Re: MMC cap bank

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Crow Leader by way of Terry Fritz <teslalist-at-qwest-dot-net>" <tesla-at-lists.symmetric-dot-net>

Tesla list writes:
>Original poster: "Terry Fritz" <teslalist-at-qwest-dot-net>
>Hi Ken,
>150nF at 60Hz 2000Vp-p will draw 80mA.  It would be best to run them in 
>parallel as you say.  Then you could time them all and have 5 data 
>points.  Needs a 1500V 400mA transformer to run 5 caps.  Not sure if a MOT 
>would just overheat...  Would take 600VA to run it!

600VA is not too bad, may a blower on a microwave oven transformer might be 
in order. As usual, the more one wlooks the more interesting things get. 
Take  my proposed 5 parallel caps. This introduces possible problems as 
well. When one cap faults and self heals, the other 4 caps can dump some 
extra energy into the faulting one. This might make for a more realistic 
test if you run 5 strings of capacitors.
Other issue. How do you know when a cap is failing? Film caps can short 
out, or just burst into fire, especially when you have a large amount of 
available power running it. They can also gracefully age where the 
capacitance drops over time. From the literature I've read, people like to 
consider a self healing cap end oflifed when it loses 10% of initial 
capacitance.
Whih 5 caps in parallel, you cannot really tell when one has worn out 
because of capacitance drop without some sort of current measurement always 
taking place. A drop in current, or increase (short in the cap) would mean 
the cap died. One might measure the current the whole set of caps is using, 
and if it goes under or over a set point, everything is powered off. Time 
could be recorder of a elapsed time meter. Line variations might throw 
things off. A ferroresonant power conditioner could regulate the power to 
the microwave oven transformer.
KEN
>Cheers,
>         Terry
>At 09:36 PM 7/4/2003 +0000, you wrote:
>>Tesla list writes:
>>Why is series, ad at a higher voltage? The failure of one will affect the 
>>voltage the others receive. You're going to need more than 5 caps to at 
>>one testing to declare a sound spec for the rest. I just had a new light 
>>bulb burnout after a few hours. I guess that means lightbulbs last 3 hours.
>>KEN
>
>