VFD motor leads and induced voltage.

[Mporter4]

Application:
I have two pumps, a 20 Hp and a 10Hp 3/ph 460v in the same conduit.
The Vfd's are located in each mcc bucket.

I have been getting overcurrent faults on one of the Vfd's every now and then.
This just started this week, these drives have been in service for years.
So far I have taken the following steps:
Meg motor and all phases are about 7.6 Meg ohms to GND.
Line voltage is 479 phase to phase, 280 phase to GND.
With the vfd running I used a true Rms amp probe and saw A=8.3 B=6.9 C=9.3.
also at the same time I measured the voltage with a DMM and saw A-b= 320v
A-C= 327 B-C = 320
A-GND = 212
B-GND = 204
C-GND = 212
also I found that the ground does not come back to the vfd but to the common GND on the mcc and from there to the vfd.
Approximately 125ft of wire.
I replaced the drive and while I was doing this I noticed some arcing on the motor leads when I was disconnecting the leads. I assumed this was induced voltage so I held it to ground and the arcing never stopped.
I turned the pump off that was ran in the same conduit and the voltage went away (20vac).
I realize this is not the correct way to wire a vfd and that running the leads in the same conduit with no reactors or shielded cable is not ideal, but I found it strange that when I grounded the leads out that the induced voltage would not go away for at least a short time.
Also could this be the reason my original VFD failed?

 

[OkiePC]

It sounds like you have a ground fault (perforated insulation in the pipe or elbow somewhere) that is not being detected by the megger as a dead short. 7.6 megohms to ground is relatively low.

I like to megger motor circuits from the VFD out through any (hopefully undisturbed) closed disconnects and connected motor. Never apply megger voltage to a VFD. Use at least the 500v setting and the 1000v scale if available. I support a shedload of AC drives wired through pipe in groups using THWN with multiple grounds to a wet production floor, so I have to use the 500 volt scale and I expect to have to replace sections of motor wiring every few years. I have not yet sold the boss on VFD cable, but its on my short list of things to do. Anyway, don't be surprised if you get radically different readings on 600v insulation with 1000v megger but the wire can experience up to 1400v peak to peak voltages hence the call for good VFD cable.

Measure leg to leg resistance while you have the leads disconnected from the drive. If your megger has an Ohms scale, use it in favor of a regular meter. My Fluke electronic megger at work has a much better ohmmeter than most DVOMs.

Your true RMS meter is still not going to be accurate. The best way I have found to externally measure current from a VFD is with a Load Controls UPC. It takes current and voltage readings simultaneously and puts out a 4-20mA scaled signal. I think you can get a unit that detects current imbalance.

I have some Toshiba drives (75 and 100HP) that will not run if the megger reading falls below 120 MegOhms. I have some old VFDs that will still putter along with as little as 2 megohms to ground. You want infinty to ground, and a balanced ohm reading leg to leg. The VFD display should be more accurate than a clamp on meter for current readings as a general rule. I've seen some drives with certain set that seem to allow a clamp on to work okay, but most of the time they are "off" by too big a percentage to be useful.

 

[DamianInRochester]

If it truly induced energy it is not simply going to go away just because your grounding the wires.

Sounds like the wires to both pumps are closely coupled and tangled with each other for a great distance. No, the original installation was definitely not done well.

I would be tempted to at least pull the wires back and twist the three wires of each motor together before running them back in the conduit. You can't over do it though or else you will shorten them. You should find that just this small change will greatly reduce the amount of inductive coupling between the two pumps. The twisting will also reduce the RF given off as well.

Even better, gut it and do it right. Get some good VFD cable and be done with it.

Your original drive could have failed due to excessive standing waves, high output terminal voltages, and corona discharge. You could actually be getting corona discharge from a wire of one motor to a wire of the other motor. It might also help to lower the carrier frequency of the drive if possible. If one or the other pumps always runs at base speed I would be tempted to just eliminate the VFD and put a contactor back in. Judging by your measurements that does not appear to be the case.

 

[dmargineau]

Unfortunately, like the previous two posters suggested, it looks like you will have to pull new wiring for both motors.
The culprit is the PWM (Pulse Width Modulation or "carrier frequency") of the VFDs output signals, which, over time, tends to deteriorate the THHN class conductors insulation, especially in the sharp bends of the conduit runs.
We went through quite a few of this situations and, after multiple device replacements, additional grounding and line side supply isolation, the only remedy was the replacement of the field wiring, this time using inverter duty conductors.
I belive there are white papers addressing EMI Mitigation and web posted lab tests which prove that the ozone created by high-frequency magnetic fields will alter the electrical properties of the Teflon coating and the PVC insulation of the THHN class conductors.

 

[DickDV]

First point: It is futile to measure amps or volts in any form on the drive-to-motor leads. I've not seen any device including some hugely expensive ones that will interpret the PWM pulses with any accuracy. Best to use the VFD display for output amps and volts.

Second point: You really have no practical option but to rerun the drive-to-motor leads separately and with new wire. We can talk around it for days and you still will end up fixing the faulty initial installation. Probably best to get started now on running new motor leads.

 

[jmiller78]

First point: It is futile to measure amps or volts in any form on the drive-to-motor leads. I've not seen any device including some hugely expensive ones that will interpret the PWM pulses with any accuracy. Best to use the VFD display for output amps and volts.

I would have to disagree. We use this all the time http://www.wtb.tue.nl/woc/ptc/education/ogo21/Fluke%20voltmeter.pdf

Using a good fluke true rms amp clamp on drive output leads is very close. I have watched many of them compard with the reading on the HIM and it is almost always within .5 amps of the HIM. Most of our drive to motor cable runs are between 50 and 200 feet.

The 190 series 4 channel scope is very accurate and we find bad IGBT's and load issues with it all the time.