Electrical conduction through air

[mdaugird]

TL;DR Summary

How to make air conductive without reaching breakdown voltage.

I am open to ideas of a way to test a bus bar for insulation flaws that has a 3D shape that varies in a production environment.

I have a DC bus bar that when it was just straight we would run a high potential test on a metal table that is the reference plane, and then flip it over to test the other side.
Now the bus bar is bent to have a 3D shape, and there are 14 different variants of the bends, and production could change the shape at any time.
The test voltage will be 3KVdc (Megger Test). The 3D aspect of the bus bar means that parts of the bus bar will be above the metal table that I am using as the reference plane.
The distance the bus bar is above the reference plane is about 18" in some cases.

The concept we are exploring involves a way to create a conductive atmosphere around the bus bar to avoid missing a gap in the insulation that was protected due to being so high above the table. Imagine a 2'x3' test cube that we can box in.

We didn't want to submerge the bus bar in liquid, but we could atomize a water(liquid) spray in the chamber, and use a high velocity drying air after to remove any droplets.
We have also considered using some kind of ionization technology, but it seems like ozone is a by product that could degrade the bus bar insulation.

I am interested to see if there is a different way to go than the above.
Thank you for your time and have a good day

 

[berkeman]

Summary:: How to make air conductive without reaching breakdown voltage.

I am open to ideas of a way to test a bus bar for insulation flaws that has a 3D shape that varies in a production environment.

Interesting problem. How 3-D is the shape? For example, if they are just bent into S-Curves, can you bracket your test bus bar with two other bus bars of the same shape, and apply your test potential between the middle test bus bar and the outer 2 reference bus bars?

 

[berkeman]

Interesting problem. How 3-D is the shape? For example, if they are just bent into S-Curves, can you bracket your test bus bar with two other bus bars of the same shape, and apply your test potential between the middle test bus bar and the outer 2 reference bus bars?

Wait, the outer reference bus bars have to have their insulation removed. Only the center bar being tested will have its insulation on it. Is that possible?

 

[mdaugird]

There is a single bus bar, I am not sure what is meant by "The center bar being tested"
The shapes change top dynamically making a bracket or JIG impractical.

Only thing insulated is the single bus bar.
this bus bar can bend back onto itself like a U shape, or S shape.

The hope is to create a conductive gas that the single bus bar sits in, so that the gas becomes that JIG dynamically to whatever gets stuck into the chamber.
RF is something we are considering as well, but I am open to ideas.
Thank you Berkeman for your thoughts, have a great day.

 

[DaveE]

How about a clam shell fixture with conductive foam shaped for each design. Drop the DUT onto the foam, close the lid with another foam shape to sandwich the DUT, and do the hi-pot test from DUT to foam? Of course some shapes will be difficult. Foam is cheap as far as tooling costs go, and the rest of the fixture could be common to all types.

 

[mdaugird]

Thank you all for your input but the time for the test does NOT allow for any JIG/Clamshell/Bracket structure.
There are 14 different SKU # coming from the assembly lines into this testing fixture. There is no way someone can select the correct JIG/Clamshell/Bracket install it, run the test and then put the part on the good or bad line and be ready for the next quickly enough, and then make a new JIG/Clamshell/Bracket when a new product comes down the line.

To be somewhat to the point, we tried JIGs already and it failed. Foam, Pins, Conductive balloons, saltwater, ect have been tired. Saltwater damaged the bars because the water creep under the insulation wasn't dried. The thought of atomized mist is viewed as something we can dry away with an air blade.
The produce line manager simply will not accept a solution based on the above.

My hope was to work smarter and find a unique solution.
Since this is a production environment, the concerns about toxic gases or oxygen displacement are also a consideration.
Neon, or Argon is a possibility because the volume is small, but still their breakdown voltages are higher than I would like.
thank you again.

 

[DaveE]

I agree, ionizing O₂ isn't a good idea. You would want an inert gas. The cost and complexity would be a big problem.

Conductive water spray sounds like a real pain for the manufacturing people and will likely lead to corrosion of the test fixture without careful mechanical design.

The problem with most Ions is that they like to attach themselves to other atoms.

 

[jrmichler]

Is this a megger test or a hipot test? They are not the same. If it is a hipot test, do not even think of doing this.

If a megger test, is the megger current limited to a safe value for humans? If so, you could connect the bus bar to the megger, have an operator wear a grounded chainmail glove (https://www.mcmaster.com/chainmail-gloves) and run their hand over the insulation.

Search engine tells me that megger current is limited to values that may shock, but are not harmful. I would want to get that in writing from the megger manufacturer before trying this idea. Then I would put double ground wires on the glove.

 

[berkeman]

I don't know enough about this test, but I was thinking along the same lines. An operator-moved probe of some kind that could be safely swabbed over the test bus bar surface to look for excessive leakage current. Our Hi-Pot testers all have sealed test chambers with safety interlocks, but maybe there is a way to make this safe.

