bafco
Electrical
- Dec 22, 2004
- 9
Hello all,
I've been visiting a facility over the last two days trying to troubleshoot a problem with thirteen 460V 2HP VFDs; package class Deltas. These are very compact, but they do have quite the feature set. When I was first notified of the problem, I was informed that on multiple occasions all of the drives would trip at the same time. First blush is a mains supply issue; transients courtesy of equipment within the facility or a neighboring facility. After personally visiting however, I've found nothing.
What to know:
600' x 1500' facility, typical distribution with 12,470/480 service @ 1200A (estimated). 480 sub feeds to multiple 480 panels with 30 to 45 kVA pad mounts for 208/120. All transformer secondaries are grounded and voltages are balanced.
These 13 drives are all fed from different sub panel locations with the exception of 2 which share a panel.
Each drive is supplied via a dedicated circuit breaker properly sized at 10A per manufacturer's recommendations.
Distance between drive and breaker averages 200 to 250 feet, #12 THHN. Distance between drive and motor is 2 to 3 feet, same #12 THHN.
Everything is solidly grounded.
Drives are rated for 4.2A output, with actual motor loads running at 2.7 to 3.2 depending on location. All motor phases are balanced. All incoming AC voltages are balanced phase to phase and phase to ground.
DC bus levels remain steady be it at idle or at full motor load, centered around 665 (all perfect so far...)
IGBT temperatures are all well below limits according to the HIM modules. (It's been in the mid-90s here)
Yesterday, we verified all drive parameters were sound per the application; no aggresive acceleration, controlled deceleration to stop. DC bus never spiked above 715 when stopping, but that's not when they're having the problem.
Three of the drives were given %5 line reactors, located within 8" of the drives.
Still, suspect of mains supply transients. These drives have a published trip at 810 DC bus or 575VAC equivalent. So, on one of the drive's circuit breakers, I placed a fluke 43B and set it up for 20% transient capture. Throughout the day yesterday, I only captured three events which were each less than 2ms, very low amplitude relative to the normal AC sinewave. The drives all rode through these events while we were there. Before we left for the evening, we cleared the meter memory and started a transient capture again, leaving it there overnight. Went back this morning and was told that all of the drives tripped again at around 8:30AM, even the three drives with line reactors. Checked the meter; nothing captured. Not a one. Before we arrived the supervisors had reset most of the fans, including the one that we were metering. All fans same story, over voltage.
I would have thought that an event large enough to trip multiple drives across multiple sub panel locations would present itself clearly, even for the meager 43B.
So far, I'm just plain stumped. One thing that I'm suspect of, for no real good reason, is the lighting controllers. The facility is lit with these little T5 4 bulb fixtures. The lights are controlled by Lutron remote lighting contactors. These contactors are all co-located to the 480 sub panels. To test them out, we made sure that all the drives were running, then proceeded to start cycling lights on and off. Nothing seen visually, nothing captured on the meter. We also suspected some exhaust fans in the facility, whose feeders are in the same conduits as the drives; these fans are just on motor starters however, and again, no ill effects on the drives when started and stopped. The drives run all day (usually), the exhaust fans run all day, the lights run all day.
Another kicker: sometimes, only two or three drives trip on over voltage, at random times, at random locations.
Could this be a harmonics issue? Could my 43B have missed a cap bank switching upstream somewhere? The line reactors tell me that whatever is happening is in excess of a few ms, but I'm somehow missing it.
What I don't know is if the drives trip while at idle. I've seen drives with load reactors give a false over voltage trip due to output shorts or over current conditions; rapid output shut down spiked the drives backwards through the IGBTs because of the field collapse on the reactor. Removal of the reactor revealed the true fault. Could this somehow be happening to these drives on the input side?
What if I just oversized the drives? Larger DC bus caps may extend ride through times and provide additional buffer?
I didn't see any HF currents on the grounds that I would call excessive. These drives include EMI filters. The MOVs to ground can also be isolated. Not sure what to try next here guys. Could this be a nuisance trip in the HF realm on a periodic basis? What about when they all trip at once?
Reactors were our cheap fix attempt. Isolation transformers at 13 different locations aren't so cheap. New drives? That lands somewhere in between. I've thought about cutting the MOV to ground jumpers; theory being that I've got some kind of circulation there that's reflecting on the DC buss.
Maybe I'm just tired.
