Testing transformers with a megger 1

[jonathn1]

Recently had a boss tell me to look over a procedure. he had us testing using a megger with red and black leads. We are asked to use the red lead to ground and the black lead to the power connection for each phase.
My recommendation was to switch the leads locations so the black was on ground and red for the power connections. Stating that altho it wouldnt change the test results it did enable good habits for any later tests, since we are not working with electricians doing these tests.
He came back saying a test was done in France on this. They actually proved that changing the leads so the black was on the power, or phase connections, and the red was on the ground connection gave a significantly better reading. it had to do with aligning the electrons and providing a secure field.
Has anyone heard of a test like this or any kind a validity to it?

 

[IRstuff]

Is it April in France already?

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[electricpete]

Your boss isright. If the red and black lead are different polarities of dc voltage, the megger reading can be different. I think it has to do with the way water responds to the different polarities. There are some test sequences that use that behavior to assist diagnosis of the transformer condition. I'm going from memory, don't have any references at the moment.

=====================================
(2B)+(2B)' ?

 

[MacGyverS2000]

Aligning the electrons

Dan - Owner

http://www.Hi-TecDesigns.com

 

[electricpete]

There was some related discussion here:
thread237-90352

One excerpt from the SD Myer's Transformer Maintenance Institute book posted in that thread was:

"The applied voltage should preferably be in the ratio of 1 to 5 or greater (500 and 2500 V, for example) both applied for one minute. Results to date show that a decrease in the insulation resistance of 25% or greater at the higher test voltage is usually due to the presence of excessive moisture. Why is this? Water in an insulating system is polar positive and will be attracted to areas of high negative electrical intensity. Therefore, when a Megger or similar equipment is used the negative lead is attached to the copper and the positive lead to the ground system."

=====================================
(2B)+(2B)' ?

 

[electricpete]

That quote might be interpetted to mean that paying attention to the lead polarity is more relevant for dc step test, depending on your viewpoint. At any rate the relevant point for me is we do expect some slight difference. This applies to oil-filled transformers, and not to dry transformers or motors to my knowledge.

=====================================
(2B)+(2B)' ?

 

[MacGyverS2000]

Pete,

I'd appreciate it if you would explain this one... how would the polarity of water affect measurement in one direction and not the other?

Dan - Owner

http://www.Hi-TecDesigns.com

 

[electricpete]

I found another reference, this one for motors:

http://ewh.ieee.org/cmte/pes/materials/WorkingGroups/P43/P43-2000_Draft_May-08R1.pdf

IEEE 43 - 3.3 electroendosmosis effect : A phenomenon occasionally observed, more often on older windings, when, in the presence of moisture, different insulation resistance values may be obtained when the polarity of the tester leads are reversed. Typically for older wet windings, the insulation resistance for reverse polarity, where the ground lead is connected to the winding and the negative voltage lead to ground, is much higher than for normal polarity.

Note this survived beyond the draft and remains in the 2006 version of IEEE43.

Why does it happen? I lost my original reference, so I don’t have any unique insights on that question. It does not seem unreasonable to me that the diffusion of a polar molecule can be influenced by an electric field (does that seem unreasonable to you... are you looking for atomic type explanation ?). I wouldn’t have any idea of the timescale of the diffusion, but I gather it can be enough to affect results from the above quote and my vague memories of that older reference.


=====================================
(2B)+(2B)' ?

 

[electricpete]

Some more thoughts about "why"

If conductor is negative electrode (and tank/core is positive), then water collects directly around the conductor (configuration A)

If conductor is positive, (and tank/core is negative), then water tends to be move either toward the outside of the insulation (configuration B) or perhaps more uniformly distributed (configuration C) since the negative electrode / ground plane may be less well defined.

So A, B, C all represent different distribution of water within the insulation and we might expect they would all have different insulation resistance.

Any of those explanations seem reasonable to me, but I’m just offering why it seems plausible to me ....and apparently observed directly by other.

=====================================
(2B)+(2B)' ?

 

[MacGyverS2000]

I'm trying to imagine water being pushed in/out of an insulator quickly enough due to the small motive force offered by a megger to affect the measurement. I'm not saying it can't happen, I just can't come up with a logical scenario in which it might...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[electricpete]

Well, I certainly will stipulate that your intuition counts just as much as mine. However we have a number of references above suggest it can be relevant for some (not all) equipment and insulation types. Some more:

http://www.dpi.nsw.gov.au/__data/assets/pdf_file/0006/259071/18---Roger-Devery.pdf

As a result insulation test sets were manufactured with the high voltage output negative with respect to earth so that the worst case reading is obtained. This configuration became very much standard on all insulation testers and high potential test sets and remains the standard configuration for today's insulation measuring instruments

http://www.biddlemegger.com/biddle/5kV-DiagnosticTesting.pdf

With modern insulating materials there is little, if any, difference in the reading obtained, regardless of which way the terminals are connected. However, on older insulation, a little known phenomenon called electroendosmosis causes the lower reading to be obtained with the positive terminal connected to the grounded side of the insulation being tested. If testing an underground cable, the positive terminal would normally be connected to the outside of the cable since this will be grounded by contact with the soil, as shown in Figure 3. Please note that you do not connect directly to the insulation but rather to the cable’s neutral or ground.

=====================================
(2B)+(2B)' ?