 

[rbelli1]

Summary:: How to make air conductive without reaching breakdown voltage.
We didn't want to submerge the bus bar in liquid

What about a fluidized bed of conductive particles?

Example with sand. Your material will need to be https://ph.parker.com/us/en/conductive-plastic-pellets-and-injected-molded-plastic-solutions from sand.

It might be a good idea to fluidize with an inert gas.

BoB

 

[jrmichler]

A fluidized bed of conductive (copper?) particles would not conduct electricity while fluidized. But, you could make a bed of conductive particles, fluidize it, insert the bus bar, stop fluidizing, run the megger test, fluidize, and remove the bus bar.

 

[mdaugird]

That is a great idea, the fluidized bed can pump air while inserting the item and then we turn off the air and run the test.
thank you very much rbelli1 for a great idea

 

[Baluncore]

You might consider building a layer of chroming into or onto the insulation. Then you can test the insulation leakage without air getting in the way. The layer could be a fine copper woven mesh, a foil, or a conductive paint.

 

[tech99]

What about a fluidized bed of conductive particles?

Example with sand. Your material will need to be https://ph.parker.com/us/en/conductive-plastic-pellets-and-injected-molded-plastic-solutions from sand.

It might be a good idea to fluidize with an inert gas.

BoB

I don't think a bed of metal particles will conduct until a certain voltage threshold, which I think is due to an oxide film. Maybe just a few volts will be enough, but not sure how the bed will make contact with a pin hole in the insulation.
I think you could alternatively use a 5kV lab supply (having 1 MOhm resistance) with a probe to search for corona from pin holes.

 

[tech99]

I don't think a bed of metal particles will conduct until a certain voltage threshold, which I think is due to an oxide film. Maybe just a few volts will be enough, but not sure how the bed will make contact with a pin hole in the insulation.
I think you could alternatively use a 5kV lab supply (having 1 MOhm resistance) with a probe to search for corona from pin holes.

A second thought, maybe raise the bus bar to high voltage and then search with an RF probe for noise of any corona. Or use a UV camera to search?

 

[anorlunda]

A second thought, maybe raise the bus bar to high voltage and then search with an RF probe for noise of any corona. Or use a UV camera to search?

All good thoughts, but in practice it would take some laboratory experiments to measure the actual parameters. I don't think you will find quantitative answers here.

 

[DaveE]

A second thought, maybe raise the bus bar to high voltage and then search with an RF probe for noise of any corona. Or use a UV camera to search?

OK, or perhaps scherling photography? The problem is what @anorlunda said: you want a clear pass/fail result and predictable test conditions, you can't have the manufacturing techs "reading tea leaves".

 

[rbelli1]

an oxide film

I took a small aluminum spacer and tested continuity with a multimeter. The open circuit voltage was 2.7V. just barely laying the probe on the test article did not allow any current to flow. A very small movement or slight pressure broke the oxide layer and allowed current flow. I got the same results with one probe pressed tightly and the other lightly.

During fluidization there will be sliding of surfaces throughout the bed. After stopping there will be pressure between the particles.

In metal filled polymers there is a certain threshold where the material becomes conductive.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.462.8385&rep=rep1&type=pdf

As an experiment I took a 23cm x 17cm box filled to about 2.5cm with a random jumble of metal hardware. I placed a nail in one corner connected to a 10VAC transformer. I placed a screw in the opposite corner with an LED and resistor connected in series to it. I completed the circuit to the transformer and the LED lit brightly. I placed the screw in several other location in the box and the LED also lit brightly.

not sure how the bed will make contact with a pin hole in the insulation.

The original method was to lay the test piece on a flat conductive surface. I don't see how there was contact in a pin hole in the insultion in that condition. The 3kV test voltage must have been sufficient to arc through any pinholes. The selected test voltage will need to be high enough to arc over the maximum possible gap caused by the particles used.

BoB

 

[tech99]

Thank you for an interesting reply. I think that 10VAC is sufficient to break down the oxide film. I have done experiments with loose contacts like this. At voltages of a fraction of a volt I think you will find there is no path. This is the same action as the radio detector called the coherer, and it seems that the conduction is by electron tunneling in the metal oxide layer. It is non linear and bi-directional. With carbon particles, which do not have an oxide film, there is no such action.

 

[rbelli1]

loose contacts like this

I was actually surprised at how good the circuit was. Shaking the box caused flickering but once steady the contact was steady.

I tried it again. 9mA is the current of the LED. Fluffing up the pile led to a bit of flickering. A gentle shake stabilized it.

BoB

 

[tech99]

I was actually surprised at how good the circuit was. Shaking the box caused flickering but once steady the contact was steady.

I tried it again. 9mA is the current of the LED. Fluffing up the pile led to a bit of flickering. A gentle shake stabilized it.

BoB

As I mentioned before, I would be interested to know if you have conduction with voltages below about 1 volt.

