. . .I've never understood how power spikes can cause an overvoltage trip on a device with enough bus capacitance to maintain it's logic supply for 15 seconds after the power is disconnected. A voltage spike is the electronic equivalent of a 1-second blast from a pressure washer into a bathtub. That certainly wouldn't overflow the bathtub unless the level were right to the rim already. If the overvoltage is the result of a ground plane shift than you will probably need something a little more involved than a line reactor. You would probably want a power filter.
Keith
The tech at a/b said that the voltage spike might only be microseconds in duration (too fast for my meter to catch) since the first thing it hits is a rectifier which instantly turns it into DC voltage and applies it to the bus. I am with you, though, that the bus should immediately drop that voltage down to an acceptable level. The PF40 book says that the fault level for high DC bus is 810vdc which equates to a line voltage of 575vac. I can't think of what we have that might cause that sort of noise...although there are two old 1336 drives on this same MCC center that don't have line reactors and also never fault for bus volts...
He recommended line reactors as a solution.
I have ruled out all the load size causes. I tried a longer decel time (3.0 seconds) for 24 hours but it did not help. Today, I set the decel rate to 0.5 seconds and tried over 20 times to make the drive fault. I tried short cycles, long cycles...the PLC limits the short cycles (drive must stop for 5 seconds before it can start again in either direction).
Sometimes the drive trips when sitting enabled with no run command (stopped). I left my Fluke on peak hold for 2 hours and the range it picked up was 439 to 459.9 vac, even though the drive fault queue shows that it tripped for high DC Bus volts once during that test.
I can't use one reactor for both drives, since they're installed in MCC buckets and drive equipment located on opposite sides of a production floor.
I still haven't found the line reactors we are supposed to have on the shelf, so if I have to order some, I will try not to oversize them, but it would be nice to stock as few as possible, since the cost difference between a 1hp and a 5hp is only a few dollars.