Re: Dialectrics in Plate Capacitors
You are absolutely correct --- there are no absolutes. But there is
definitely "the final straw that broke the camel's back!". My suggestions
are based on good engineering practices. You can cheat a bit if you don't
run your system very much or can tolerate failures on a regular basis, but
IMHO if you are going through all the work to wind up plastic caps and
solder them together and then oil-immerse them why not do a great design to
start with. Why take short cuts especially when building caps is very time
Pro cap builders use .002 inch (2 mil) or less for all their dielectrics in
caps running in the 7-50 kV range. They then series these units up to
acheive the capacitance they require. Amateur builders can use thicker
dielectrics for reasonable handling but pushing a single 60 mil thickness at
15 kV Erms with over 20 kV peaks is not reasonable. It won't last unless
you get very lucky with statistics and by chance get a near perfect piece of
plastic. -- it can happen. Running 4-5 mil units would be more reasonable
and give you a better survival chance.
From: Tesla List <tesla-at-pupman-dot-com>
To: tesla-at-pupman-dot-com <tesla-at-pupman-dot-com>
Date: Sunday, December 19, 1999 11:39 PM
Subject: RE: Dialectrics in Plate Capacitors
>Original Poster: "Coiler" <mycroft-at-access1-dot-net>
>First and foremost, I will say this: There are no absolutes!
>After all, we "know" we have to go to 3xDC rating for building caps.
>But then Terry comes along and actually runs over the DC rating for some
>time, in his EMMC :)
>That being said:
>> 60 mil is not adequate for 15 kV Erms which will hit well over 20
>> kV peaks.
>It can be. Depends on how it is constructed. While I personally now
>use 90 mil as 3 30mil units in series, I know of some that are running
>at less just fine. But a 90 mil single unit *will* fail. It's just too
>much potential to prevent corona even under oil.
>My rule of thumb is to aim for <5KV RMS potential for each unit in a
>> Recommend at least 90 mil. Professional manufacturers use .002
>> mil and then
>I'll assume you mean 2 mil? or even .2 mil? I doubt you mean 2 microinch.
>Myself, i use 5 ply 6mil.
>> 1 inch will not provide enough creepage path across the plastic to avoid
>> tracking especially at RF frequencies. Go at least 2 inches.
>I am living proof that this is not an absolute. Flat plate capacitors.
>1 inch (or less) around the foil. Been running fine at various power levels
>for quite some time. I have *NEVER* had a capacitor failure due to creepage
>or blow through. Have had one fail because a helpful student dropped it :/
>Again, many units in series is the solution. My caps are used 3 in series
>Coiler, Homebrewer, Nerd...mycroft-at-access1-dot-net