[Amps] hi pot versus RF voltages

[John Lyles]

Work was done in the early days of high voltage RF to try and correlate these different breakdown conditions. I have a collection of papers on the subject that I have researched over the years, due to my own need to understand this for high power RF designs. To summarize the findings, RF high voltage is strange! It is difficult to find exact matches, and conditions vary among experiments. If you are using standardized ball gaps at sea level under fixed humidity, and making breakdown studies, peak RF voltage and DC are not far apart. The balls must be 1 inch or larger for this. As they approach needle points (sharp edges too) then RF breaks down much easier than DC. Add the presence of ionization influences from cosmic rays, UV light,  or a nearby radioactive source, and it changes again. 

My own figure of merit has been to use DC high potting and assume that the RF peak voltage breakdown is similar, as long as generous allowances have been made, such as keeping clearances, adding corona rings around sharp conductors, acorn nuts, buffing surfaces smooth, radiusing conductors and nearby grounds. Otherwise, if you are wondering about a band switch or coaxial connector, then all bets are off that there is a 1:1 correlation of DC to RF breakdown. As frequency increases, it is said that ionization takes place at lower voltage than breakdown voltage. Negative ions to accumulate at the gap. They are swept from the the field each half cycle. Positive ions, however, are not so easily swept out as they are heavier. At 60 Hz they can be swept out, but at higher frequencies, they add to the field stress at the gap and cause a lower voltage breakdown. 

But to be conservative, use a DC high potter, and then add a figure of safety to your design for RF. I don't have a ratio to quote, maybe others have experience with that. Standing waves on lines, of course, make knowing the exact RF voltage complicated. Or use an AC high potter, even a neon sign transformer with a variac driving it and do your own testing to prove the correlation. It isn't simple. 

Some of the better old reports on this are:
L. E. Reukema, "The Relation Between Frequency and Spark-Over Voltage in a Sphere-Gap Voltmeter". Pacific Coast Convention, AIEE, Sept, 1927.
F. W. Peek Jr., "Dielectric Phenomena in High Voltage Engineering", book, 1929.
P. A. Ekstrand, "Radio-Frequency Spark-Over in Air", Proc. IRE, June, 1940.
A. Alford,"Radio Frequency High Voltage Phoenomena", Pacific Coast Convention, AIEE, June, 1939.

73
John
K5PRO



> Message: 8
> Date: Sat, 24 Jan 2009 17:43:05 -0700
> From: Larry Benko <xxw0qe at comcast.net>
> Subject: [Amps] Breakdown voltages

> Has anyone seen a study or have any knowledge of the relationship 
> between the RF breakdown voltage and the 60Hz breakdown or DC breakdown 
> voltage?
> 
> I would like to test several parts such as switches, relays, vacuum 
> capacitors, and air variable capacitors in the HF range.  Hi-pot testers 
> are always (nearly always) DC or 50/60Hz testers.  I would hope to be 
> able to do a 60Hz test and then modify the answer by some percentage 
> (perhaps based on frequency).