Measuring Reflected Wave Voltage at Motor Terminals 1

[toygasm4u]

Greetings,
I am trying to satisfy myself that a VFD/motor combo isn't going to suffer damages induced by CIV. Some notes on the application:

*2HP 460 VFD with 4kHz carrier
*2HP 460V motor
*VFD rated cable @ 250~500' length
*Output reactor installed at the drive

I'm using a Fluke 196C scope-meter in conjunction with a 7000V differential probe. My question is this:

Should I be measuring Vp-p between phases, or Vp to ground? Currently when measuring Vp-p between phases, I'm getting readings in the neighborhood of 2100 volts. I was always of the understanding that the main area of concern with CIV was between phases, due to the possible break-down of phase paper installed between phases. NEMA MG-1 requires that a motor withstand 1200V I think, and our particular motor is rated with a CIV of 1600V. 1600V what? Between phases or from phase to ground?

 

[dpc]

I suspect they are referring to phase-to-phase voltage, but volts are volts, so probably you'd be concerned with voltage from your winding to either phase or ground. One common failure mode is turn-to-turn failure in the same phase. In a random-wound motor, if the first and last turns are in contact with each other, there can be a large differential voltage at this point.

There will be some actual motor experts along shortly, I suspect.

http://ieeexplore.ieee.org/iel5/28/33428/01583837.pdf?arnumber=1583837&htry=7

 

[DickDV]

Given the data in your posting, even without any data, trouble can pretty well be predicted.

At 250-500feet at this low hp, you will definitely need at least a dv/dt filter and possibly a sine filter to properly protect an MG1-P31 rated motor.

 

[electricpete]

What does CIV stand for?

I know the biggest concern as was mentioned above it the turn insulation which is sensitive to rate-of-rise of voltage (not just the peak).

My guess is that votage to ground would be of interest although I'm not positive. I'm pretty sure that is the way that a surge test is conducted. The high frequency components are capacitively coupled to ground.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

CIV = Corona Inception Voltage = voltage when corona starts => when ozone is produced. Ozone attacs insulation and accelerates aging.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[LionelHutz]

Well, line to line you should see the carrier frequency waveform at approximately 650V amplitude. You're seeing that plus a ringing voltage up to 2100V?

If you're seeing the voltage ringing up to 2100V then you've got a problem. I would be really surprized if the motor survives much longer. Install a good dv/dt filter with the goal to get rid of most if not all of the voltage ringing.

 

[dinkelja]

What make and model is the VFD?

 

[DickDV]

It's not going to make much difference what make and model drive is involved.

Given a 2hp motor, a PWM output, a 4khz carrier, an IGBT drive output section, and 250-500 foot motor leads, you are in trouble, period.

In my judgement, given these conditions, motor leads up to 60 feet are ok without any extra. Beyond that, figure on either reactors, dv/dt filters, or sine filters in increasing order of severity.

 

[toygasm4u]

Thanks for the replies so far guys. Our equipment is manufactured to go anywhere in the world, for applications that range from 100VAC 1 phase to 600VAC 3 phase. Our application also requires that motor leads vary in length from 50 to 400 feet. I don't write the rules, I just follow em'! ;-) We've got a drive reactor package that's been working well with lengths up to 500 for 3 years now; so it's a given that it is do-able with just a reactor alone on the output.

When I measure the combo that's been working from phase to phase, I'm seeing surge voltages in the neighborhood of 2kV, which is 200 over the published limit. This makes me think I'm measuring it wrong, considering that the unit's been in service for 3+ years and still operating normally. So I guess the root of my question is: How do I measure RWV properly? Phase to Phase? Or Phase to Ground?

Does a CIV rating of 1600 mean between phases? or from phase to ground? or both? If between phases is the case, then I'm stuck in some parallel dimension where exceeding the limit has been fine for 3+ years.

 

[DickDV]

Phase-to-phase is where the ringing and reflected waveforms are observed. Those are what are normally destructive to motor insulation.

Unfortunately, it's not always that simple and phase-to-ground voltages can be destructive too. I generally think of phase-to-ground voltages being the source of common mode noise which, while not a motor insulation threat, is still a messy problem in buildings with poor or high impedance grounding systems.

As for not having any problems with the long motor lead lengths, you must be buying very good grade motors. Otherwise, the basic ingredients are all there for premature motor insulation failures.

 

[Skogsgurra]

It is physically very difficult to get more than 2000 V from a 650 V source. Ringing/reflection usually have a factor < 2 and even if you measure phase-phase, you should not be able to get more than 650 V + 1350 V, which is quite close to 2000 V. But, the factor is usually not equal to two. A more realistic factor is 1.7 - and that should give you around 1750 V.

So, I can think of two explanations:
1. Your mains voltage is on, or above, upper limit. That could give you the readings mentioned.
2. Either your differential amplifier or probes needs to be HF compensated. A probe that peaks when measuring a square wave will definitely show too high a value when measuring high speed signals (less than microseconds, I presume).

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[toygasm4u]

see that's been confusing me too. I wasn't expecting to see more than about 3x buss voltage.

 

[Drivesrock]

To be able to see the dV/dt pulses on the PWM of the drive output you must have a scope and probes with a high frequency. I am using 75MHz probes with a 100MHz scope.

Drivesrock.