Help tuning a G320?

Short version: could I get some advice on tuning a G320 (not the newer G320X), without an oscilloscope?

(Very) long version...

I have a 3 axis machine, running 24VDC, 60W, servo motors, US Digital E5S-250-250 encoders, and Gecko G320 servo drives.

I have a fourth G320, and I've recently realised the control box for my machine has the ability to run a fourth axis, so I want to use that fourth G320 with an additional motor and encoder.

That fourth G320 (and one of the other existing three) went back to Geckodrive in 2015 as I thought they were faulty, but they were returned with a clean bill of health.

To test the extra G320 before I drop cash on a fourth motor, I've simply swapped the connector block from the G320 controlling my Z axis onto that "new" G320, and set the limit, damp, and gain pots to match. And... it doesn't work.

Basically, I get the same problem I first had with this G320 drive - any movement of the axis will almost immediately trip the drive out (showing the fault light). Having dropped the axis acceleration in Mach down a ludicrously low setting (approx 2 inches per second per second), I can jog the axis up and down for a while, but it'll still sometimes fault. Also, when left alone, the motor will sometimes oscillate back and forth, slowly judder in one direction, or be still but then suddenly "jump" a short distance and then fault out again.

From previous experience I know the fault will follow the G320; it's not related to the machine axis.

I've also tried adjusting the damp and gain pots, but with no success. So... how do I move forward from here?

If it would be of any help I could maybe try to provide some videos of the behaviour?

Have you checked that the encoder phases are correct? You may need to swap the channel A and B .
Dan Mauch

Dan - good question. I haven't, but then all I'm doing is swapping the whole green connector block from one G320 to another. In the photo below you can see the "bad" unit on top, with a good one underneath. Unless there's an element of randomness (and I'm just lucky in that three of my G320 units are identical) I'd expect them all to behave the same with the same connections (component tolerances and real world issues accepted).

Attachment 416020

Note that the issue isn't the physical placement - I can put the "bad" unit where any of the good ones are, and it'll still trip out pretty much immediately an axis is moved.