 

[mdaugird]

I took a small aluminum spacer and tested continuity with a multimeter. The open circuit voltage was 2.7V. just barely laying the probe on the test article did not allow any current to flow. A very small movement or slight pressure broke the oxide layer and allowed current flow. I got the same results with one probe pressed tightly and the other lightly.

During fluidization there will be sliding of surfaces throughout the bed. After stopping there will be pressure between the particles.

In metal filled polymers there is a certain threshold where the material becomes conductive.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.462.8385&rep=rep1&type=pdf

As an experiment I took a 23cm x 17cm box filled to about 2.5cm with a random jumble of metal hardware. I placed a nail in one corner connected to a 10VAC transformer. I placed a screw in the opposite corner with an LED and resistor connected in series to it. I completed the circuit to the transformer and the LED lit brightly. I placed the screw in several other location in the box and the LED also lit brightly.
The original method was to lay the test piece on a flat conductive surface. I don't see how there was contact in a pin hole in the insulation in that condition. The 3kV test voltage must have been sufficient to arc through any pinholes. The selected test voltage will need to be high enough to arc over the maximum possible gap caused by the particles used.

BoB

There never is "CONTACT" when doing a High Potential Insulation test. We are trying to create a uniform electric field around the test time, We put 3-6KV on the conductor and measure the leakage current.
There is always some leakage current, and we graph that over time.
https://www.prioritywire.com/specs/Single Conductor 15KV, Shielded, MV-105.pdf
This wire is similar to the comment Baluncore made. Baluncore is correct that would solve my problem, but as a test equipment designer I can't change the clients product to fit my needs. However that is exactly how the real world tests work every day. we test 15KV cable at 69KV for 15 minutes. There is a capacitance that we see the charging current level off. We then evaluate that "steady state"
If there damage in the wire insulation, we will never see the current level off, bad terminations result in a higher leakage current per foot than we expect.
I do not know of a UV camera that can see the Corona reliably, what frequency is the UV spectrum from a corona in the 3-6KV range. I can adjust the humidity if that helps.
thanks again for the feedback.

 

[rbelli1]

I would be interested to know if you have conduction with voltages below about 1 volt.

Using the diode test function on the multimeter I see between 1.5V to 3V and the continuity beeper beeps occasionally when the box is shaken. Open circuit voltage is 5.1V. Short circuit current is 1mA.

I placed a 1n914 diode in series with the leads to get 0.66V. I can't get the mixed nuts to activate the continuity beeper when only 0.66V is present.

Touching or lightly rubbing the screw and nail together does not usually set off the beeper. Some pressure is necessary. More pressure is required at the lower voltage.

BoB

 

[tech99]

Using the diode test function on the multimeter I see between 1.5V to 3V and the continuity beeper beeps occasionally when the box is shaken. Open circuit voltage is 5.1V. Short circuit current is 1mA.

I placed a 1n914 diode in series with the leads to get 0.66V. I can't get the mixed nuts to activate the continuity beeper when only 0.66V is present.

Touching or lightly rubbing the screw and nail together does not usually set off the beeper. Some pressure is necessary. More pressure is required at the lower voltage.

BoB

Yes, the oxide film on metals ij light contact is an insuator below about 1V and this is quite reliable. Even shaking does not break the film very easily. This is the basis of the coherer radio detector, which is biased to about 1V and utilises electron tunneling to create a threshold effect.

 

[Tom.G]

Just Blue-Skying here.

Using the diode test function on the multimeter I see between 1.5V to 3V and the continuity beeper beeps occasionally when the box is shaken. Open circuit voltage is 5.1V. Short circuit current is 1mA.

That's why low-level signal circuits use Gold contacts in connectors, relays, switches.

Of course Gold marbles are out of the qustion here, perhaps Gold plated particles would do the job. (They still might walk out the door though. )

Another idea is instead of a water bath, there are various metal alloys that are liquid at close to room or body temperature. Try this link:
https://en.wikipedia.org/wiki/Fusible_alloy
Be sure to check toxicity!

Cheers,
Tom

 

[rbelli1]

I did one more experiment.

I placed into a trough a number of aluminum PEM inserts. I applied an AC voltage across them with a 100k resistor. At about 28VAC they started conducting. They remained conductive down to 0VAC. The electrodes were 27cm apart. Shaking the trough reset the system.

I repeated this with stainless steel inserts. They were slightly larger but the same basic shape. They did not conduct until 35VAC. However the first time I tried this I got 15VAC across the 100k resistor. After some amount of shaking the voltage went up to 30VAC.

I also repeated this with some galvanized self tapping sheet metal screws. The results were similar to the aluminum parts except that only 28VAC was necessary.

If using metal particles to do this test, material is important. I would recommend not using stainless steel and probably anything that uses chromium for the passivization layer is not suitable. Chromium can be toxic anyway.

BoB

 

[Windadct]

Having specifically worked for a company that make bus bar insulation systems, I feel your pain.

We wrapped in foil
We used a conductive brush electrode (yes a manual process)
We did water immersion when feasible

Still depends on the insulation levels needed and how you need to "certify" the test to your clients.