Kinetix E64 Fault

Has anyone else experienced nuisance E64 faults on AB Kinetix6000 drives? The fault is an encoder noise fault. We've struggled with these on and off for almost 5 years on our equipment. It never occurs frequently enough that we can do any troubleshooting but every month or so another customer calls in an E64. AB has come in and looked at our equipment and says we've gone above and beyond all the noise mitigation bonding requirements they recommend. We had a testing facility come in with an ESD gun to see if static was the issue and it never faulted. I'm curious if we are alone or if anyone else is having similar issues.

I remember this happening on a system that I installed but the plant electricians did manage to sort it out themselves

apart from the shielding (which you seem to have in control) I remember:
machine earthing had an effect (especially machine frame back to the cabinet)
I remember firmware versions had an effect (but that may have been a power up error)
cabinet temperature (ie drive temperature)
A clean power supply for the 24 Vdc - use a oscilloscope to check

A few ideas:

1) It is possible noise is coming into the machine from outside? Do you have a filter installed at the mains?

2) Do you have more than one identical drive? Are they all faulting? If not, then try swapping drives.

3) If there is real noise, you should be able to observe the failure by using RSLogix5000 to trend the position value of the servo axis. You can set up the trend to record a fixed amount of data continuously until a trigger, where it will stop. Later you can go see the results.

4) You have incremental/absolute/stegman encoders? If you used a different type perhaps the problem will go away? You might get AB to lend you a motor and cable for that.

5) This last idea is going to sound impossible. I never would have believed, but this really happened with four of our systems. We built some machines where a couple of pulleys were being rotated by timing belts at around 600rpm. An AB 845GM absolute encoder coupled to the shaft is used to trigger a Kinetix 350 with an MPL servo. Occasionally, as often as 5 minutes, and as long as two weeks, the servo would miss a trigger. We troubleshot for months, believing that the 845GM encoder was at fault. After all, the trends clearly showed that the outputs of the encoder would periodically drop out, or stick at a single value for 30-40ms at a time. Mysteriously, the problem only occurred in our shop, and never at the customer plant, where they were running five other identical systems. The problem was ultimately traced to static electricity built up on the belts, and pulleys, periodically discharging across the face of the encoder's chrome plated disk, interrupting the encoder feedback. It seems that when the shaft is turning, the shaft and ball bearings begin to "float" on a thin film of grease, disconnecting the assembly from the ground provided by the outer race of the bearings. We figured that out by measuring continuity while the system was running. The belt flex generated a field which accumulated until the discharge. The clue was the low humidity in our shop during February. We purchased a conductive brush (McMaster Carr) and installed it against the shaft to discharge the field. Our mystery problem disappeared. So, if you have something that could create a static field (a non-conductor or two) and maybe get more complaints in the winter, then you might have a static discharge. A conductive brush on a shaft surface or a spring-loaded ball at the end of the shaft are the two solutions I found that are commonly used for solving static-related motor bearing failure.

Hope this helps!

I used a ferrite core on the motor and feedback cable on Kinetix 350 and it helped.

Thanks for the advice everyone.

We've tried a line filter on the main incoming 480, the 120 to the drive control power, and on the 24V - still got faults. This was done on different machines, not all on the same one.

We also installed the cleanest (lowest ripple) dc power supply we could find (AB & Sola seemed to be the best. We are using AB.)

The faults occur on the IAM and the AM. They have all been the double wide BM03 or BM05 drives, and only with the motors that have the absolute encoders (M style). Never had a fault on the H style encoders.

We've got the latest firmware on everything.

One other interesting thing to note is that when the fault first started appearing we discovered that if we changed our move profile from trapezoidal to s-curve the fault would go away (kinda points to a firmware issue). But lately we've had the fault occur during s-curve moves too.

Your experience with static doesn't surprise me one bit. Someone else mentioned something about this in a reply to a post I made a few years ago (maybe it was you). The static theory was what drove us to have the ESD testing done but I suppose our test could have been flawed somehow and static is still the issue. On the other had we've had machines fault in S Carolina, Texas, Mexico, & Puerto Rico which I wouldn't expect to be areas with low humidity. Regardless of that I guess I shouldn't throw the static theory out the window yet.

Jothiswarup said:

I used a ferrite core on the motor and feedback cable on Kinetix 350 and it helped.

Click to expand...

Where did you put the ferrites? At the recommendation of AB I also tried small ferrites on just the A, A-, B, & B- wires inside the encoder breakout board housing on the front of the drive. I had the wires wrapped around each one about 5 times. It actually made the problem worse. It was faulting every 5 minutes that way.

Sorry for the late reply. This is on a Kinetix 350, I put them on Motor power cable as well as in the Motor feedback cable where the cables are connected to the Kinetix. It did solve my issue which proved there was some noise in the system.

I've fought many Kinetix battles. Many of the "noise" issues I've fought turned out to be mechanical issues. Any flex/slop in a mechanical coupling, linkage, gearbox will look like noise to the servo. Our company had been using in house made mechanical coupling. Sometimes there was no slop in them (bad) and sometimes there was .060" slop in them (worse). On a system with high gain, trying to position to .001", .060" is a mile.

Also had a machine where the mechanical coupling that would flex, A LOT, during accel and decel. Occasionally the drive would fault. To fix this issue, not my idea, we installed a larger servo, larger gearbox and larger belt actuator so the slop looked like less of an "problem" to the motor. I was out of town for weeks trying to resolve this. Even had the AB servo guy in 3 times. THAT was expensive.

Bottomline look into mechanical issues. It's one of those things where if it isn't BROKEN then it's not a mechanical problem. HAS to be programming/electrical