Extreme insulation resistance

Go To Last Post
9 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

This is not AVR related, but I am looking for all tips and tricks available to achieve a high insulation resistance between conductors in a circuit for some automatic safety-test equipment. The tester in question measures leakage resistance of a finished product from live/neutral to the chassis using 500V DC.

The problem is that the tester, followed by a relay switching matrix, followed by test leads reads about 500Megaohm when unconnected to the product, and the product must pass a test at several hundreds of MOhm more. Quantifying the "no-load" leakage accurately enough to do a parallel resistors calculation is impossible since the reading at "no-load" varies wildly.

The way I see it is that there are three leakage components in parallel, one in the tester, one in the relays and one in the test leads. The relay matrix and the test leads have at least 1.2GOhm individually as tested by a different device (full scale reading) while together they are somewhere between 800MOhm and 1GOhm. I cannot measure the leakage of the tester directly since I don't want to connect two outputs together....

A dessicant in the enclosure for the switching matrix did not help much.... :cry:

Is there any genius here who feels he's capable of offering a few ideas to a frustrated guy? :x

NxP

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Is there any genius here who feels he's capable of offering a few ideas to a frustrated guy?
Would moving to the space station help? :)

John Samperi

Ampertronics Pty. Ltd.

https://www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You might look here:
http://us.fluke.com/usen/products/categoryinsul.htm

Fluke is one of the leaders in High Voltage testing equipment.

Dranetz is another one. Look at:

http://www.dranetz-bmi.com/products/prodspec.cfm?prod=40

EDIT:
Fluke also has many tutorials and other documents that my provide some further help. Give Fluke's web-site a browse, I think you'll find it interesting.

You can avoid reality, for a while.  But you can't avoid the consequences of reality! - C.W. Livingston

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Unfortunately freight costs to space are slightly high for our admin to approve.... a vacuum chamber with superconducting electromagnet levitated relays to avoid direct contact might be cheaper in the long run... hehe.

I'm looking at the Fluke application notes right now and I'll see what I can gather from that. Unfortunately new equipment can be only a last resort since it would involve investing in a complete redesign of the system (currently based on a Schleich GLP2-i). If percentage volumes of this item are small compared to standard devices it might be more cost effective to do tests with the standard manual testers.

I hope I'll be coming back soon to say something works!

NxP

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You might look into using "guarding" techniques on the high resistance sections of the tester, along with being sure to use relays that allow generous creep paths, etc.

Tom Pappano
Tulsa, Oklahoma

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello NxP,

Note that the Megger brand meter shown here http://www.byramlabs.com/product_info.php/products_id/8467

Shows a 3% variance at 500 volts. I think the relays are the problem and your statement:

Quote:
Quantifying the "no-load" leakage accurately enough to do a parallel resistors calculation is impossible since the reading at "no-load" varies wildly.

IMHO is of no consequence as this value at such a sensitive (meter) setting cannot be relied upon. And although I have done high potential testing using both AC and DC test voltages, I can say that we often wrap any exposed areas in poly and seal them as best as possible. This is more to prevent moisture absorbtion into the cables/test leads. Often we have to leave the tester connected to "dry" out the cable. However, the DC cable testing you mention usually does not follow a formal time sequence or a "stepped" increase in the test voltages that is typical of AC high potential testing. Is this a correct assumption on my part??

Also, testing parameters should include the ambient atmosphere unless this device is going to be on the space shuttle:-) Otherwise you could encapsulate the entire device in an inert gas!

Seriously you might want to try barriers between the HV potentials in different spots to isolate the weakest dielectric. If possible silicone sealant on exposed connections may push the M-Ohm value up to your required value.

John

Just some guy

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

nxp wrote:

.....
The tester in question measures leakage resistance of a finished product from live/neutral to the chassis using 500V DC.
.....
Quantifying the "no-load" leakage accurately enough to do a parallel resistors calculation is impossible since the reading at "no-load" varies wildly.
.....
The relay matrix and the test leads have at least 1.2GOhm individually as tested by a different device (full scale reading) while together they are somewhere between 800MOhm and 1GOhm.
.....
I cannot measure the leakage of the tester directly since I don't want to connect two outputs together.

Greetings N*P,

I'm unclear about your testing methodology.
Are you qualifying the tester at 500VDC?

You may find that the real world leakage
in the tester, relays, and leads, increases
at 500VDC.

I also didn't understand your reluctance to
"connect two outputs together". Can you post
a schematic?

Why do you have a relay matrix? Does the relay
switch in the leakage test in place of another
test (or normal operating) voltages?

You may have little choice but to run the
insulation leakage test on dedicated equipment
that can meet your expectations for parasitic
leakage. Obviously, the more simple the
tester the less chances of error.

Comments Welcome!

Peter

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi Peter,

I hope I can explain this somewhat more clearly.

The insulation resistance test is specified to be done by a tester that outputs 500V DC. I have manual testers which already do this, and the new test equipment which also checks in this way. I am aware that leakages at high voltages tend to increase.

Since other products that have been tested on this device are specified to have only tens of MOhms or greater at 500V DC instead of several hundreds of MOhms the leakage was not apparent previously.

Regarding why I did not "connect two outputs together", I have been trying to find the insulation resistance of the individual sections of the circuit individually using a manual tester. Trying to connect the output of the manual tester to the output of the automatic tester to check the leakage of the tester alone against a calibrated instrument isn't very safe in my opinion (risk of damaging something?)

As to the relay matrix it is necessary since the automatic tester has only a small number of test outputs, and some tests can only be done between just three of these outputs. Since a variety of tests have to be performed (the product has up to five terminals) including functionality, insulation resistance at 500V DC, high voltge withstand at 1.5KV AC, earthing lead continuity, etc... the matrix is necessary to connect the tester outputs to the clamps connected to the product. Some relays are switched by the same control lines since the system is based on kelvin clamps with separate drive/sense lines.

I agree with you that a simple tester with low parasitic leakage might be better, and I actually have one such thing with a go/no-go bulb and buzzer. Unfortunately to comply with the requirements from up high we need to have a test report generated by the tester with no manual intervention by the operator (to avoid possible mistakes). Also I am required to use the ATE to avoid having four different devices for each item tested, which is horribly slow and inefficient.

I will be calling the manufacturer for the ATE on monday and see what they comment.

Thanks for the interest and comments!

NxP

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

nxp wrote:
I will be calling the manufacturer for the ATE on monday and see what they comment.

Greetings nxp,

Should be interesting to hear, I suspect they
will indicate that you are trying to operate
outside of their "normal" specs, and they can't
guarentee you'll get better performance.

Your choices are to use the manual tester, as
a special test beyond the "normal" product
requirements, and your customer should expect
to pay a premium for this service. Or, use
the ATE to drive an adapted manual tester
with the extended range to comply with your
expectation of operator error free testing.

You may have to route the ATE connections
through a new relay box of your design with
extended insulation resistance, to get the
required test performance.

Interesting problem, not seen every day!

Comments Welcome!

Peter

--
Peter J. Stonard
www.stonard.com