Overvoltage Fault caused by regeneration on acceleration

[milldrone]

Originally Posted by bfitz

My thoughts are that you could probably do the first level tune (P0.05 set to 01) with the load connected. I believe that this test doesn't actually move the motor (_please_ wait for confirmation on this though).

Unfortunately I'm not at work so I cannot confirm if the #1 auto tune does not move the motor.

I know for a fact that the ABB ACS800 has two different autotunes. One with no load connected to the motor, this one is needed for ultra low speed (1 & 2 hz) response. The other does not move the motor, and can be used for most other non critical control. But this does nothing to help you.

ABB reps have told me that their sensorless vector is the best (can you imagine that!) and that some "sensorless vector drives" are not much more than torque boost. I cannot confirm either way. I suspect there is some truth here.

I believe that Brian is right (except for one little caveat)

Originally Posted by bfitz

This is critical. You are changing the load on the drive. A V/Hz drive will just keep putting out voltage at the specified frequency without knowing that the motor slip is decreasing rapidly. An S-curve will help you approach a set speed with minimal drama, but the situation you referenced is encountered at a steady state. S-curves won't help you here.

I believe that some drives (low end sensorless vector) actually increases the frequency in an attempt to keep the motor speed the same as the desired speed. When the load is gone suddenly and the speed is "high" there is a potential for overvoltage. Because the drive will then try to reduce the frequency to keep the motor speed equal to the desired speed.

You mentioned earlier that this was a high inertia load. Did you make any adjustments in the P 2.00 to P 2.05 area?

[bfitz]

Originally Posted by milldrone

I believe that some drives (low end sensorless vector) actually increases the frequency in an attempt to keep the motor speed the same as the desired speed. When the load is gone suddenly and the speed is "high" there is a potential for overvoltage. Because the drive will then try to reduce the frequency to keep the motor speed equal to the desired speed.

I'm not an expert on VFDs, but I believe what you are talking about is called slip compensation. The frequency output can vary to keep the motor speed equal to the desired speed regardless of load on the motor, or slip. Yes, when the load is gone suddenly, the motor has more torque than necessary and slip starts to decrease and the extra energy is put into speeding up the connected load. However, I would wager that a sensorless vector VFD would be able to detect this increase in speed and compensate by lowering the output frequency almost instantly. This should prevent an overvoltage by limiting the amount of extra energy stored in the rotating mass in the form of overspeed. In my mind, this is exactly the purpose of slip compensation; well, that and minimizing speed errors in general.

Brian

[milldrone]

Originally Posted by bfitz

I'm not an expert on VFDs, but I believe what you are talking about is called slip compensation. The frequency output can vary to keep the motor speed equal to the desired speed regardless of load on the motor, or slip. Yes, when the load is gone suddenly, the motor has more torque than necessary and slip starts to decrease and the extra energy is put into speeding up the connected load. However, I would wager that a sensorless vector VFD would be able to detect this increase in speed and compensate by lowering the output frequency almost instantly.

Yes you are correct. This is merely slip compensation. This is also ABB's point (it's merely slip compensation) and the "almost instantly" part also.

Check out what the drive does to the defaults on P 2.01 (slip compensation page 4-25) when in vector mode.

Edit: I'm hoping to have some time this afternoon to try the #1 auto tune on one of my units.

[milldrone]

bblank,

I just tried a #1 auto tune on a early firmware version GS3. It does not move the motor and takes about 1/2 second to perform the test.

To perform a #1 auto tune on my only current firmware GS3 I would have to put on my fall protection gear, and separate an integral motor from it's gearbox. I'm not wanting to do that!