1394 GMC - actual position feedback problem

[zitam]

Hi,

I have problem with one of axis in 1394 GMC servo system. I've two axis configured, first one was configured without any problems, but I can't configure the second one.

Marker Test is ok. During Motor/Encoder test servomotor doesn't make distance inserted in Test Increment window, it makes moves in both direction making weird noises, but information after test says it's OK. Tuning gains stops at 50 % after little movement with information that "Axis fault occured. Please examine and clear fault before reexecting the test" but there isn't any fault in Axis system variables.

I checked all of the setings for second axis and they are same like in the first one. I also checked DC bus connection and it seems ok. I changed feedback cable betwen axis and situation doesn't change. Any idea? Yesterday when I was testing the system everythnig seems OK.

Best regards,

Matt

 

[zitam]

What I noticed is fact that actual position in Axis system variables changes very fast without any movement of axis, so during normal operation I have Position Error Tolerance exceded.

 

[OkiePC]

Autotuning doesn't work in GML.

Actually I shouldn't say that. It did work for me about 5% of the time.

I suggest putting your tuning numbers back to "normal" if you had numbers that were more stable, or refer to the reference manuals which provide some guidlelines for manual tuning. Or, try the auto tuning again with more carefully calculated (or just educated adjustements) for some of the parameters.

Sometimes Autotuning gets pretty close and small adjustments at runtime, or with a test program can polish it off. Other times autotuning can turn a fairly stable but sluggish servomotor into a nervous jittering faulting basket case.

Once you can see (chart) how the motor actual position follows the command and how its repsonsiveness is affected by the gains, you can "see" what your changes cause to happen.

You can adjust the gains on the fly, but there are tricks to doing it right and saving the numbers. It has been a few years...I might be able to find an old thread here with some pics to guide you:

http://www.plctalk.net/qanda/showthread.php?t=66336&highlight=GML

http://www.plctalk.net/qanda/showthread.php?t=19911&highlight=GML

 

[zitam]

Thanks for a fast reply. Problem with tuning doesn't concern me, because now I don't need to tune them correct. biggest concern is position feedback value that is changing without axis motion and feedback turned off. I go online via gml commander and axis 0 position is stable a axis 1 position changes ( in both directions ). I've downloaded program that I created yesterday, which worked fine and situation looks the same (position is changing )

What may have caused that ?

 

[OkiePC]

biggest concern is position feedback value that is changing without axis motion and feedback turned off. I go online via gml commander and axis 0 position is stable a axis 1 position changes ( in both directions ). I've downloaded program that I created yesterday, which worked fine and situation looks the same (position is changing )

What may have caused that ?

How much is it changing? Jumping all over the place? Just a few counts or fractions of units?

Here are some possibilities:
1) the motor encoder is actually moving (loose from motor rotor, damaged, coasting, machine vibration). The higher the resolution, the more noticeable this may be?

2) noise. the low level signals from the encoder have noise superimposed on top of them causing the controller to actually think the encoder is moving when it really isn't. There are encoder loss and noise features that may not be turned on, but the 1394/s-class is pretty good at faulting for encoder noise if configured to do so. If noise is your issue, I highly recommend enabling the detection features as well as adding suppressors, or rerouting cables. Noise on encoder cabling or even high speed inputs on the 1394 can be a big problem. I have had several instances of noise on digital inputs with the 1394. I have only had a few problems with encoder noise, and it always seemed to end up being a broken shield or ground, damaged cable, or damaged motor/encoder.

3) code. Anything running when this happens? Any logic (in GML or externally connected devices) that writes to this parameter?

4) bad module...don't all the motor encoders connect to the system module on the 1394? Can you hook it up to a different (spare) channel just for a test? Maybe swap the configuration for the two axes temporarily? Easy to do if they're the same size...different sizes, not so easy but you can put in a dummy program and just move the feedback connector.

Also, you may already understand this but it's confused folks in the past. The "Axis Feedback" On/Off has nothing to with the signals 'feeding back' from the encoder. It is just a dumb way of saying "Axis Enabled" On/Off. Even with feedback off, if that encoder moves (and the system has power) the controller will continue to keep up with the position.

 

[zitam]

thanks again, I'll try to answer best I can

It changes for 400-500 position units with every refresh going up and down from 5000 to -5000 or going only in one direction, when 1 position unit is 1 motor rev (8196 counts).

1. this is motor with resolver (1326AB-B410G, don't remenber now) and resolver looks ok. I also change motors, without any effect.

2. whole drive is at my desk in office so there is no noise there

3. It happens without any logic running. When I launch simple program (home, feedback on, move) and the good one works fine and this one goes in fault state becasue of position error tolerance exceded (feedback stays off, what should hapened automaticly)

4. Yesterday it worked OK, and nothing special hapened what could damage inputs.

Can't say nothing more

Any other ideas

 

[OkiePC]

So, it's not the motor, and unlikely to be noise, then I would try a different resolver channel next.

 

[zitam]

I will check it tommorow. fingers crossed