PowerFlex400 Application Questions

Recently I was called into a plant as a consultant on a machine that was experiencing a large number of drive failures. All eight failures were on 250hp normal duty PowerFlex400 Allen-Bradley drives. The system was powered by a 460V floating delta power source.

I was advised by the Allen-Bradley techs on-site that their standard procedure on floating delta power was to disconnect and remove the input MOV's on the drive input terminals. Apparently, if they do not do that, under imbalance-to-ground conditions, one or two of the MOV's will heat up and destroy itself. (The MOV's must be rated at lower than 460VAC)

I also found that these eight failed drives had seen an intermittent ground on the motor leads which had destroyed most often an input SCR but occasionally an output IGBT.

In none of the eight drives were any faults logged or any apparent attempt made to protect the drive's power circuit components.

I have two questions: first, is it really standard AB practice to remove the input MOV's on floating or unbalanced power? Second, on floating power are the output ground fault and short circuit protections disabled or inoperative?

I did a small job with pf40 an ev I had to rem ok ve thay jumper on the board because the integrator before me smoked 4 of them. He said it didn't say I couldn't use them on a flaring system.
Don't know about the second question.

"To prevent drive damage, the MOVs connected to ground shall be
disconnected if the drive is installed on an ungrounded distribution
system where the line-to-ground voltages on any phase could exceed
125% of the nominal line-to-line voltage. To disconnect these devices,​

remove the jumper shown in Figure 1.6."

Nothing is saying that ground fault or short circuit faults become unoperative
PaulB

DickDv will usually chime in with a solution. Maybe put an ATTN: DickDv in the subject line. Wait, What?

Yes, blues, it's me! There's plenty out there that I'm unfamiliar with or don't understand. Every day, I get reminded of that. It motivates me to want to help others with questions.

But, in this case, I am a bit amazed that a responsible supplier like AB would remove all transient voltage protection on the network that has the worst transient problems and not replace it with protection rated for 460VAC. Why use MOV's rated at less than 460VAC at all?

And, how can you, with a straight face, market a drive that has no functioning output short circuit or ground fault protection on floating systems? Even the fault log is incapacitated so no record of the event is captured!

I was so surprised at my findings at this facility that I wanted to pass it by some of the AB folks on here to see if this was a local anomoly or whether this truly was the "official" policy of the company.

So far, I'm not liking the answer I'm getting!

Just out of sheer curiousness and ignorance, how would you do ground fault detection on an ungrounded system? I'm familiar with the 3 pilot light approach but not how a drive would look at it.

In VFD's, generally the current in the + and - legs of the DC bus are compared and, if out of balance by a set small value, the fault is activated. On ungrounded, unbalanced, or high resistance systems, this still works but, due to sudden shifts in balance on the supply side causing current to flow into and out of the motor leads, there is a tendency for nuisance faults. For this reason, the ground fault can be turned off in the software and often is.

Output short circuit fault or overcurrent fault is similar but is activated, not by imbalance but by the absolute value of the current in any of the motor leads. This is critical for protection of power circuit components in the VFD.

Floating networks are often thought of as having no reference to ground but this is not strictly true. Each phase conductor has at least some capacitance to ground and on large networks this can be fairly large. When a ground fault occurs on the drive's motor leads, this capacitance attempts to discharge thru the drive into the grounded motor lead. That's the current that is potentially damaging. In the incident I mention above, these were 320amp rated drives and all eight failed drives lost either an IGBT or SCR.

In addition, with no input MOV protection, there is no control over where the network voltage to ground can go. Quick ground faults or other disturbances in the network can cause the phase voltage to ground to spike and fly-back. We observed voltage spikes in excess of 1400VDC in 48hrs of monitoring. We also saw DC offset voltages in the network, first about 400VDC negative and then bleeding away over 5 minutes time and then 400VDC positive and bleeding away over 5 minutes.

So, while VFD's on floating networks, in my view, are never a good idea, I'm still interested in what AB's practice is and whether I was seeing things with the "no faults" issue.

As for the troubled application, we installed drive isolation transformers with grounded wye secondaries, 460VAC rated MOV's phase to ground (750VDC clamping voltage), and cleared up the source for the output ground faults. They have now run a week with no further issues.