Powerflex 527 motor overshoot.

[PiperJohn]

I've got motor that ran fine for years, but now it's sarting to overshoot its target position a few times a day. Is this indicative of a motor aging and needing the PID loop tuned?

This is a belt driven horizontal cutter with no brake and an incremental encoder for feedback. No noticeable mechanical or load changes.

When is tuning necessary? Any advice for someone who's never done it?

Thanks,
John

 

[Rob...]

Likely the belts stretched or slipping.

 

[Steve Bailey]

Overshoots and then tries to reverse to the correct position? Or when it signals the move is complete it has gone too far? Are you moving the product to a position and cutting it off with a fixed knife, or are you moving the knife itself? Is the encoder mounted on the motor or on a wheel driven by the product? Is the belt a timing belt or a V-belt?
Lots of things could be causing an imprecise move. Retuning is only one of many possible remedies, and not necessarily the most likely to correct your specific problem.

 

[PiperJohn]

Motor drives a cogged belt with an encoder mounted on a pully. The knife cuts moving across the part. On occasion it stops about an inch or two too far and faults. It makes one pass each cycle but doesn't reverse on its own to correct.

Example, it normally moves from 10 to 100. Then it might move to 130, triggering a target fault.

The motor does not loose its position. After an overshoot I can bring it to home position and the encoder position is correct. The load of each cut is the same load on the motor.

I don't see a mechanical problem with it.

Any ideas?

 

[Rob...]

What type of encoder are you using? Incremental or absolute?

I presume this is a system working for some time? If so, any recent software changes?

 

[PiperJohn]

It has been running nearly 3 years and uses an incremental encoder with a home sw.

The motor does not have a brake and the drive doesn't have a dynamic brake resistor.

The problem seems to be getting worse, but might have been there from the start.

It will often overshoot slightly and won't trigger an alarm. The operators will usually just hit a reset button to recover unless it overshoots far enough and hits the hard stop.

No software changes to the program.

What to do?

Occasionally in the past it would give a buss overvoltage fault.

 

[jstolaruk]

Does the VFD have a brake resistor connected? That buss overvoltage fault may be a tell-tale.

 

[PiperJohn]

It doesn't have a dynamic brake resistor.

It hasn't had that buss overvoltage fault for at least a year.

I believe it's a design issue. Perhaps a under sized drive that can't respond fast enough to consistently stop it? Expecting too much control of a VFD instead of using a servo?

Regardless, it is getting worse. When and where is tuning warranted?

I could probably mitigate over/undershoot by lowering the speed or acc/decc but would like to figure out what is changing.

 

[jstolaruk]

Lower the Decel setting and see if the problem goes away or its occurrence frequency reduces. Easy change at no cost.

 

[PiperJohn]

Lower the Decel setting and see if the problem goes away or its occurrence frequency reduces. Easy change at no cost.

I can try lowering it a little and widening the positions. Slowing too much may effect the cut quality.

Thanks

 

[PiperJohn]

IWhen is tuning necessary? Any advice for someone who's never done it?

I was hoping someone could answer my questions about ;
Tuning, and when it's necessary.

 

[kalabdel]

I realize you are looking for a specific answer re PID re-tuning and I don't have one.


It does look like:
1-Pres-existing problem: You seem to be unsure if it is or not.
2-Mechanical: You do not believe that is the case.
3-Drive: How is the drive commanded, discrete or via some communication?

 

[kalabdel]

Can you clarify the following:

• Overshoot a few times a day: In a row, random, cumulative?
• Normally moves from 10 to 100. Then it might move to 130: What unit is this, encoder pulses? What is the encoder resolution? Do those 30 extra units equate to one or two inches (you stated too far by one or two inches)?
• Overshoots too far and faults: What faults, machine as in program fault (what you referred to as target fault)? The question is to make sure that the drive itself does not fault.
• PID loop where, in the drive or PLC? Is the encoder connected to the PLC or to an interface on the drive? How is positioning implemented via the drive (drive is setup of positioning) or strictly PLC?
• I can bring it to home position and the encoder position is correct: How do you know the encoder *value* and what is correct? Does it always return to the exact same value?

 

[PiperJohn]

Thanks Kal,
The overshoot happens randomly but but becoming more frequently. I am usually not called unless it overshoots enough to crash the drive where I have to reset an overcurrent fault. Otherwise operator can reset program driven fault because motor position is outside of anticipated positions. The target positions (back & forth) are entered on HMI screen and actual position is displayed on screen.
I arbitrarily picked 10-100 as an example. It cycles from 30mm to 1250mm at the moment, these are the target positions in program.

The VFD is set to position loop. The Incremental encoder homes on switch at startup. Feedback counts are 6552 per rev. Scale is 180mm per 5 motor rev.

It moves at velocity 1800 with 18000 acc/dec. Encoder hooks to powerflex 527 module and commands are communicated over CIP network via controller.

Yes, I believe it's a design problem but something is changing to make it worse.

I know it is holding position mechanically because after an overshoot I can manually bring motor to its hard stops and the positions match read what they normally do. 0mm to 1600mm.


So what is changing? Belt tension and slides good. Motor wear? Drive caps leaking? Program has not been changed.

I don't know.

 

[kalabdel]

Thanks Kal,
The overshoot happens randomly but but becoming more frequently. I am usually not called unless it overshoots enough to crash the drive where I have to reset an overcurrent fault. Otherwise operator can reset program driven fault because motor position is outside of anticipated positions. The target positions (back & forth) are entered on HMI screen and actual position is displayed on screen.
I arbitrarily picked 10-100 as an example. It cycles from 30mm to 1250mm at the moment, these are the target positions in program.

The VFD is set to position loop. The Incremental encoder homes on switch at startup. Feedback counts are 6552 per rev. Scale is 180mm per 5 motor rev.

It moves at velocity 1800 with 18000 acc/dec. Encoder hooks to powerflex 527 module and commands are communicated over CIP network via controller.

Yes, I believe it's a design problem but something is changing to make it worse.

I know it is holding position mechanically because after an overshoot I can manually bring motor to its hard stops and the positions match read what they normally do. 0mm to 1600mm.
So what is changing? Belt tension and slides good. Motor wear? Drive caps leaking? Program has not been changed.

I don't know.

This is going to be fun!
Either get much more information from the operator or better yet put in a camera to capture those events. Otherwise it pretty much could be anything. Really, mechanical issues should be eliminated first.

What is the drive IP? So I can locate that specific drive and its related Axis in code.
There are three 0.55KW with 1638 PPR x x = 6552, and have pretty much the same configuration.