The Sequel - Oh Blimey I bought a CNC Lathe (Beaver TC 20)

[BillTodd]

I have fixed simo-drives and lathe turrets in the past, but they were relatively simple servo failures (99% down to capacitors - large and small )  and mechanical (operator driving the turret into the chuck at a bazilllion rpm)

From my experience positional errors are down to misreads of the encoder , we (a very experienced colleague and I that used to maintain about a hundred cnc machines here in essex) would mostly just replace stuff but occasionally the older stuff would be irreplaceable so i would dive in to component level, with mixed success.

1)could be mechanical  - loose or broken couplings - fixed an AGIE wirer where the coupling had turned to cheese
2)knackered bearings in the encoder  - very common on spindle encoders
3)Older 'analogue' encoders used on older bridgeports would just start playing-up then  eventually stop - probably weakening photo-diodes or bulb failures : the machines were replaced before I got to the bottom
4) broken cables - mostly X or Y axis  usually due to being clogged with swarf ,chaffed bare and eaten by coolant

[BillTodd]

After chatting to my colleague Chris. He has had a problem where the servo motor ref is set at , or too near to,  the turret reference which can cause the location to jitter by a whole motor turn. 

Heidenhain recommend timing the motor reference 180 degrees from the turret reference before setting the grid (coupling) offsets .

Other problems can be caused by absolute encoder (ones that store their position) getting corrupted by loss of battery power. ( I  remember having to manually 'walk' one by turning on and off the lathe to get the axis awaay from the roll-over / overflow because no seemed to know how to reset the encoders)

[awemawson]

Thanks for that Bill - food for thought.

So yesterday afternoon I had another session and totally mucked things up - must have turned to pot the wrong way or something as the system just locked up. I walked away for a think.

This morning I dug out one of my spare servo cards (6SC6100-0NA11) - extracted it's parameter card (switches and pots) and measured the setting on the offset pot . Nominally 10 K with 15 turns end to end. I measured pot overall as 9.98K and bottom leg as 4.35K and top leg as 5.98

NB I had tried this card  a few weeks ago and knew it basically worked.

Then I pulled the same card out of the machine and firstly measured where I had set it - wildly wrong as bottom leg = 7.61K and top leg as 2K - miles out.

Then I set is as close as I could ending up at 4.36K bottom leg 5.37K top leg. (Interestingly adjusting the pot also altered it's over all resistance)

Re-installing the card and re-initialising the machine it starts up like a champ and indexes remarkably well - don't fiddle any more Andrew!


https://youtube.com/shorts/UyyZrmppEEk

[vtsteam]

Yay! 

[tom osselton]

 

[awemawson]

Thanks for the encouragement chaps but I really do think I'm going to end up in a mental asylum 

This morning I went to the workshop with the good intention of fully loading the tool disk with it's full complement of tools and holders and making sure that they are evenly balanced but first as is good practice repeat the run with the existing  load of six tools representing a 'half load'. NO - running program #13 that quasi-randomly selects tools it was back to significant rotational positioning errors, so much so that the curvic coupling wouldn't come together.

Trying to bottom out what had changed (nothing!) I then had an issue that I couldn't jog the machine at all. A quick look at PLC error messages told me that the automatic lube pump for the main machine ways was in fault. Check it's fuse - return fuse to wrong slot in holder - why has a good fuse got 240 volts across it? - find the error and take fuse from spare location where i'd put it back into the live location - OK pump now has power -but no oil! Oil float sensor on bottom of reservoir - fill up with 1 litre of Magna 68 way lube. Hooray we can now jog!

But where were we - what were we doing - oh yes program #13 Turret exerciser. Back to it !

But it now runs - if anything turret needs driving slightly further clockwise but in each and every attempt the curvic coupling is mating OK.

And before you ask - no the automatic lubrication system doesn't feed the turret !

[modeng200023]

Real progress 

[AdeV]

But where were we - what were we doing - oh yes program #13 Turret exerciser. Back to it !
But it now runs - if anything turret needs driving slightly further clockwise but in each and every attempt the curvic coupling is mating OK.

Was the machine switched on for most/all of the time between your first unsuccessful run, and the successful run after mucking about with the oil pump?

If yes: Could be something thermal - maybe there's an iffy PCB trace near a component that gets warm/hot - when cold it's not conducting reliably, once warmed up the expansion is enough to get it working?

Literally clutching at (copper) straws here...