Certainly it is difficult to discover "what changed", and even more challenging when the problem has existed for some time, and probably others have been involved, possibly making changes that are not obvious and not documented.
AND, I am guessing that the machine is in fairly constant use, so an investigation of EVERY CONNECTION is not a reasonable option.
Using a clamp-on ammeter to examine the current balance between phases can reveal a poor connection, and that has allowed me to solve a problem. Checking the current in the ground wire connections, and comparing that with other machines, may also reveal a problem.

I am aware that in some organizations there are rules forbidding any electrical access to any machine electrical system while the machine power is present. Certainly these checks should not be done without another person standing by who knows and is able to rescue another who is being shocked. AND PPE is appropriate for this sort of checking.

 

@cawthrawilliam , where it is was economically possible, I made a point of fitting a suitable 3 phase line reactor or choke between VFD and motor.
Helps especially where non-VFD rated motors are used.
Usually mounted in the control cabinet.

 

I have worked with this drive thru out its life cycle and I am very familiar with it. Based on everything I have read, you have eliminated everything except the gear box. My initial thought was mechanical and I still believe that to be the case. The torque accuracy of the drive is very good, but it is dependent on the motor model created during the motor ID run. I am a big fan of the standard (rotating) ID run and it ideally should be performed without being coupled to the gearbox. I would like to make some suggestions/recommendations. You don’t mention the size of the drive. It should be based on the motor FLA. Either a normal or heavy duty drive rating is ok, nothing larger. For monitoring purposes, I highly recommend using ABB’s Drive Window commissioning software - with the fiber optic interface to the drive. If it were me, I would remove the auger from the second system and monitor the speed/torque/current while adjusting the drive speed from 0 - 100%. Use a 60 second accel ramp. You can command 100% speed in Local mode and hit the Local Start button. This will give you a benchmark that you can use to review the performance of the system in question. Good luck.

 

@cawthrawilliam , where it is was economically possible, I made a point of fitting a suitable 3 phase line reactor or choke between VFD and motor.
Helps especially where non-VFD rated motors are used.
Usually mounted in the control cabinet.

I have not done my diligence to check for grounding yet, nevertheless, on the comment you made about the reactor - a colleague of mine and I were discussing the potential for trying an isolation transformer.. Then if no benefit showed, going with what
prairiemystic mentioned with the encoder.. At this point, it is becoming a costly investigation and converging on figuring out whatever works to stop the bleeding, but trying potential avenues without knowing the root cause is driving me insane.

 

I have worked with this drive thru out its life cycle and I am very familiar with it. Based on everything I have read, you have eliminated everything except the gear box. My initial thought was mechanical and I still believe that to be the case. The torque accuracy of the drive is very good, but it is dependent on the motor model created during the motor ID run. I am a big fan of the standard (rotating) ID run and it ideally should be performed without being coupled to the gearbox. I would like to make some suggestions/recommendations. You don’t mention the size of the drive. It should be based on the motor FLA. Either a normal or heavy duty drive rating is ok, nothing larger. For monitoring purposes, I highly recommend using ABB’s Drive Window commissioning software - with the fiber optic interface to the drive. If it were me, I would remove the auger from the second system and monitor the speed/torque/current while adjusting the drive speed from 0 - 100%. Use a 60 second accel ramp. You can command 100% speed in Local mode and hit the Local Start button. This will give you a benchmark that you can use to review the performance of the system in question. Good luck.

We removed the gearbox and ran a ROT ID. We ran it in local mode as well and it was still showing nominal torque values between 8-13%, but jumping all over in between.

We tried two drives for the 5 hp motor in question. The first was rated for 5hp and the second (and current) is rated for 10hp. There was no change in perfomrance after running the auto tunes.

Removing the auger from the second system is not so straight forward as it is an industrial sized extruder, but can be done. The crammer motor, gearbox and auger assist in forcing material down into the extruder. This is the high-level process of the system we are dealing with. When I was first put on this project, I looked at all the mechanical parameters between both and looked for a relationships, especially after seeing that FFT spike, but there was no correlation. Then when i saw how the torque signal was soaring randomly, yet in a noisy way, it made me think of a marginally stable PID controller trying to loosely chase a reference input. I will go back and look at the current the drive is seeing. I know it was pulling ~3.5A under no load based on teh HMI output, which seemed high to me... but i cant remember if it was dancing around like the torque was.. you would think since they are proportional to each other..

