EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: taydin on October 27, 2019, 07:52:03 am
-
Here is a puzzle for you guys:
There are two pictures below. One is a pair of twisted 0.25 mm2 multi strand wire, the other is a single wire (same type) and I'm measuring insulation resistance on both.
1) Which one will have a higher insulation resistance?
2) Which one will have a higher breakdown voltage?
3) Is the insulation resistance the only factor determining the breakdown voltage (I don't know the answer to this one, but I did measure the above and know the answer)
-
1. Probably the second provided the teeth of the clip are not digging into the insulation.
2. The second, unless you have extra insulation on the end, the spark only has to jump the gap at the end of the twist,
3. The breakdown voltage is determined by the dielectric strength, vs any air path, if you give enough voltage a certain thickness of dialectic will break down.
-
If a spark bridged the gap on the first test, you need to separate the ends, the test is invalid as an insulation test.
The second test is invalid, because the teeth on the test probe may puncture or weaken the insulation at a single point. You should tightly as possible wrap the single conductor in aluminum foil or tape, this will test the insulation over a much larger area. Do not get too close to the bare ends with the foil.
-
In both cases I am referencing insulation not air gap.
I would expect the insulation resistance test would max out each of the ranges on both single & twisted.
Your twisted pair will have the higher breakdown voltage. I would expect it to be about twice that of the single conductor.
-
If a spark bridged the gap on the first test, you need to separate the ends, the test is invalid as an insulation test.
There are no strands protruding off the end, they are well inside the insulation.
The second test is invalid, because the teeth on the test probe may puncture or weaken the insulation at a single point. You should tightly as possible wrap the single conductor in aluminum foil or tape, this will test the insulation over a much larger area. Do not get too close to the bare ends with the foil.
I know of no megger in the market that has teeth in their probes. Think about it, if the probe bites into the insulator, then IT damages the insulator instead of the high voltage :) Kinda defeats the purpose of using a megger. And besites, we are measuring Gigaohms, who is concerned about the nicer contact and a few ohms less resistance that teeth would provide?
-
Ok, the twisted arrangement had an insulation resistance of 70 GΩ and it broke down at about 4700V.
The single cable had an insulation resistance of 885 GΩ and it didn't break down up to 5 kV.
I can see how the twisted wire would have much less resistance, because all the insulation can be thought of as parallel resistances. But it isn't intuitive to me how 2X thickness of insulation breaks down at 4.7 kV, but X thickness survives 5 kV :-//
-
Maybe I was a little too vigorous in the twisting and it thinned out the insulation a little. But still, it is definitely more insulation than the single cable ...
-
Ok, the twisted arrangement had an insulation resistance of 70 GΩ and it broke down at about 4700V.
The single cable had an insulation resistance of 885 GΩ and it didn't break down up to 5 kV.
I can see how the twisted wire would have much less resistance, because all the insulation can be thought of as parallel resistances. But it isn't intuitive to me how 2X thickness of insulation breaks down at 4.7 kV, but X thickness survives 5 kV :-//
So for sure if you separate the exposed ends further away from each other, the result is the same? Although I guess 5kV is less than 1mm or so gap, so it should take more than that to jump across.
-
So for sure if you separate the exposed ends further away from each other, the result is the same? Although I guess 5kV is less than 1mm or so gap, so it should take more than that to jump across.
I did a gap test and 5 kV arced when the tips were between 1 - 2 mm apart. But the strands were all inside the isolation, I had checked that.
Here is another possibility that would explain the result: This isn't a very good quality cable. If the insulation is porous, the twisting might have introduced many air pockets inside it, which will make the breakdown easier.
-
I am going to test some conductors from a telephone cord 26AWG. Will post back with the results.
-
Two layers of insulation does not guarantee that the breakdown voltage is the sum of the individual layers breakdown voltages. For a number of reasons such as material type, dye type, manufacturing differences, voltage dependancy and contamination, the actual resistances of the layers will be different. Since this is a series circuit, the greatest voltage is developed across the highest resistance. eg., let's say one layer is 100 Gohm and the other is 200 Gohm and both are rated for 1 KV. If we apply 1 KV across the stack one layer sees 333 V and the other 667 V. Doubling the applied voltage to 2 KV nets 667 V across one and 1333 V across the other. If it's only rated for 1 KV, then it's being operated outside of its rating. I don't know what the resistance spread is for typical jackets, but I could easily envisage one batch of insulation being so much better than the average and ultimately failing because it took most of the voltage.
I don't know if there's a name for this effect, but it must have something to do with Murphy.
Many wiring standards require that all wires in a cable or raceway have to be rated for the highest voltage experienced by any one of them - unless there are special considerations such as grounded shields around the HV wires.
-
Confusing the issue is that these resistances are not linear. 100 G ohms in series with 200 G ohms will not measure 300 G ohms due to leakage, polarization, and other stuff.
-
Using old (10-15 year) 24 gauge telephone line conductors.
My DC insulation tester is a Fluke 1507 & only goes to 1000VDC. Each scale would just max out. With 1000VDC all samples read 11 Gig Ohms.
