I have a client that is in the food industry. The motors they have, 4160 V, open windings, all have flour caked up inside them. Sometimes the vents in the middle of the stator get plugged, and the motor overheats. Recently we had one get so hot it gave up the ghost by shorting two of the form wound coils together.

I have used the pdma machine to test these motors for four years, once a quarter. I quit a year ago. We had failures during the times I was testing them, and I couldn't find anything that led me to believe the motors were going to fail. It was only after they had failed that I went back and kinda said "oh yea, I believe there it is"
I was led to believe that the capacitance value would change over time and indicate dirty windings. I had more luck with taking temperatures of the air coming out of the stator.

Is there something I'm missing with the pdma values? Is it possible to detect dirty windings? is anyone else having success doing this?

Dave

This message has been edited. Last edited by: RRS_Dave,

I don't use that type of test.

A comment on the physics of the situation - if it is just flour, I would assume it is non-conducting and has dielectric constant not much different from that of air - the mere presence of flour is not going to be detectable by electrical test.

Now, if the flour was wet AND formed a conductive mixture AND if that conductive mixture extended from endwindings back to the core where it electrically contacted ground, then it would be expected to change the capacitance to ground slightly.

As a side note, it might be an interesting excercize for someone with a lot of time on their hands to try to measure the characteristics of the flour on the surface of the winding.

The question whether localized thermal insulation aging causes any reliable noticeable to capacitance measurement... my guess would be no, but I'll leave it to others to comment.

Your motor is 4160 Volts. If you have good information how is the winding configured, you can calculate what the capacitance to ground is and what the capacitance turn to turn is. I am attaching a file with the sketch of the slot of a 2500 hp, 4160 V motor. Calculating those capacitances is a very little problem. In the attached file for the 2500 hp, the capacitance to ground is 0.054 micro-Farads per phase and the equivalent capacitance turn to turn is 47 pico-Farads per phase.
The motor that burned out on you is being rewound or was rewound. All the data necessary for the calculations are known (the rewind shop has it). If you supply the data we can calculate the capacitances. You also have the impedances (or inductances) from the past surveys. In other words we can calculate if it were possible to find out that the winding deteriorated by changing the capacitances. We can do it under the watchful eyes of thousands of readers of this site and under the watchful eye of the low voltage equipment manufacturers. Anybody can jump in and correct any error. Why don’t we do it? I bet, that is the thing that you are missing.

In the meantime we have to use the data from my 2500 hp motor. Since the low voltage testers usually use higher frequency, I will supply the calculated data for 200 Hz (All Test Pro frequency).
The reactance of the capacitor 0.054 micro-Farads (with 47 pico-farads in parallel) is Xc=1/wC= 1000000/2*pi*200*0.0543=14663 Ohms.
The motor reactance (or impedance) measured phase to neutral at 200Hz would be 5.33 Ohms. Sorry, you have to take this number for granted. Just a note of explanation: This reactance would give the motor 5.75x nameplate current at the locked rotor and 60 Hz. A very reasonable number.

So now we have 2 impedances in parallel: One is 5.33 Ohms, the other is 14663 Ohms. The capacitive reactance would have to drop at least thousand times to affect the locked rotor impedance. But in fact the changes of the capacitance are very moderate, below 30% (factor of not 1000, but 1.3). That is why nobody uses the capacitance as an indicator of winding insulation condition.
So here is your “missing link”. But, please, do not trust me. Get the relevant data and post it.
It seems to me that people are finally realizing that it may not be them who are wrong. There is something fishy about the low voltage testing! All the above is a classical example that something is wrong, indeed.
jank

turn_to_turn_and_to_ground_capacitance-2.doc (59 Kb, 15 downloads)

I agree with Jan, at such a low test frequency it is hard to detect a moderate change in reactive impedance.

I thought that the test frequency is in the 20 kHZ range. At least nothing prevents from using much higher test frequency, so that a 50% change in capacitance will influence total current to a greater effect.

David

you can trend the change in winding capacitance by measuring impedance across the windings. I am thinking the caked flour will actually cause the windings to overheat due to lack of airflow. capacitance of windings will decrease as the insulation dergrades due to heat. Impedance will increase in one or more phase as the capacitance decreases