Re: yet another cap for MMCs

[Terry Fritz <twftesla-at-uswest-dot-net>]

Hi Jason,

At 07:48 PM 8/12/00 -0400, you wrote: 
>Hello,
>     I have a question about some caps. I was looking for a supplier of high 
>current MMC caps for use with my dual MOTs (4500 volts -at- around 2.5KVA), 
>when I flipped open my Allied Electronics catalog. The caps are on page 683 
>for those of you who have the catalog. If you don't the URL is:
>        http://www.ascapacitor-dot-com/page7.htm
>    What I was wondering was if anyone had used these caps before with good 
>results, and what people think about their power handling capabilities, etc. 
>The 2kv variety run from 100nf- 1.5mfd and I would be using 4, 470nf 2kv 
>caps in series. They look like they would handle the current, the type I 
>would be using is basically a 1.5" cube with terminals that are .5" wide. 
>The only problem is that these are a bit pricey being $11.47 a peice. If 
>they do work they would be good for low voltage/high current setups using 
>MOTs and other low voltage transformers. 

IGBT snubbers are excellent high-power/high-voltage caps.  They can make
great MMCs.  However, their price scares most of us away.  But in your
low-voltage and high-current coil application, the snubbers may be a good
match.

The "data" for these caps does not say much :-(  You probably need and RMS
current of about 15 amps.

WIMA has data at:
http://www.wima-dot-com/snubber.pdf

That is probably pretty close to yours.  Although not obvious, WIMA has
done the work for us.  The graph in the lower right on page 5 of the PDF
file says that the 2kV 470nF caps can handle 20 volts RMS at 200kHz.  That
may sound tiny :-)) but the RMS current is what we are after.

Xc = 1 / (2 x pi x F x C)  == 1 / (2 x pi x 200000 x 470x10^-9) = 1.7 ohms

The current is 20 / 1.7 = 11.8 Amps.  So at 200kHz, WIMA is saying you can
push about 12 amps through these caps.  Of course they are derated and all
that...

So you are in the ball park.  I can't positively say they will not burn up
but I am fairly confident they won't.  Without more accurate numbers, that
is as close as I can say.  The numbers say you are right on the line.  I
would try it.  I "think" they will be fine.

>    Also has anybody ever tried putting their MMC under oil, or with forced 
>air cooling to keep them cool? This could be a new field for EMMC 
>development. It would be interesting to build two identical MMCs,place one 
>under oil or use some other form of cooling and leave the other just plain 
>and then test the two to destruction. I'm willing to bet that the cooled MMC 
>would stand up at least a little better. You could probably just do this 
>with a single cap for each setup and it would be alot cheaper.

Originally, I was going to put the first MMCs under oil for cooling.
However, as time went on, it simply was not worth the mess for the little
cooling you would gain.  Poly caps heat from the center and all the poly
layers provided high thermal insulation from the center to the case.
Cooling is not that much of a help.  A fan is easy and may help if MMCs get
warm over a reasonable length of time.  However, you should really not let
them get hot no matter what.  Snubber caps do rely an their leads to carry
off heat.  So in your case, a fan blowing on those big copper tabs may be a
big help.  "I" would not use oil.

For the record, an 8 string by 4 in series array of 0.056 Panasonic caps
would run at +5C and cost about $45.  So it is all the same in a way...  8
strings of these caps will handle 24 amps.  "I" would probably go for a 6x3
EMMC :-))  That's only $26 for the current (18 amps) and voltage (2200V
each cap, kinda high :-)) that should work fine.  But I like to push MMCs
hard...

Cheers,

	Terry