Failed servo/ encoder help

[Bongo]

Hi Friends. Struggling with fixing the machine I built a few years back. It has a twin rotating ball nut design on either side of the moving gantry. here's the machine build video : https://youtu.be/cqYJS27aC4w?si=ir_XA7-AvJQFJMMD

One of the 400W servos that drive the gantry is playing games. I have been round the houses tracking down the problem (different sides of the gantry moving different amounts, is scary!). At first I thought I had dirt binding the ballscrews but not so. I only found the suspect servo after getting two dial indicators on either side of the gantry. I can see one side gets consistently out of true - eventually faulting the servo drive as it attempts to rack the gantry relative to the side that stays true... I have tried a different servo drive (no change) different cables (no change), switching the output pins to eliminate software and cables from the computer to servo drives... etc etc..

In fact, in 'smart jog' (the servo tuning software, I can see the servo's mechanical angle changing as the faulty servo returns to what should be 'zero' after moving about, while the recorded position feedback stays the same. I think it's therefore probably a problem with the encoder that looks like it screws off the back of the servo...

It's a long shot I know, but has anyone got one of these servos cluttering up a draw somewhere?? It's a CSMT-04BB1ANT3

A more reasonable question- can anyone offer guidance on the repairability of the encoder and any pitfalls to watch out for when dismantling?

[Bongo]

update. I've got the encoder casing off and it is a Tamagawa encoder TS5563N1. I can't see anything obviously wrong, so am a bit stumped. Any suggestions? The encoder itself isn't coming off the motor easily. It's held onto the motor body with some very small fastenings that have been liberally glued in place.

[routercnc]

If you have access to an oscilloscope and make some breakout leads between the encoder and the driver you might be able to check what is happening. By this I mean intercept the encoder wires while it is connected to the driver. I would guess there is at least power, ground and probably 2 pulse trains (slightly out of sync with each other depending on the direction of rotation). Maybe a google search will tell you the pin outs, plus if it is incremental or absolute encoder.

Or perhaps you could figure out the wiring by arranging a breakout connection from a known working encoder to a driver. But this would only confirm that it had stopped working. Then it’s a decision on whether to buy another motor with encoder, or hunt around for a compatible encoder if they are sold separately.

Without a scope, assuming you know which are the pulse train pins and ground you could connect a multimeter and check the RMS Voltage of a pulse train at different motor speeds. The values would not mean much but a reading would increase or decrease with speed.

[m_c]

Generally encoders fail completely, not intermittently.

Have you checked all the wiring is OK?
A broken wire (or one making intermittent connection) would give those kinds of symptoms.

RouterCNC's suggestion of connecting an oscilloscope would give a far better diagnosis.

[Bongo]

Hi guys, thanks for the suggestions. I don't have immediate access to a scope, though I think I could borrow one. Here's the 'progress' on this:
After rummaging about under the encoder's casing, and concluding that I was way out of my depth, and it was fixed to the servo in a way that was NOT letting me take it off without damage, I put it all back together as best I could and... The problem has disappeared! Leaving me both pleased and with the sense that it will probably come back when least convenient. For now though it's been working for some time without any signs of missing a beat.
A tentative hypothesis is that maybe there was a speck of dust or other build-up on the scales that got dislodged, or probably more likely, there is some loose connection that got pushed about and is slowly working its way loose again...
At this point, as I have tested the servo with a known working drive and cable and it was still playing up (back when it was playing up), I think I'll just order another second-hand servo as a spare ready for if/when any of the four fail.
Thanks again for your input guys.