Epoxy losing its insulation?

[Skogsgurra]

A large synchronous motor had its rotor refurbished. The insulation was replaced with new material and epoxy resin was used to impregnate it. The insulation tested OK after curing. Around 100 Megohms.

When delivered and installed, an insulation test was made before energizing and the insulation was down to about 100 kohms or less. Hot air for a day and night didn't help. No extra circuitry present, only the open winding - still around 100 kohm, maybe less.

Don't know what epoxy was used, but these guys have done hundreds of rotors before and done so for at least fifty years. They know their thing. Anyone had this problem before?

Gunnar Englund

http://www.gke.org

 

[rotman]

For the new, epoxy impregnated insulation I would expect much higher resistance value than 100 MOhms, as stated (if measured in cold state) - so it might suggest something wrong in insulation/impregantion process. Maybe the rotor's drying before impregnation was insufficient?

 

[Skogsgurra]

Thanks for the answer.

I actually do not know what the resistance was. It was said to be more than 100 Mohms. It may have been much higher. These guys (one of the better known rewinders in this country) have done this kind of job over and over again and have never had a problem like this before.

Is it possible that insufficient initial drying causes a problem like this?

Gunnar Englund

http://www.gke.org

 

[TheBlacksmith]

On my latest project (3 phase generator stators) we are seeing about 500 Megohms after VPI. 100 seems very low to me too for a brand new modern insulation system.

 

[Skogsgurra]

The initial 100 Mohms are not the problem. The insulation was probably a lot better than that. I shouldn't have mentioned that level. Sorry.

The problem is the lower than 100 kohms measured on site and that it cannot be improved by drying the rotor out. Any experience from that kind of situation?

Gunnar Englund

http://www.gke.org

 

[electricpete]

are you able to split the winding to see if one portion measures lower than the rest?

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

 

[jraef]

That's quite an attrition rate. I think you need to consider a better system by which you encourage the Ohms to stay. Perhaps an increase in minimum wage, or better benefits?

Seriously though, you have apparently eliminated the possibility of hygroscopic intrusion, but what about temperature? Was the temperature of the material the same when both measurements were made? Maybe the epoxy has a negative coefficient of resistance and was cold when they first measured it.

Any chance of measurement error in the first instance, i.e. the probes weren't touching hard enough? I would never accuse you of such a thing, but it appears that you were not there for the initial measurement. ;-)

Other than that, what you are implying is the possibility that as it cured, the electrical properties of the compound changed. That would mean that it was the volatile components that provided the insulation values and as they evaporated in curing, the remainder is more conductive. Or possibly something crystallized in the process, I remember from (long ago) chem labs that some materials change their electrical properties upon crystallization.

Just guesses though. I only added this to lend legitimacy to my irresistible joke in the beginning!

Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[Laplacian]

Did the repair shop pull a sample of epoxy from the source tank?

Some grades and types are hygroscopic while in the tank and required various heating and drying schemes to keep the tank uncontaminated. There may also be some other source of contamination in the source resin tank for the VPI process; perhaps from a previous job

 

[electricpete]

I certainly agree with all previous comments.

I would also keep an open mind to the possibilities of contamination or hidden mechanical damage that occurred after the first test (perhaps during shipment or testing).

Some things to check if not already tried
- Separating the windings and recheck.
- double check selector switches and scales.
- repeat with different instrument.
- very proper test voltage was used.
- was a hi-pot applied at the field? what voltage? (tracking during hi-pot can create tracking path that will lower insulation resistance later check by megger)
- Were proper techniques for discharging winding after test used (slowly ramp down voltage and maintain at zero or discharge thru grounding resistor). Shorting a charged winding directly to ground can create a traveling wave transient which can damage the insualtion.
- contact rotor rewinder for advice

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

 

[stardelta]

I once saw a similar thing with a stator that had been rewound. After VPI the stator was placed in the curing oven WITH another stator that was in there to be dryed out after the winding had been steam-cleaned. It was assumed that the moisture from the dry-out stator found its way into the rewound stator and dramatically lowered the IR before it fully cured.

