle grinder also at the top of the headstock to allow the quill bore to become flexible in diameter.


Hi Phil
This really does beg a picture :smile: :smile: :smile:

Robin

[swarf-boy]

Hi Phil,

Robin is correct. I would love to see a more indepth view of this too. (PLEASE).

[M250cnc]

Well as I was asked so nicely I will start a new thread.

Here are some pics to whet your appetite.

The first is the complete assembly.

The second shows fine lines on the quill that emphasises the smooth but tight fit.

The third shows a close up of the ballscrew way deep inside the housing.

Don't forget that estop button in a convenient place :smile:

Phil



Sent from my HTC Desire using Tapatalk

[swarf-boy]

Ohh Phil , that looks realy interesting.

Can't wait for you to start other thread.

What do you recon, Robin.