[Prev][Next][Index][Thread]

Re: Cap Question



Reinhard,

My responses are interspersed below.

Tesla List wrote:
> 
> Original Poster: RWB355-at-aol-dot-com
> 
> Hi All,
> 
> Iīve got a really wierd problem. H E L P !!!!
> I decided to "oil in" one of my test caps. So I bought myself a piece of PP
> pipe, two end caps and proceeded on.
> 
> The cap specs:
> 
> Al foil: (2x) 78.74" long and 11.8" wide
> PE sheeting: (2x) 24 sheets of 1.75mil PE (42 mil per set)
> 
> I started with 24 sheets of PE. Next is one AL foil. Then again 24 sheets of
> PE and stopped with the last AL foil. I used a wooden rod to roll the
thing up
> nice and tight. I didnīt have very much slippage (PE sheets) and it ended up
> rather nicely.
> 
> Hereīs a bad ASCII Drawing:
> 
> _________________ AL foil
> _________________ PE (24 sheets)
> _________________ AL foil
> _________________ PE (24 sheets)
> 
<SNIP>
> 1.) I had calcīd this cap to be aprox 21 nF. It turned out to having only
10.5
> nF. Thatīs only one half. I scratched my head, unwound the whole thing and
> tried to rewind it tighter. It didnīt help at all. Why? Is there a
possibility
> that the two AL sheets are acting capacitivly on each other, because I
didnīt
> cover the top AL foil with more PE sheets effectivly reducing the
expected 21
> nF to the measured 10.5nF ?!?. I donīt think so.

The appropriate formula for a cap is:

     C = 0.2248*k*A*(N-1)/(d*x)  (in pF)
    where k = 2 (the "effective" value for LDPE and mineral oil)
          A = Effective plate overlap area in square inches
          N = Number of conductive plates. 
              (NOTE: use 3 for a rolled cap)
          d = individual dielectric film thickness
          x = number of stacked sheets of dielectric between plates
    
When the sheets are rolled, the assembly behaves as though you had a
3-plate capacitor since you actually end up using both sides of the
plates. It appears that the Tesla Coil design programs were predicting
the right theoretical value, since when I plug in the values for your
capacitor in my own design spreadsheet, I also estimate about 20 nF!

After thinking about this a bit, I suspect that what you may be
encountering is the effects of entrapped air. Since you're using a large
number of very thin sheets, with no kraft paper to aid in "wicking",
this will have the effect of substantially reducing the "effective"
dielectric constant so that it's closer to 1 instead of 2.
Unfortunately, it will also have the effect of creating internal corona
discharges which will, over a period of time, destroy the LDPE
dielectric. 

> 
> 2.) I plunged it into my PP container and filled her to the rim with mineral
> oil. The capacitance didnīt change at all. I ran a line from the vacuum port
> to a glass full of oil. From the glass I went to my trusty vacuum pump
(an old
> fridge pump...works pretty good for the $10 it cost me). Next mystery. I
> pumped it down and measured the capacitance while pumping down. I figured it
> would go up as trapped air was being removed. Nope, it didnīt!! It actually
> went down aprox 2nF in the beginning. After 30 min of running I removed the
> vacuum pump and let the cap suck back the oil in the glass jar. There was
> still oil left after the vacuum was back to zero, so I figured it was
(almost)
> full. The cap now measured 10.8nF. !!#~-at--at--at-##+!! I reconnected the vacuum
pump
> at let her have it another 30 min. As the pump started making vacuum again,
> the cap value started to drop again, stopping at aprox. 10.2nF. I redid this
> twice more. I let the cap stand for two full days and measured it again. Now
> it has 10.95nF.

Assume that there's air entrapped between layers that cannot escape,
even when you apply a maximum vacuum. Thinner LDPE sheets tend to entrap
more air simply because they're more flexible and there are many more
layers to provide the opportunity. Under a vacuum, the entrapped air
expands, sort of like a balloon, taking up the "slack" in your winding.
This expansion increases the spacing between plates, and would account
for the decrease in capacitance you're observing. And, if the air is
truly entrapped and cannot be extracted with a vacuum, then standing a
few days will not help either. 

Upon disassembly, you'll most likely find "dry" areas between dielectric
layers. BTW, this is a common manufacturing problem that commercial
manufacturers avoid by using hazy film (not smooth), and by the addition
of kraft paper to force wicking of oil between layers. I think the above
may explain the results you're seeing and addresses the four questions
you posed below... 

> 
> 1.) Why doesnīt the cap have 21nF or at least somewhere near this?
> 2.) Why didnīt the cap value increase, as it was being filled with oil. I
> figured the PE and the oil should have a K of around 2 at least ?
> 3.) Why did the cap value drop as I removed the air ?
> 4.) Why didnīt the cap change itīs value (not much at least) after
standing a
> few days?
> 
<SNIP>
> Reinhard

Safe cappin' to you!

-- Bert --