Looking for some basic information for Orac electrics rebuild

[Nealeb]

I have just acquired an Orac which seems in reasonable condition mechanically, but I'm less sure about the electrics!

I understand that it came from a school originally; the next owner stripped out the original electrics and replaced it, and the last owner (now deceased), I was told, had it running and then decided to once again renew the electrics. So, apart from the steppers and spindle motor, there is nothing left of the original setup to give me a guide. It is working, after a fashion.

The spindle drive looks OK - new VFD driving the original motor and the spindle turns at varying speeds. The original spindle encoder(?) has disappeared and there is a single hall effect sensor which is not currently wired to anything. The plan is to put in an encoder to do "proper" threading. I can get speed and so on sorted once I have an encoder in there.

Motion control is currently via some almost-undocumented Chinese USB board driven from Mach3. By trial and error, I have both axes moving. The stepper drivers are on a 3-axis TBA6560 board driven from 24V. Even at the maximum current setting, the slides move somewhat jerkily and lose steps except at very slow speed and there's no way they would achieve the brochure's claim of 1200mm/min. I suspect that the combination of those drivers and too low a voltage means that the steppers just don't develop enough torque. The PSU in there is 24V/20A so I doubt that there is any current limiting there. The steppers also seem to buzz rather a lot - is this normal? Would anyone like to suggest suitable voltage and current settings for the original steppers to get the machine moving properly - I don't mind sourcing appropriate drivers and PSU to suit. I have built a couple of CNC routers and rebuilt the electronics in my CNC mill so this kind of thing is straightforward for me - if I know what I'm aiming at! Do the original steppers cope with microstepping or is it full steps or nothing?

Is it worth upgrading the steppers as well? What should I be aiming at in terms of NEMA23/34 and torque ratings if so? The original motors seem a bit on the low power side, but maybe they are OK?

Motion control. I gave up on Mach3 on my router and mill in favour of UCCNC and UC300ETH cards, but it seems that they still don't have the lathe side of that sorted and there is no predicted date for it. At the moment I'm thinking of PlanetCNC which does seem to support lathe mode, proper synchronised threading, and so on. Is this a reasonable choice? I have seriously looked at LinuxCNC (I have used this in the past, but on a CNC router) together with Mesa cards, but from what I have read LinuxCNC can be a challenge to set up for lathe use, and the Mesa cards seem to be unavailable anyway. My machine does not have an ATC and I doubt that I would ever find one, so that's not really a deciding factor.

A lot of questions at once, I know, but it would be great to get a steer on what direction to take without any clue as to what was in there to start with!

Thanks,

[Nealeb]

When I managed to find out the port and pin numbers for the undocumented Chinese USB motion control board in the machine, managed to get Mach3 to talk to it, and added a missing wire between it and the driver board, things started working. However, it quickly became apparent that the 24V supply plus TBA-something driver chips were not delivering - too many missed steps at anything above painfully slow feed speeds. I have now replaced the drivers with CWD872 drivers which are happy with a 50V AC supply (it was easier to go for the larger drivers and use AC than the next model down which would have needed a rectifier/capacitor setup). 50V 500VA toroidal transformer to supply them. I now seem to be getting reliable operation at 1000mm/min although I'm sure that there is still some tuning to do. To answer my own question - yes, the original steppers are fine, run more smoothly than with the TBA drivers, and stay cool to the touch even with a nominal 5A or so from the drivers. Running happily at 8x microstepping.

Next step - getting rid of the motion controller/Mach3 pair. Probably PlanetCNC as that seems to be the only choice that supports proper threading apart from LinuxCNC and Mesa cards - but seem to be all out of stock...

[DavidB]

There is this thread viewtopic.php?f=43&t=6223 where LinuxCNC has been used with the standard spindle encoders and a simple BOB.

[kfm36]

Also worth checking out Centroid Acorn as a controller as it has a good set of features for a lathe, including threading and constant surface speed. You would have to factor in the software as well as the Acorn card in the overall cost.
Have a Mirac up and running on it and an Orac part-way through conversion.

[Nealeb]

Thanks for the comments. I decided to go with the PlanetCNC hardware and software. Every time I look at the Mesa site the desirable cards seem to be out of stock, and despite my IT background (and some Linux experience) it seems like quite a learning curve and project to get it all properly configured - even if it might be the best overall technical solution. Centroid Acorn looks like it would do the job I want, but it is quite expensive (especially with the needed extra for lathe threading) and it's not clear what it offers over PlanetCNC. I would have liked to use UCCNC and a UC300ETH as I already use a couple of these on other machines but their lathe option is not yet available - and has been "just round the corner" for several years...

