In article ,
Ignoramus31221 wrote:

On 2008-07-18, Joseph Gwinn wrote:
In article ,
Ignoramus31221 wrote:

On 2008-07-18, Wild_Bill wrote:
You don't quantify the amount of motor insulation leakage current, or
describe how it is that you became aware of it.

I measured it yesterday with an clamp on ammeter, it came out as 9
amps (!). Really a shocking amount.

Nine amps?...!! If that's true, the motor is toast, and should be
replaced. Immediately. When we talk of leakage, we mean a few
milliamps.

Probably a measurement error due to spikes.

No leakage without VFD.

How did you test this?


The VFD is not causing this.

Unless the clamp-on is confused by the high-frequency output of the
inverter drive.

That's my feeling.

OK. When I say that the VFD is not causing this, I mean that ordinary
high-frequency current coming through the capacitance between windings
and frame are not nearly that large, by orders of magnitude.


One way to tell is to hook all three windings of the motor together and
connect them to 110 volts through a light bulb (not through the VFD),
and measure the current to ground. In a perfect motor, the current will
be zero. (The motor will not even try to run, or be in any danger, and
single-phase is what to use.)

Do not use an ordinary ohmmeter for this. The test must be made with
full power voltage, or more. Or a megger, if you have one:
http://en.wikipedia.org/wiki/Megger.

I think that I have a megger.

I suspected as much. Don't know why.


If the current is in the amps, at least one winding is shorted to the
frame, and the motor must be replaced or rewound.

This is *exactly* why one firmly grounds the frame of a machine tool.

I think that the motor is not leaking 9 amps and it was a measurement
issue. I will try using a analog clampon next time. But there is
leakage of some, unspecified amount.

An analog meter should be less affected by the high-frequency noise, but
analog electronic meters (I have a few) are affected by HF noise.
Passive analog meters should be immune.

Some kinds of high-end DMM are designed to make accurate measurements of
true RMS currents and voltages in inverter-driven motors. I know Fluke
makes such meters, and I imagine that they are not alone. I think the
key is "true RMS" and a RMS bandwidth that exceeds the chop frequency by
a factor. But I have not tried this yet.

You can also open the green wire and put a passive analog AC ammeter in
series. Or a power resistor, and look at the waveform with a scope.

I would be tempted to test a few other motors, for comparison, as well.

Joe Gwinn