Thanks for any inside guys,
~Mark
I've been visiting a facility over the last two days trying to troubleshoot a problem with thirteen 460V 2HP VFDs; package class Deltas. These are very compact, but they do have quite the feature set. When I was first notified of the problem, I was informed that on multiple occasions all of the drives would trip at the same time. First blush is a mains supply issue; transients courtesy of equipment within the facility or a neighboring facility. After personally visiting however, I've found nothing.
What to know:
600' x 1500' facility, typical distribution with 12,470/480 service @ 1200A (estimated). 480 sub feeds to multiple 480 panels with 30 to 45 kVA pad mounts for 208/120. All transformer secondaries are grounded and voltages are balanced.
These 13 drives are all fed from different sub panel locations with the exception of 2 which share a panel.
Each drive is supplied via a dedicated circuit breaker properly sized at 10A per manufacturer's recommendations.
Distance between drive and breaker averages 200 to 250 feet, #12 THHN. Distance between drive and motor is 2 to 3 feet, same #12 THHN.
Everything is solidly grounded.
Drives are rated for 4.2A output, with actual motor loads running at 2.7 to 3.2 depending on location. All motor phases are balanced. All incoming AC voltages are balanced phase to phase and phase to ground.
DC bus levels remain steady be it at idle or at full motor load, centered around 665 (all perfect so far...)
IGBT temperatures are all well below limits according to the HIM modules. (It's been in the mid-90s here)
Yesterday, we verified all drive parameters were sound per the application; no aggresive acceleration, controlled deceleration to stop. DC bus never spiked above 715 when stopping, but that's not when they're having the problem.
Three of the drives were given %5 line reactors, located within 8" of the drives.
Still, suspect of mains supply transients. These drives have a published trip at 810 DC bus or 575VAC equivalent. So, on one of the drive's circuit breakers, I placed a fluke 43B and set it up for 20% transient capture. Throughout the day yesterday, I only captured three events which were each less than 2ms, very low amplitude relative to the normal AC sinewave. The drives all rode through these events while we were there. Before we left for the evening, we cleared the meter memory and started a transient capture again, leaving it there overnight. Went back this morning and was told that all of the drives tripped again at around 8:30AM, even the three drives with line reactors. Checked the meter; nothing captured. Not a one. Before we arrived the supervisors had reset most of the fans, including the one that we were metering. All fans same story, over voltage.
I would have thought that an event large enough to trip multiple drives across multiple sub panel locations would present itself clearly, even for the meager 43B.
So far, I'm just plain stumped. One thing that I'm suspect of, for no real good reason, is the lighting controllers. The facility is lit with these little T5 4 bulb fixtures. The lights are controlled by Lutron remote lighting contactors. These contactors are all co-located to the 480 sub panels. To test them out, we made sure that all the drives were running, then proceeded to start cycling lights on and off. Nothing seen visually, nothing captured on the meter. We also suspected some exhaust fans in the facility, whose feeders are in the same conduits as the drives; these fans are just on motor starters however, and again, no ill effects on the drives when started and stopped. The drives run all day (usually), the exhaust fans run all day, the lights run all day.
Another kicker: sometimes, only two or three drives trip on over voltage, at random times, at random locations.
Could this be a harmonics issue? Could my 43B have missed a cap bank switching upstream somewhere? The line reactors tell me that whatever is happening is in excess of a few ms, but I'm somehow missing it.
What I don't know is if the drives trip while at idle. I've seen drives with load reactors give a false over voltage trip due to output shorts or over current conditions; rapid output shut down spiked the drives backwards through the IGBTs because of the field collapse on the reactor. Removal of the reactor revealed the true fault. Could this somehow be happening to these drives on the input side?
What if I just oversized the drives? Larger DC bus caps may extend ride through times and provide additional buffer?
I didn't see any HF currents on the grounds that I would call excessive. These drives include EMI filters. The MOVs to ground can also be isolated. Not sure what to try next here guys. Could this be a nuisance trip in the HF realm on a periodic basis? What about when they all trip at once?
Reactors were our cheap fix attempt. Isolation transformers at 13 different locations aren't so cheap. New drives? That lands somewhere in between. I've thought about cutting the MOV to ground jumpers; theory being that I've got some kind of circulation there that's reflecting on the DC buss.
Maybe I'm just tired.
Thanks for any inside guys,
~Mark