I now have the PlanetCNC controller in front of me; spindle encoder (100PPR) is on order, and I need to sort out the other bits and pieces needed to wire it all together. My Orac came with an MPG and couple of switches for axis/jog speed already fitted although not wired, so that's a good start.

I built all the electronics for my CNC router myself and I have upgraded a CNC mill already, so that side of it doesn't worry me. Wish me luck!

[Nealeb]

Quick update, just to show I am still working on this one!

Stepper drivers replaced with something significantly chunkier, running off 50V AC. Could have used a size smaller and slightly cheaper drivers but because these run off AC I avoided the cost of a rectifier and capacitors, just added a toroidal transformer. Worked out slightly cheaper overall. Unlike when I bought it, I can now run the steppers at a sensible "rapid" speed.

Motion control now by PlanetCNC. Seems that UCCNC do now have a lathe-compatible update but I found out too late. Motion control card I/O all by ribbon cable connectors, so also installed a bank of IDC connector to screw terminal adaptors, plus PWM to analogue converter. Steppers and limit switches wired and working. VFD needs 12-24V for input signals so about to install a relay board to take 5V outputs from motion controller. Rotary encoder installed but not yet wired, so need to get that side finished, VFD connections in place, and it should be usable. When I bought the machine it had a couple of rotary switches and a manual MPG wheel fitted, so they will get wired in as well for manual jogging with switches to select axis and step size.

Even have a backplate casting ready to machine to add an ER32 collet chuck!

It's getting there...

[Nealeb]

Well, it was all going well until...

Electronics all finished, once I had sorted out motor speed issues. Hadn't realised that the spindle motor needed to be driven at 100Hz to get full speed, but that's done now. Jogging MPG dial plus axis select and jog speed switches all wired, configured and working.

So, on to first test cut. Didn't quite look right, and close examination showed that there were ripples in the surface of what should have been a parallel shaft, maybe a couple of thou/.05mm deep. Did lots of measuring and eventually found that the pitch of the ripples was about 5mm. What does this suggest, given that the Z ballscrew pitch is 5mm? Yes, that's right, the Z ballscrew is slightly bent, moving the whole saddle in and out as the saddle moves. I have been able to partly disconnect the ballnut from the saddle although I cannot get it quite clear of the saddle yet - need to dig a bit deeper into the hardware - but once the nut is free and the screw is turned, the "bentness" of the ballscrew is very obvious. Next step, I guess, is to take it off the machine and see what can be done about straightening. If not, I shall have to replace it but given that any replacement will come with its own nut that's not going to be trivial to refit. Ho hum...

My best guess is that someone at some time in the past has lifted the machine by the balllscrew and while the nice springy cover protects it from dust and swarf, it doesn't protect against careless handling.

If anyone out there has any good ideas, please let me know!

- Brian

[Nealeb]

Not too difficult to remove the Z ballscrew - and, as usual, having removed it you discover how you should have done it!

I held the ballscrew on V-blocks on my milling machine table and measured about 0.2mm deflection at the centre. Given how easy it is to deflect the ballscrew this amount with gentle hand pressure, I'm surprised that this would have caused the machining defects I was seeing as I would have assumed that the V-guide on the bed would have resisted this. However, while everything was set up, I did what I saw shown on a YouTube video which was to put a stub of steel in the milling machine chuck and, with a piece of wood packing, use this with the Z table movement and DRO to bend the ballscrew straight. It needed quite a lot of movement to take the screw past its elastic limit - that is, to put in a permanent bend - but after a few bend-and-measure attempts, I brought the deflections to about a tenth of what it was. That is, about 0.02mm. Reassemble, and a test cut showed that the machine was now cutting as it should and giving a ripple-free surface. I am still not quite sure that the ballscrew was the problem and maybe I reassembled it more accurately in some way. Who knows?

I have now started exploring gcode/CAM with Fusion 360 using constant surface speed, use of tool table and offsets, and so on. Previously I have just used a CNC mill and router. The mill uses a toolsetter that measures tool height automatically on tool change so no need for tool table but the lathe (with Multifix toolpost) is a different beast.

So far, It's looking good and I'm starting to understand what the machine can do and how I persuade it to do it! One concern is that the spindle motor seems to get quite hot. With the spindle at around 600rpm, the motor case is reaching about 60C or a bit over (just about hot enough to touch). It it a TEFC motor, the ASEA original, and I have taken the fan cowling off and cleaned dust out of the air holes and motor fins. However, it is running at something less than its designed 50Hz speed and I wonder how effective the fan is at this speed and if it is worth adding a separate fan somehow.