Re: Poly Cap Tolerances (fwd)
---------- Forwarded message ----------
Date: Sat, 23 May 1998 13:02:44 -0500
From: Jeff Corr <corr-at-enid-dot-com>
To: Tesla List <tesla-at-pupman-dot-com>
Subject: Re: Poly Cap Tolerances (fwd)
>I built four caps and targeted the value at .0053uf and 7500wac. I used 8"
>wide AL flashing and .0625 poly sheeting in the caps. Unfortunately, I
>didn't keep accurate records of the following but I ended up with an actual
>value of .005180 with the four caps connected (2 pair in series, then
Your lucky to get so close, this small value of error is totally
>1. Once you have computed the cap value for your setup, what amount of
>deviation from this value can be considered acceptable?
Quite a bit - much more than here.
>2. What kind of differences in computed and actual value have others
>experienced when building homemade caps? (Especially poly/oil types).
I computed my cap to be .015uf, and it ended up to be .01256uf.
>3. Having built this first set, it appears to me that in general you can
>expect any difference in the actual/computed cap values to be a lower
>value? (It would seem less likely that construction techniques would "add"
>to the computed value. It seems obvious to me that construction techniques
>could obviously cause lower values through misalignment, tightness of the
Tightness of the roll is the most possible explanation here. It should be
that my caps, before being rolled were only 800pF, and the first time I
them, it was only .01uf. I rerolled them for a higher value.
>4. If #3 is true, and one can expect say a 2-5% lower value, should we
>target a value 5% greater than the calculated need? (Actually, I'm
>surprised my construction techniques worked as well as they did. With some
>practice, I expect I'll get closer to my target.)
That is what I do now. But keep in mind it is better to have too small of a
by 5% than 5% higher than the recommended value.
>I also don't know what the tolerance on the poly thickness is but mine
>should be close to advertised since I used a single .0625 sheet in each
>layer. Using multiple thinner sheets introduces more room for calculation
>errors I would think.
Calculation errors are more frequent with more layers I am sure, but I used
90 mils between each plate using 3 mil poly and it came out pretty close
>BTW: I rushed my first cap and must build another. While it is obvious to
>me now, I didn't ensure that the cap could hold a vacuum prior to filling
>and it would seem impossible to stop the leak now.
I always get griped for this... :-) but I have made two large caps for my
without vacuuming them whatsoever, and after 9 months both still run fine.
I think I went on the conservative side for voltage rating, perhaps that
compensate for the need to vaccum. While you might have to construct
cap, use the one you have made just to get an idea of how long they will