High Voltage High Frequency Insulation

Best reference I know of is a book, "Insulating Materials for Design and Engineering Practice", Frank M. Clark, published by John Wiley and Sons, Inc. New York and London. I believe it's still considered the bible of electrical insulation materials.

You may find a copy in a university technical library if there is one nearby or, perhaps arrange a library loan.

Might also be some references on the web.

Good luck.

Thank you Bluestone.

I will check our local libraries and online booksellers. I'll let you know the results

BigBirdAustralia

BigBird

I am a bit uncertain as to what you are trying to cause or prevent from happening but here are a couple of thoughts.

Your 80 kHz is only a ways above audio freqs so it may not be very significant.

First defense against corona discharge is to eliminate pointy places in the mechanism.

Next is to use some anti-corona gunk like they do on television fly-back transformers. TV shop should have some.

Third is to look at a material called Kaplan--it is used in high powered transmitters and hi-pots to well over 15kv while operating at fairly elevated temperatures.

Hope I didn't miss the point of your post entirely.

Lonnie

Thank You Ishurtle.

I have to wind high frequency transformers with 15 kV between windings. Points and spikes have been eliminated.

At 50/60 Hz the corona inception is around 500 Volt per .001 inch. To some extent it is frequency (dV/dT) dependant. Therefeore it is reduced at higher frequency, but I have not been able to find the figures for higher frequency or the formula concerned.

The whole unit will be oil immersed but the corona withstand of the actual solid insulation between primary and secondary is my main concern.

BigBirdAustralia

BigBird:

After seeing additional application details, I believe your real concern should be the dielectric strength of the oil. That, of course,depends upon the characteristics of the oil, electrode shape, area, orientation, oil gap, etc.

I suspect your question really relates to insulated wire you plan to use for transformer windings.

The same reference I mentioned earlier includes good information on these subjects. However, with regard to frequency vs dielectric strength, I believe you will find there is essentially no difference between power frequencies (50 or 60 Hz) and 80 Hz.

For insulation systems, frequency is a major concern at radio frequencies, for example, primarily due to dielectric heating.

For high voltage apparatus bushings with uninhibited mineral oil-paper or film-oil-paper insulation between condenser foils, I limited the continuous longitudinal voltage gradient to 14 kV rms/inch (which is the weakest direction). The allowable radial gradient was 100 kV rms/inch. Also had maximum design limits for one minute withstand, impulse and switching surge.

Hope these comments help.

Thanks Bluestone,

The frequency is 40 - 80 kiloHertz. The 1 minute withstand voltage on polypropolene (similiar structure to teflon) actually drops by a factor of around 4 from 2.6 kV to 560V for a thickness of 12.5 micron (ref IEEE conference paper). I suspect that the corona inception will do something similar.

Melinex (polyester film) gives 700 V corona inception for 100 micron (.004 inch) or 150 V per .001 inch (ref ICI industry note MX202) which would require a max of around 100 V per .001 inch or 0.15 inch at power frequency for 15 kV.

The 15 kV across the 1 layer secondary will be across 200 mm winding width so the oil is not highly stressed.

I need to determine the thickness of the insulating former between the primary and the secondary. At 200 V per .001 inch recommended maximum for teflon at power frequency (ref Reuben Lee, John Wiley & Sons 1988) this would be 0.075 inch, but the voltage stress will need to be reduced at 40 - 80 kHz. If I use the factor of 4 then the insulation thickness would be 0.3 inch. But is this significantly below the corona inception voltage at that frequency?

Of course the more insulation the worse the leakage inductance as well.

Cheers

BigBirdAustralia

I apologize BigBird. I missed the k (as in kHz).

With regard to Teflon (PTFE), we know that the dielectric constant and power (dissipation factor) are unaffected by frequency (at least up through 1 megahertz) due to its non-polar structure.

However, the dielectric strength is affected by frequency, as you noted. The actual reduction depends upon the surrounding media (air or oil, for example).

In the Clark reference I previously cited, there is a curve of PTFE dielectric strength in air versus frequency of the applied voltage from 60 - 10^8 Hz in both parallel and perpendicular directions. The curve is based on 30 mil thick samples with a time to breakdown of 40 seconds. The curve indicates that PTFE is weakest in the perpendicular direction. In that direction, the curve shows a dielectric strength of 650 vpm (volts per mil) at 60 Hz, 400 vpm at 10 kHZ and about 320 vpm at 1 MHz.

On the same page, there is also a discussion of the effects of testing under oil and indicates the breakdown value "may be increased to as much as 7000-7500 vpm" when tested under oil (along with the reasons).

While briefly looking, I also noted a corona-resistant PTFE is available (CRPTFE).

If you haven't already done so, I would think an electrostatic field plot of your geometry would be in order. I used to do this sort of thing with boundary element analysis (as opposed to FEA) software from Integrated Engineering Software.

Hope these comments help.

Best regards.

you can use poly-f4-ethylene ( I dont know how to spell in english). it has feature of low dielectric coefficient and high temp endurance. auitable for hfhv transformer.
maybe its teflon. I can help you to get it in mass.

dry air transmission (nitrogen) is your best bet PTFE, polyprop, etc. all have a curve involving dielectric breakdown. Where as nitrogen is rock solid through the VLF and ULF frequency ranges. So corona rings and EMF bleed will not occur until it leaves your helix and gets on the stick.

If I missed your point. Sorry to waste your time. I briefly scanned through the posts. 2:30am -7:00GMT I don't have the necessary constitution to read through all of the psuedo intellectual pandering that responded to your query.