Servo Troubleshooting

[TimothyMoulder]

Hello All!

I'm working on a basic troubleshooting procedure for servo systems. Nothing brand-specific, and not strictly PLC, but related Anyhow, attached is my document so far. I'd like some opinions on it, and any tips, tricks and techniques you would recommend including.

Thanks!

TM

 

[Peter Nachtwey]

The beginning is excellent, "assume nothing"

I would add "divide and conquer". I would also list symptoms. For instantance:

1. The motor doesn't move. Then I would divide the system up with test that show each part works or doesn't work.

     A. is the actuator against a mechanical stop?
     B. Can the actuator be moved with a manual voltage source into the drive?
        1.  is the drive disabled by limit switches?
        2.  is the drive enabled?
        3.  is the drive configured?
     C.  If an error is induced manually by turning the encoder, does the servo card generate a corrective output?.
     D.  Is the corrective output the correct voltage in both sign and magnitude?

You get the idea. I would like to see your finished product. I would recommend that you also download manuals from various manufacturers to get ideas.

 

[Steve Bailey]

Tim,

I'd suggest that an early step in your process (after the advice to assume nothing) should be to check the diagnostic indicators on the servo amp and controller. I'm fully aware that there are times when these codes can be misleading. Sometimes they merely identify symptoms and not the underlying cause of the problem, but they are generally a good starting point.

Also, don't overlook the operators of the machinery. A good operator who knows how the machine behaves when it's working properly will often spot subtle differences in operation that can provide clues to what's wrong.

 

[Greg Gauper]

Since you indicated 'servo' system and didn't specify electric vs. hydraulic I'll chime in with the following for hydraulic systems:

If you are using servo or proportional valves with built-in valve spool feedback, bring this feedback signal into your control where it can easily be monitored and compared to the valve command. Probably 95% of all hydraulics problems are due to contamination which results in sticking or valves with poor shift times. A complete failure can be detected with a little logic (compare command & feedback for large abnormal difference, with short time delay to allow for valve shift time).

If the system has blocking valves and an internal pilot source, build in a simple means of checking the step response of a valve. We generally offer a software means of selecting a 'test' mode which applies a 2Hz 10% square wave to the valve so that the response can be observed on an oscilloscope or chart recorder.

Provide lots of quick-disconnect test points on the manifolds and power unit.

Provide simple pressure transducers at critical points in the system so they can be monitored with a chart recorder or something similar.