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

Re: Ganging Caps





---------- Forwarded message ----------
Date: Thu, 11 Sep 1997 11:25:15 -0500
From: "Barton B. Anderson" <mopar-at-mn.uswest-dot-net>
To: Tesla List <tesla-at-pupman-dot-com>
Subject: Re: Ganging Caps

Harri, thanks for your reponse and suggestions. You bring up some very good
points I hadn't thought about. Yes I was calculating equal voltage distribution
and yes assuming equal field distribution. I just figured I'd deal with the
tolerances as they occur (as they effect the working system). I'm going to take
your suggestions and make some actual measurements. I'll post my findings,
although it might not be for a couple weeks (vacation time coming up).
Regarding the electrical field distribution: You talked of encasing each roll
it's own conductive housing and connecting one side of the cap to this casing
to limit E-field. Possibly there's an easy way to do this? maybe simply
wrapping each roll in aluminum foil and connecting one leg to the foil will
acheive this E-field limiting. Of course I know roll to roll breakdown must not
occur. What do you think? What does anyone think?

I do have specs on these caps from CSI and the inductance is in a range from
100 to 200 nH. I'll have to measure this as well with dummy terminals since the
inductance will be influenced largely by the terminals. Also, William Noble
reported he tried DC caps but the series resistance was too high as you noted
was an area of concern. I'll have to check this out as well.

Thanks again Harri and William,
The Bart Dude


Tesla List wrote:

> ---------- Forwarded message ----------
> Date: Fri, 26 Sep 1997 09:24:32 +0300 (EET DST)
> From: Harri Suomalainen <haba-at-cc.hut.fi>
> To: Tesla List <tesla-at-pupman-dot-com>
> Subject: Re: Ganging Caps
>
> On Thu, 25 Sep 1997, Tesla List wrote:
> > of these apart, I found 5 cap-rolls paralleled. Each roll of course is
> > it's own cap measuring 20uF. I did the math and found I had enough to
> > get down to 0.001uF if series'd up (yes, I have a bunch). I could even
>
> It most certainly can be done. However, there are few points of consern.
> Do not assume equal voltage distribution! If you can estimate the
> tolerance for each roll calculate the worst case. Assume all other caps
> in the series string are at high limit and only one at low limit. Make
> sure the voltage calculated that way is ok for the worst case smallest
> cap.
>
> Do not assume equal electrical field distribution! Say you put one roll
> just on top of another. Ground the lower end. The high point has
> air or whatever around. Then the conductor foil is actually capasitively
> connected to the air. It may well have too high E-field. Make sure this
> never can happen. Pay attention on getting E field right. One easy
> method is to have each cap roll in a conducting cylinder etc connected
> to one end of the roll. Then the foil in the cap will be only capasitively
> connected to the conductur around and E-field will not be huge *inside*
> the cylinder. Do take care no cylinder-to-cylinder breakdown occurs.
>
> Do not assume equal leakage currents in the rolls. If leakage currents
> cannot be measured parallel each roll with a resistor making the leakage
> current high enough. Then calculate the worst case voltage for the
> largest resistor according to the tolerances. This is very essential
> if there is any dc bias. For general purpose use dc bias should probably
> be assumed.
>
> I think that after those matters have been taken care of it should
> work pretty fine. However, I hope the ESR/ESL of each roll is low
> enough. Do notice that if one roll has ESR of say 0.01 ohms 100 of
> them in series will make a cap with 1 ohm ESR.
> --
> Harri.Suomalainen-at-hut.fi - PGP key available by fingering haba-at-alpha.hut.fi