Smoke and cascading failure with servo, G320

[leo.dearden]

I've been running a K2 router with servos (3.5A, 24V), a home grown PSU (26V, 99,000uF smoothing), and G320s for about 6 months now. I bought the drives 11 months ago.

I recently left them energised but idle for a few days (not by any means for the first time), and when I returned to them my servos remained in the position that I left them in, but the main fuse had blown on the servo PSU. I replaced the fuse, and immediately one of the G320s sizzled, glowed bright orange under its cover, and produced a plume of nasty dark brown smoke. I hit the mains power emergency cut off within a couple of seconds, but of course the damage was already done.

I think that the included photo shows the failed component clearly (the small upright rounded cuboid near to the large electrolytic). The board was a REV 7.

Unfortunately, my troubles don't seem to have stopped with one failed drive.

I swapped the Y axis servo on to the previously unused servo drive intended for the C axis, and reconfigured EMC2 to drive that as the Y axis. The motor remained dead, though the encoder was clearly working (I could turn the Y axis ball screw by hand, and after a fraction of a turn the ERR light would light on the drive). From now on I'll refer to this new G320 as the Y axis drive.

Not thinking clearly enough about what I'd observed, I decided to try the Y axis servo driven from the known good Z axis drive, and got exactly the same behavior as on the Y axis drive. Switching the set up back so that the Z axis drive drives the Z axis servo now gives the same broken behavior as the new Y axis: dead motor, working encoder.

The X axis drive and servo are both fine (they work together). When I put the Z axis servo on the X axis drive, that works too.

So now I have:
servo drive
X: OK OK
Y: unknown bad
Z: OK bad


I don't want to try the Y servo with the X drive in case it blows up the X drive too.

I have made the following measurements:

Z motor DC resistance ~= Y motor DC resistance ~= 1ohm

X G320
PSU ground to ARM + = 61kOhm
to ARM - = 56kOhm
PSU +V to ARM + = 3.7MOhm
to ARM - = 540kOhm

Y G320
PSU ground to ARM + = 5.1MOhm
to ARM - = 5.1MOhm
PSU +V to ARM + = 13MOhm
to ARM - = 13MOhm

Z G320
PSU ground to ARM + = OC
to ARM - = 5.1MOhm
PSU +V to ARM + = OC
to ARM - = 12.8MOhm


So, what now?

It looks like I have damaged the Y & Z drives, and something may be wrong with the Y axis servo to cause it. Does that sound right?

How can I find out if my Y servo has the touch of death?

How can I stop this happening in the future?
I've already learned two things:
- Don't plug and unplug servos while the power is on (I'd done that many times without incident, but I've not read the FAQ which says not to).
- Add per controller 20A FF fuses. I only had a 1A slow blow mains input fuse.

Mariss, which if, any of the drives should I send to you?

Any help apreciated

Leo

[leo.dearden]

Firstly: until now my G320s have worked very well for me, and appart from my current troubles I'm very happy with them. I'll either be replacing them with more G320s or G380s if they come out in time.

Secondly: Another measurement. When the servo motor is disconnected and I jog the axis a few mm, the X G320 drives its output to +-22V, but the Y and Z G320s drive theirs to about +-200mV.

It's looking pretty terminal for them . I'm just waiting to hear back from Gecko now as to whether to order replacements and/or send the problem drives back.

Cheers,

Leo

[Mariss Freimanis]

The power supply is in common to all your drives. Something went seriously south in your power for that to happen. Like a fried bridge rectifier for instance.

Mariss

[leo.dearden]

Mariss, thanks for the reply. If there is a problem with the PSU then it must be something subtle:

Nothing else has burned up.

The DC voltage as measured across the PSU smoothing caps is a steady 26.35V according to my digital multimeter. AC voltage at the same point is 155mV.

The X axis G320 is fine. Still works, the axis moves as commanded.


Marcus replied to my email to support@geckodrive (pretty much a copy of the thread start post) saying that I should send the drives in, so I'll put them in the post tomorrow. I'm looking forward to finding out what is wrong, and what if anything I've done wrong.

Marcus also said: "I am guessing that you may have over-powered your drives, and they are dying because of that. However, I can not say without seeing them myself."

I must have misunderstood something here. I didn't know that it was possible to over-power the drives.
- Does that mean over-current? I thought that the 20A limit was safe: the drive would never allow more than the adjustable current limit (max 20A) to flow and if the limit amount flowed for too long then the drive would error and cut the power. Isn't that right?
- Does that mean over power dissipation? ie: The drive itself dissipated too much power and overheated? Why would it have done that? I thought that the PWM output meant that the G320's power dissipation was pretty much constant, regardless of how much power the servo motor was dissipating?


I realise that until you've had a chance to look at my G320s you won't know what happened to them, but I'm interesed in the general question: the more I understand, the better my chances of avoiding problems in the future.

All the best,

Leo

[Mariss Freimanis]

Two separate issues with two possible scenarios.

A single drive failing like that could be a single-axis motor precipitating the failure. You mentioned you had left them on for several days without moving the motors. If there was any static torque load on the motor without it moving for days, the commutator on the motor could have bridged and shorted, blowing the drive.

You also mentioned the other axis failed as well, cascaded as you described. I took that to mean more than one drive was damaged. The scenario then is entirely different; it involves something that's common to all drives which is the power supply.

Send the affected drives back. We will analyze them and determine if it was door #1 or door #2.

Mariss

[leo.dearden]

First the Y axis drive failed spectacularly (glow, sizzle, smoke).

I (effectively) replaced the Y axis drive. The replacement didn't work, with no obvious sign of trouble. I tried connecting the Y axis to the Z drive, and it still didn't work. Then I found that the Z drive didn't work with the Z axis any more.

So, three drives in total were damaged, but (I think) not all in the same instant.

Through all this, the X axis remained AOK.

Perhaps I should have called the failure 'contagous' rather than 'cascading'.

My hunch is that the common cause of the failures is the Y axis servo or wiring, but I can't find any problem with it, or any measurable difference between the working X servo and the suspect Y servo. The screw terminals to the Y servo motor were loose, and I've tightened them up now. Could a loose connection have caused all this damage?

Anyway, the drives will go in the post as soon as the post office opens.

Thanks again,

Leo