 

[IRstuff]

While water is "polar," it, by itself is an extremely poor conductor, as evidenced by the fact that DI water has extremely low conductivity. Conductivity requires IONS to carry the current, not just "polar" molecules.

It might be possible that the presence of water causes dissolution of materials in the insulation that results in conductive ions being in solution with the water. These would result in a measurable reduction in resistance, and if they have work function differences with the electrode materials in the circuit, there might be a noticeable difference in resistance with reversed voltage.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[ScottyUK]

It's possibly worth noting that the cellulose-based Kraft Paper has a significant percentage of water held within it, some of which occurs as the cellulose itself breaks down over time.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jonathn1]

Well all this has been wonderful and i appreciate all your time and knowledge. I noted 2 contradictions that confused the issue even more. IEEE stated that the insulation reading would be higher but the biddle quote stated the reading would be lower. Then again if i am testing and dont like the reading can I change the leads to get a more favorable reading? In the dpi quote it states that the high voltage side is already switched inside the megger. So if I switch the leads I just changed the test to the opposite of what I wanted. Since my testing is on an isolation transformer. I am testing phase to phase not to ground. Does the same rule of moisture apply?
What do you think?

 

[IRstuff]

Sure, why not; it's the same insulation.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[OperaHouse]

This makes sense if you stop looking at it as resistance and instead look of it as a semiconductor. I had heard of this and saw one at the US Air Force Museum, a diode for a crystal radio made from the blue oxide layer of a razor blade (if anyone remembers those). An electrolytic capacitor is another example. The wet electrolytic conducts rather well when the reverse polarity is applied. A photovoltaic cell can be made with a sheet of copper and forming an oxide layer. Think about old selenium rectifiers. An oxide layer formed by any metal is bound to form a diode that will increase the withstand voltage in one direction. Out gassing of some older insulation could easily provide the chemicals needed to form a layer on the nearest frame metal.

The atomic level is messing with us all the time. Hold a clip on a resistor lead on one side and the resistance goes up. Hold the other side and the resistance goes down. That thermoelectric effect is easily seen on any meter over 3 1/2 digits.

 

[MacGyverS2000]

electroendosmosis

I couldn't figure out why this term seemed so familiar, yet didn't quite fit for me, and the it hit me. This is a term used quite often in biology for running electrophoresis plates (done a few of those in my time). It's used to separate chemicals (and DNA), and you can "watch" these different stained biological elements climb a plate at different speeds (over a period of an hour or two, so it's not fast by any means).

But I still fail to see how that would cause a difference in measurement from one direction to the other. Seems to me if it's applied in one direction for 5 seconds, then swapped around for another 5, the measurements should be the same.

Sorry guys, but it just doesn't make sense to me. Now I'm wondering if electroendosmosis is what is really to blame, or if someone chose the wrong word to describe what's going on.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[electricpete]

While water is "polar," it, by itself is an extremely poor conductor, as evidenced by the fact that DI water has extremely low conductivity. Conductivity requires IONS to carry the current, not just "polar" molecules.

As a point of clarification, the reason I brought up the polar nature of water was to explain their apparent tendency to move in the presence of a field, not to explain why organic insulation which is moist has lower insulation resistance.

IEEE stated that the insulation reading would be higher

I think all the references including IEEE are consistent in suggesting that you get a lower reading when applying negative to the conductor and positive to ground. I suspect what may be a point of confusion may the word “normal polarity.” in paragraph 3.3. IEEE is using the word normal polarity to mean negative to conductor. (See section 5.5: ”Insulation resistance tests are usually conducted at constant direct voltages of 500–10,000 V having negative polarity”)

Then again if i am testing and dont like the reading can I change the leads to get a more favorable reading?

That’s up to you and your conscience. (I’d think the difference probably won’t be big except for a very wet winding with paper or asphalt insulation). The logic of the test equipment industry’s approach seems to be to always test it the same way so valid comparisons can be made. One could argue the acceptance criteria are based on the standard polarity (although really in most cases we shouldn’t feel that much different about an insulation resistance reading 10% above an acceptance reading than 10% below... the acceptance criteria are somewhat aribtrary binary cutoff in a continuous world).

Since my testing is on an isolation transformer. I am testing phase to phase not to ground. Does the same rule of moisture apply?

The rule was geared to testing to ground. If both windings are floating I’m not sure how you could decide which winding to play the role of ground and which one to play the role of conductor. I would tend to say it does not apply, or at least not worth worrying about for that special situation.


=====================================
(2B)+(2B)' ?

 

[electricpete]

For Doble members, there is a very detailed discussion of this topic in Doble’s “Dielectric Theory and Practice”, document “72A-1974-01 Rev. B 3/05.” Also it seems the “Evershed Effect.” is a related phenomenon.

=====================================
(2B)+(2B)' ?

 

[jonathn1]

I certinly received more than I expected. I know RTD testers for communications check system impedence not just resistance. Thinking of in those terms simplifies the way I was looking at the test. Certainly makes sense once you put it in Electronic terms rather than electrical. I now understand the reasons more clearly on developing a "NORMAL TEST CRITERIA" and then sticking to it. (is the 1 minute test). The changes in the readings would more than become apparent over time if I had degradation or contamination take place.
Thanks everyone I have the answers I needed this time around. I will certainly look up the information provided and pass on what I learn.
Jonathon