 

[Skogsgurra]

Thanks all,

It is good to have you available on-line.

No, I wasn't there either first time or second time. A guy just phoned me and asked me if I had any idea. I had not.

The temperature was probably around +20 C in both cases. The rotor had not been started when the second measurement was taken.

No, no sample left.

Jeff, these ohms were bought together with the epoxy from some overseas manufacturer. Difficult to negotiate - they do not speak our language or any language we know. They resist every attempt to communicate.

I was thinking that, perhaps, something can happen with the epoxy after curing - getting bad like some wines do when getting older (do not really mean that, but something similar). They are taking the rotor back to the workshop and will probably have to do the whole thing over again.

Gunnar Englund

http://www.gke.org

 

[waross]

Hi skogsgurra;
I would consider shipping damage or an unreported incident during reinstallation.
It reminds me of a 600 kw generator end we shipped to the US for a cleaning, redip and bake after severe hurricane damage. It was crated in a plywooud box by the repair shop.
It was repaired and returned to us, but when it was being unloaded from the ship, the lifting slings were placed around each end of the box. When the generator dropped out through the bottom of the box, they considered themselves lucky that it didn't hole the old freighter. It did shift the copper strip windings, and a complete rewind was required.
Some inquiries as to the shipping and handling methods of your machine may be in order.
Respectfully

 

[itsmoked]

I'd have liked to have seen that .. I would agree too with your suggestion.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Skogsgurra]

Bill and Keith,

That's a whole new idea. Might very well have happened. I will fwd the thinking to the winder.

Gunnar Englund

http://www.gke.org

 

[IRstuff]

Never assume that the shipper didn't drop or otherwise damage the shipment.

We shipped an electro-optical tester to a Navy shipyard for installation into an aircraft carrier. When installed in the test bay, they couldn't get the system to perform calibration alignment. Everyone swore that it hand not been mishandled.

Shipped the system back to the factory and the first thing we noticed after unpacking the crate was that the 2-inch aluminum base plate was bent up in one corner.

TTFN

 

[itsmoked]

You guys shouldn't have used such a big hammer when you packed it!!

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[electricpete]

Um... not quite a new idea if you look at my 6/23/06 13:31.

"I would also keep an open mind to the possibilities of contamination or hidden mechanical damage that occurred after the first test (perhaps during shipment or testing)."

But that's ok, whose keeping track ;-)

We have had two large motors that have failed (stator ground fault in one case and turn/ground in the other) within a very short time after shipping. Neither one immediately after shipping. Both tested very dry. One operated several hours before failure.

The cause was never completely determined (there was also work done on one of these motors prior to shiping) but the coincidence of timing leads us to keep an eye on shipping. We now require an impact recorder on all our large motors during shipping. It's something that Westinghouse does for Reactor Coolant Pump motors and transformer customer/shops usually do... but other than that not a common practice for motors (invariably our shops tell us they don't know where to get an impact recorder because they have never had it requested before).

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

 

[IRstuff]

Never said was a new idea, considering that Waross made a similar comment, not that anyone's keeping track ;-)


Just offering some evidence to back those assertions.

When confronted with the evidence, the Navy said, "Oh yeah, it might have fallen off the forklift."

TTFN

 

[electricpete]

I was responding lightheartedly to skogsurra who said "that's a whole new idea"



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

 

[jraef]

Tangential thought alert!

In a prior life we used to use a simple device called a "Tip-N-Tell" on the sides of MCC crates because Allen Bradley MCCs could get critical damage to the bus support system if the units were tilted more than a few degrees from vertical. These devices were simply a couple of angled plastic 1/2 tubes with a reservoir of blue glass beads and sticky glue on a cardboard placcard that was stuck to the crate on 2 sides. If the crate was tipped, the blue beads traveled up the tubes and got stuck in the glue, providing a permanent indicator that the crate was mishandled.

It made me wonder if there is something like that for indicating shock damage as well, and there is. Could be very useful for large electrical machines.

They also have a version that is reusable (keyed reset) and even one that records the time and date of the shock!

Shockwatch link

Eng-Tips: Help for your job, not for your homework Read faq731-376