Encoder noise - blower induced?

[kaiser_will]

I have never run into this problem before. 460VAC LSI 150HP motor drive, driving 150HP Marathon 460VAC motor, blower driven (TEBC). In most applications, encoder drift at zero speed would not be an issue. In this case, the machine is testing a hydrostatic transmission and the test is intermittently failing due to sensing some main drive shaft movement when the transmission is controlled to be parked. Controller parks the transmission; test machine loads up the drive shaft; test machine tests for drive shaft movement; movement outside of +/- X degrees movement fails the test. This is a duplicate build of machines from 2 years ago (identical machines, mechanically and controls-wise). Blower is on opposite end of motor; BEI encoder is tethered to opposite end shaft. We have exhaustively gone through and tested every part of the signal chain as well as analyzed the incoming and intermediate power with no known culprits found or no "A-Ha" moment encountered. One thing is true - when the VFD is enabled (software command), the blower is also energized and often we have captured an erratic encoder signal at the same time the control system starts to measure shaft movement but the shaft is physically parked. Current idea is that the plastic blower fan is inducing enough EMF through the encoder case that is not being dumped to ground (i.e., HF noise) and being measured as pulses in the controller encoder card. I am reluctant to run a $16,000 motor with no ventilation, but may have to incorporate some logic changes to disable the blower at the time of testing when zero motion is detected. Has anyone ran into noise induced on encoder signal train and how they removed/eliminated such?

[JRoss]

Encoders are generally pretty robust, but here are some ideas for identifying and eliminating the problem. What is the encoder connected to? PLC, VFD, computer? Also, is the encoder absolute or incremental, and what are the voltage levels involved (5V, 24V, etc.). For my comments below, I'll assume you're using incremental. First and foremost, analyze the Encoder output with an oscilloscope. That will help you determine if there is noise, or something else going on. You'll also be able to eliminate or prove the blower as the source of noise. If there is noise, disable the blower and see if it goes away. If you have an electrical noise problem, you could try filters to eliminate it. Also check that you have proper grounding. You might also keep an eye out for mechanical noise, or dithering. If you have enough vibration on an incremental encoder, it can fool the controller into seeing extra pulses. That will show up as an inconsistent square wave on the o-scope. You'll probably want to use a two channel o-scope, one for A and one for B, so that you can see that the pulse trains lead/lag properly. You might try a different brand of encoder. Different brands have different levels robustness and noise tolerance. I haven't used BEI, so I can't speak to their quality.

[JRoss]

Encoders are generally pretty robust, but here are some ideas for identifying and eliminating the problem. What is the encoder connected to? PLC, VFD, computer? Also, is the encoder absolute or incremental, and what are the voltage levels involved (5V, 24V, etc.). For my comments below, I'll assume you're using incremental. First and foremost, analyze the Encoder output with an oscilloscope. That will help you determine if there is noise, or something else going on. You'll also be able to eliminate or prove the blower as the source of noise. If there is noise, disable the blower and see if it goes away. If you have an electrical noise problem, you could try filters to eliminate it. Also check that you have proper grounding. You might also keep an eye out for mechanical noise, or dithering. If you have enough vibration on an incremental encoder, it can fool the controller into seeing extra pulses. That will show up as an inconsistent square wave on the o-scope. You'll probably want to use a two channel o-scope, one for A and one for B, so that you can see that the pulse trains lead/lag properly. You might try a different brand of encoder. Different brands have different levels robustness and noise tolerance. I haven't used BEI, so I can't speak to their quality.

[OkiePC]

Just a thought, but if the blower causes a false pulse when it first starts (inrush of current) then perhaps a soft starter for the blower would help?