 

Sorry for my bad grammar. I didn’t mean literally remove the auger from the barrel. What I should have said was uncouple the gearbox from the auger. Run the motor with the gearbox over the speed range while monitoring torque. The results will give you data to compare to the other motor/ gearbox. What part of the country are you located in? Maybe I can suggest someone local to you that can assist.

 

Sorry for my bad grammar. I didn’t mean literally remove the auger from the barrel. What I should have said was uncouple the gearbox from the auger. Run the motor with the gearbox over the speed range while monitoring torque. The results will give you data to compare to the other motor/ gearbox. What part of the country are you located in? Maybe I can suggest someone local to you that can assist.

I am located in Arkansas. Here's another kicker, we have two almost identical extrusion lines next to this one, that both have these crammer systems. One has an almost identical motor and gearbox (same hp, same gear ratio, same manufacturere, jsut a different manufacturer date) while the other has an even smaller motor and they purr. They never break 6-7% torque under full load.

So, I did do something similar to this, but i left the auger and gearbox attached and plotted the torque response over various speeds. While I must err on the side of caution and not share the plots, I can tell you that nominal average torque values at varying speeds stayed pretty stable, but the faster we pushed the motor did the responses become less noisy. At very slow speeds, like .1-.6 rpms did we see a deviation in torque between 1-10%

 

problem can be caused by different things. cable shielding/grounding/fbk-pwr separation/correct termination etc. are one type of problem. wrong tuneup parameters are another.

for example if P gain is too large, system becomes unstable and oscillates (continuously overcorrects). this can be aggressive leading to increased current and interference. often in such situation motor and gearbox get hot, loud whine can be heard and vibrations can be felt even if motor does not move.

wiring between drive output and motor is the dirty side. this is what generates most EMI. to minimize it, both ends of the motor cable shield need to be firmly grounded.

feedback cable is wiring between encoder and drive. this too need to be shielded AND wiring need to be twisted pair, where complementary signals are on same pair. these cables are usually grounded only at the drive to prevent ground loops because this is circuit is sensitive. if both ends are grounded, then additional grounding conductor (flexible with many strands) is needed from drive to the motor. this creates path of least resistance (up to RF, this is why condutor need to be with many strands) so ground loops will favour it over GND conductor or shield in motor and feedback cables. this should be at least 4..6mm^2 (larger is better).

note, shielded cables need to be shielded. stripping too much shield defeats the function. only remove as little shield as possible. on a motor side connectors are shielded or the cable enters shielded connection box. on a drive side there is no such box, and UVW is usually on terminals. so to terminate motor wires at the drive, shield need to be removed. i try to keep it at 2" (50mm)

encoder cable is shielded throughout but in many cases shield enters connector or strain relief but it is taped back (does not connect to anything on this side). if it is connected to encoder body, then you must have external RF grounding conductor mentioned earlier. the other side of the encoder cable should ideally enter enclosure with servo drive uncut and go to encoder connector on the drive. if this is not the case because drive has terminals for encoder, strip shield up to 2" (50mm) and wire the encoder. shield need to be tied to ground. if that GND connection is not using stud, choose a place on cable about 1-1.5ft from drive, strip only cable jacket about 20mm (3/4") and tie it toa bare metal of the cabinet using clamp or just a ZIP-tie. note ZIPtie should go twice around the cable (forming X). this is more secure, reduces stress on ZIP tie and will not fail when you least expect it. run separate gnd conductor (same as RF ground for encoder) from cabinet to building GND. do not rely on GND in the mains wiring.

if this did not solve the problem, at least you know that grounding / shielding is solid. now you can play with tuning parameters of the drive. many drives have autotune option (so read the manual). if manually tuning, monitor responses (most drives have software that includes oscilloscope). normally start with low P gain and then gradually bring it etc.