Moving on to a 50KVAC Dielectric Test Set. Single conductors wrapped in aluminum tape. The rate of rise was about 1KV per second.
The first sample broke down at 13KV AC RMS.
The second sample at 15KV SC RMS. Leads were longer past the aluminum tape in case the first sample may have failed because there was not enough creep distance.
The third sample failed at 18KV AC RMS. Semi-conduction tape was added at the aluminum tape ends to weaken the e-field at the boundary.
Two conductors twisted together.
First sample failed at 25KV AC RMS. On this sample the loose ends flashed over once & had to be spread apart more.
Second sample failed at 35KV AC RMS.
-
Same tests above were repeated with CAT 5.
Single conductor wrapped in aluminum tape failed at 10KV AC RMS.
Twisted pair failed at 25KV AC RMS.
-
Thanks a lot for the detailed tests, that was very educational!
In fact, I was so worked up when the piss poor cable I had was resisting 5 kV that I thought "would be so nice to have a 15 kV megger now to destroy this thing". But apparently I would have been much more raged to see that even 15 kV doesn't do the business :o
The blue cable I tested that resisted 5 kV has a rating of 250V! This has gotta be assuming AC voltage at worst conditions (high heat, high humidity etc).
-
@SG-1, I had repeated the insulation resistance test after wrapping a piece of that blue wire with aluminum foil, like you suggested.
First I casually wrapped the piece with foil, not caring much for the foil to be tight. The isolation resistance dropped from 885 GΩ to 95 GΩ
Then I tightly wrapped the piece of wire with foil, making sure that it is snug. The isolation resistance remained at 95 GΩ
-
Taydin, I was also amazed at the breakdown voltage. At one point wondering if the test set is working :-BROKE, then flames appeared ! :clap:
On the first test, I started on the 10KV scale, figuring it would be over fast, then had to switch to the 25KV scale, then on to the 50KV scale, WOW, what a show ! :popcorn:
A normal dielectric test voltage for 600V rated wire is either 1500VAC RMS for one minute or 2200VAC RMS for one second.
Thanks, for peaking my interest, it was fun.
I think the biggest two factors in breakdown is probably time & humidity.
As for Mr. Murphy, his legal address is the same as where I work.
-
Everything indicates that AC breaks down an insulator more easily. I tried to confirm this by googling around, but wasn't able to find a good source for this.
Here is a "theory" of mine to explain why AC breaks down insulators more easily. Would love to hear what others think or what others know about this.
For DC, as a result of polarization effect, free electrons in an insulator are drawn towards the poles of the high voltage power source, so the DC voltage need to reach a level where this polarization effect is being countered by the force of the electric field, and this level is quite high. I can confirm that insulation resistance increases with voltage (up to a certain voltage level at least).
But for AC, the polarization effect cannot occur, because the poles of the voltage source keep changing rapidly. So the free electrons exist in a more or less homogenous distribution. This reduces insulation resistance quite a bit and makes break down easier.
-
I have always been told that AC Dielectric Tests are destructive & DC tests do not harm the insulation unless you actually cause an insulation failure. This would seem to be correct because Meggers are used for preventive maintenance programs.
I'm on holiday this week, but will try to look for the IEEE test standard for this when I get back, it may shed some light on the subject.
-
What are you seeing that makes you think the resistance is changing with voltage ?
-
When testing between two wires or one wire & earth, the wire has some small amount of capacitance. Displacement current through the test sample will be continuous with AC. It will stop with the DC when the voltage becomes stable, after ramping up. The AC is always transferring some amount of power through the test sample. If enough power is transferred carbonization occurs & the sample fails.
-
What are you seeing that makes you think the resistance is changing with voltage ?
I did lots of measurements on various items. The resistance increases along with 500V, 1 kV, 2.5 kV test voltages. But at 5 kV seems to break this trend and the resistance is lower.
I remember reading about this somewhere in an application note, but can't find it now :(
-
A look at IEC standard for insulation 60664,
6.1.3.6 DC voltage test
The d.c. voltage test with a test voltage equal to the peak value of the a.c. voltage is not fully equivalent to the a.c. voltage test due to the different withstand characteristics of solid insulation for these types of voltages. However in case of a pure d.c. voltage stress, the d.c. voltage test is appropriate.
"... With regard to creepage distances, the time under voltage stress influences the number of occasions when drying out can result in surface scintillations with energy high enough to entail tracking.
"... With regard to tracking, an insulating material can be roughly characterized according to the damage it suffers from the concentrated release of energy during scintillations when a surface leakage current is interrupted due to the drying-out of the contaminated surface."
I guess you'd have to do more research on scintillation to understand what is going on.
-
What are you seeing that makes you think the resistance is changing with voltage ?
I did lots of measurements on various items. The resistance increases along with 500V, 1 kV, 2.5 kV test voltages. But at 5 kV seems to break this trend and the resistance is lower.
I remember reading about this somewhere in an application note, but can't find it now :(
I see a similar trend with the Fluke 1507, but it is just that the lower voltage tests have a lower end scale or max reading.
-
Scintillation seems to be partial discharge or corona.