Capacitor wiring questions

[Cotharyus]

I need some help with wiring on capacitors. I need a cap or group of
caps capable of delivering 50,000uf at 100V DC. I'm trying to sort
out what type and how many caps to use, and how to wire them. If
memory serves, voltage is doubled and capacitance halfed if wired in
series, and in parrallel I forget what happens - it seems like you
have to be careful charging them though, can anyone refresh my memory
and/or provide some pointers? Thanks.

 

[Quark Ng]

I need some help with wiring on capacitors. I need a cap or group of
caps capable of delivering 50,000uf at 100V DC. I'm trying to sort
out what type and how many caps to use, and how to wire them. If
memory serves, voltage is doubled and capacitance halfed if wired in
series, and in parrallel I forget what happens - it seems like you
have to be careful charging them though, can anyone refresh my memory
and/or provide some pointers? Thanks.

I would buy 125V caps and wire them up in parallel. The math is much
easier this way. When you wire them up in parallel, the capacitance
values just add.

Also, I am afraid that if you try to charge caps connected in series,
not all the caps will be charged to an equal voltage - causing over
voltage of some caps - causing explosions.

Choose 125V caps so that you have some margin of error to give you
breathing room.

-Quark

 

[JeffM]

I need a...group of caps capable of delivering 50,000uf at 100V DC

Cotharyus

Well, "deliver" is not the word, but Quark got it right anyway.

I'd be interested in what exactly you're doing
especially how you're going to charge it to 100V.
I'm betting you're going to get that wrong.

 

[Cotharyus]

I'm building a "battry zapper" for NiMH rechargable cells. Units that
do this are commercially available, however, all of them appear to
use military surplus parts (of which there are none right now) so you
can't buy one right now - I see no reason why someone with enough
sense to check and make sure they're doing it right couldn't build
one. I've also been given the name of a local EE, who appears to be
very difficult to get hold of...thank you for your help so far
everyone.

Quark - 125V caps (of lower capacitance?) wired in parallel? so if I
had 5 10,000uF caps rated at 125V I could wire them in parallel to
produce 50,000uF at 125V, and then only charge the caps to 100V? That
makes sense. I knew somewhere in my past I'd heard something funny
about charging caps wired in series or parallel, but I couldn't
remember exactly what it was. Thanks again.

 

[Quark Ng]

I'm building a "battry zapper" for NiMH rechargable cells. Units that
do this are commercially available, however, all of them appear to
use military surplus parts (of which there are none right now) so you
can't buy one right now - I see no reason why someone with enough
sense to check and make sure they're doing it right couldn't build
one. I've also been given the name of a local EE, who appears to be
very difficult to get hold of...thank you for your help so far
everyone.

Quark - 125V caps (of lower capacitance?) wired in parallel? so if I
had 5 10,000uF caps rated at 125V I could wire them in parallel to
produce 50,000uF at 125V, and then only charge the caps to 100V? That
makes sense. I knew somewhere in my past I'd heard something funny
about charging caps wired in series or parallel, but I couldn't
remember exactly what it was. Thanks again.

Yes, wire five 10mF 125VDC caps in parallel is what you want. Make
sure that you don't charge them up to over 110VDC, or you'll risk
heating up the cap and making it explode.

And just to be a geek, let me explain why you don't want caps in
series.

Say you have a circuit

A B C
100V---(SWITCH)-----|C1|-----|C2|----|C3|----|> GND

Start with the switch open. Assume that all caps are discharged and
points A, B, and C are at the same potential as GND.

When you close the switch, point A will almost instantaneously reach
100V. The only thing charging capacitor C2 is the leakage current of
C1. The leakage current is probably VERY small. So the potential at
point B would be near zero for a while. And the potential at point C
is near zero too.

This means the voltage across C1 is about 100V! So if C1 is not rated
for 100V (let's say it's rated for 50V) then it will be over-charged.
This will cause the dielectric to breakdown and allow large leakage
currents. So you're essentially shorting 100V to point B (which is
about GND for the moment) through C1.

Bye bye C1!

-Quark