DG4S Misbehaviour

If you think you've found a bug post it here.

DG4S Misbehaviour

Postby HeronG » Fri Aug 05, 2022 3:01 am

Hi guys, this is my first post.

Bought 2 DG4S drives a month ago and just installed then. They seem to perform well, fast, precise, easy to set and install, powerful, but the PID control has a problem that manifest itself more on high resolution encoders, fast motors with low inertia and high voltages.

I made a video to display clearly the problem and hopefully be part of the fix.

It is quite a big dealbreaker since this problem makes the motor always shake when stalled.

A short description of the problem is that the proportional gain on the drive has a dead band where it does not output any power, then it outputs full power like if it was cumulating output but it did not translate into a real output power. So in this band the motor gets unstable whiteout an unstable PID loop which makes it impossible to make the motor settle. All the belts and ball screws and encoders are properly coupled with no meaningful slack or play.

I think this drive has an huge potential when it comes to machine retrofit and many other applications but this is a significant problem that limits the drive usefulness.

If I can help in any way to fix this I will be happy to be of service.
Posts: 3
Joined: Thu Aug 04, 2022 11:00 pm

Re: DG4S Misbehaviour

Postby CT63 » Fri Aug 05, 2022 3:10 pm

Just a suggestion:
Might want to adjust the sampling time and see what happens. Industrial servos have encoder damping to eliminate feedback hysteresis (hunting) caused by mechanical issues or with light loading. The PID integral settings might be too high. Integral has a nasty tendency to amplify encoder feedback noise.
Posts: 108
Joined: Sat Sep 03, 2016 11:19 pm
Location: Connecticut, USA

Re: DG4S Misbehaviour

Postby HeronG » Fri Aug 05, 2022 5:39 pm

Thanks for your answer.

The sampling rate just changes the frequency of the oscillations. Maybe if it where 10 - 20 times faster it would blend the shaking, but I do not think that is the ideal solution.
As it is possible to see on the video I was only using Proportional gain, only 5 and the motor would output no torque at all within a dead band, but it seems like the PID output it is cumulating on the drive, just not been output. As it builds up with the motor been turned it suddenly pulses. this only happen when the drive is transitioning from no output to its smallest capable output.
No amount of tuning can make the hunting stop.
I do not suspect the encoder, it is fully digital, differential, and of good brand. Plus I see no missed steps.
If anyone wants to see more test I can make new specific videos.
Posts: 3
Joined: Thu Aug 04, 2022 11:00 pm

Re: DG4S Misbehaviour

Postby cncdrive » Sat Aug 06, 2022 4:30 am

I think your motor probably has too low coil inductance and/or resistance, so the current in the coild builds up too fast compared to the loop frequency, this could cause an unstoppable resonance.
Try to put some load on the motor to see if it becomes tunable then? You will have to put load on the motor in your applications anyways.
If it will not help then try to install an inductor in series with the motor coil to make the current rise rate slower.
So far we saw this issue with motors with extremely low inductance, e.g. disk motors are like that in which the motor coil is only 1-2 turns aroung a strong neodyn magnet, with those type of motors installing a series inductor is a must.
Site Admin
Posts: 4252
Joined: Tue Aug 12, 2014 11:17 pm

Re: DG4S Misbehaviour

Postby HeronG » Mon Aug 08, 2022 10:06 pm

Hi again.
So I have added the inductors, and with a touch of skepticism, tested again. I am happy to report that the problem has vanished.
On another note,
so the problem was caused by a current gain to high? My motors fall well within the min resistance range and inductance.
Is it possible that it will have an update in the future?
I am ready to declare this drive the best solution for old machines, and intent to update all my geckos to this drive. I just find the dead band displayed on the video above and the need for massive inductor to be inconvenient.
Posts: 3
Joined: Thu Aug 04, 2022 11:00 pm

Return to Report a bug

Who is online

Users browsing this forum: No registered users and 2 guests