ROBICON VFD Grounding faults 2

[toto2002]

Hi everybody,

A client of mine has Robicon VFD of the Perfect Harmony Series (PWM) 3000 V, powering 800 Hp, squirel cage star connected motor.

In each phase Robicon uses a complicated combination of 5 phase-shifting transformers (delta secondary 480 V), each connected to 6 pulse rectifier, which in turn, is connected to the transistor block. Each step can generate +/- 600 V.
In order to model a sinusuidal wave form with peak of about 3000 V the steps connect in series to each other in different combination.

The problem that we are experiencing is that the filtering capacitors after the rectifier burst ( explode). We did some measurement and it seems that the whole system is floating to ground. This means that the potential on the capacitor can go up to 3000 V.

Robicon says that this is not possible, but according to the schematic they gave the client - it is exactly what would happen. The reason is that there is absolutely no grounding connection after the rectifiers. There is also no grounding of the neutral.

Is that normal? Has anybody else experienced the same problem with Robicon VFDs, and if so how was it solved.

Thanks and regards
Ted

 

[busbar]

If you are able to determine the capacitor manufacturer and ratings, you may want to search their application literature {don’t forget temperature rating} and then pose the question directly to them.

Seems like one point in the DC circuit would at least be resistor grounded to damp some degree of what appear to be the simplest of common-mode transient effects—especially given all the constant repetitive switching.

 

[electricuwe]

The approach robicon uses for large VFDs is a solution with a lot of advantages, however it has to use several ungrounded rectifier-dc-link-inverter modules floating against ground. One essential requirement is, that every component of the floating circuits is able to withstand a much higher voltage for isolation to ground than the rated voltage of active terminals of the part. This may be achieved either by isolation in the component itself or by mounting it on an isolated base. I'm sure Robicon designs their converters this way.

There may be voltages far in excess of 3000 V due to faults and this could be the reason for the failure, but such faults could also endanger an inverter in a more common topology.

 

[toto2002]

Thanks all for the responses.
One thing I forgot to mention is that this capacitor problem occurs not very frequently ( like once in every 2-3 months).

It happens when the motor is starting. The cpacitors are the ones that Robicon supplied.

I am pretty sure that there is some voltage spike, but cannot understand why this is not happening more frequently. Maybe there is fatigue of the insulation to ground?!
Also it seems that this problem is present only with our client and the only difference that I have identified between this installation and the normal one is that the transformers are in different room from the rectifiers (about 20 feet away). I don't see how this may have an impact but still ... ???

 

[johnnyw]

Hope this helps. I'd be interested in knowing what the actual cause/solution is.

The capacitor may be overvoltaged between the two terminals either by a line surge/spike, motor regeneration, or possibly a resonance due to the transformer location. Regeneration seems unlikely in this situation and line surge/spike doesn't seem to match up with when the failure occurs - motor starting only. Did an input vacuum breaker trip previously to the motor start, which may have caused a voltage spike prior to the startup? What is the voltage ratings for the diodes and IGBTs on the DC link and why don't they fail if similarly rated? They are being stressed too if this is the capacitor failure mode. Is the drive supposed to self-protect for high DC link voltages?

If capacitors are in series, they may not be sharing the voltage properly due to bad balancing resistors or capacitor tolerance mismatch. You can measure the voltage among the series capacitors to inspect for this.

The capacitor case insulation may be failing if the voltage on either of the terminals exceeds the isolation voltage of the capacitor which could happen if the power circuit floats too high above ground. You should be able to inspect the capacitor case for this type of failure. Also, this failure would likely result in the same power module failing, or a power module failure at either end of any of the phases, since these are the furthest from ground potential. Where did the failures occur? Have you checked the ground insulation of the system - hipot the motor and drive separately (not to exceed vendors recommendation).

The capacitor may be overheating due to large current draw through it upon motor startup. What amount of overload current is being drawn on motor startup? Is there a capacitor bank pre-charge circuit? Maybe the capacitors are being stressed (overvoltage/current) when the drive energizes. Is the drive supposed to self-protect from overload currents.

You are sure the failure is not an IGBT shoot-thru?

Put a scope on the DC link of the power module in the failed location and monitor the voltage during drive energization, motor startup, motor operation, shutdown, and/or typical faults (if any). You may have to zoom in on the time-scale to pick up any transients (i.e., IGBT turn-off overshoot). Compare this with a known good power module. Also, place a current probe on the phase with the failed module and measure current on motor startup. How much are the capacitors de-rated in voltage and current?

 

[pwyane]

We experienced the same situation with a large Robicon drive we have on site. The capacitors kept blowing up every so often. Finally Robicon admitted that they had been using poor quality capacitors and they were not holding up. The Robicon rep just finished changing all the caps today and we will see how it goes.

 

[jbartos]

Suggestion: With excellent testing, measuring and recording devices available on the market, it will not be difficult to analyze the root cause of the problem. As above posting indicates, there are many causes under the consideration.
One area not addressed yet, are harmonics superimposed to dc. They need to be observed too.

 

[jbartos]

Question: Have Robicon MTBF, MTTR, and life-cycle information been available for the Perfect Harmony Series?

 

[ScottyUK]

When you replace your capacitors, are you replacing all the caps in the failed module? You may find that, following loss of one capacitor, the others in the group are damaged too even if they have not failed outright. I've found it to be false economy to replace only the failed part.

I don't know where you are - I guess the US - but a European capacitor range I would highly recommend is Evox-Rifa's PEH-169 series of electrolytics. I've found them to have excellent service life under pretty demanding conditions.

We replaced another European manufacturer's capacitors with PEH-169 equivalents. The 3-month lifespan of the original type has increased to two years and is still going. Component cost increase was about 25%.

 

[jbartos]

Suggestion: Visit

http://www.robicon.com/news/news_igbt.htm

for:
The power cells and their secondaries are insulated from each other and from ground for 5 kV class service.
///The insulation of power cells from ground reduces potential paths for unwanted currents to flow.\\\