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

Re: MMC cap failure




Hi Steve, and all,

Sorry to hear about the MMC damage.  Maybe you can use some of the
undamaged devices ?

In going from the 100mA NST to a pig (500mA ?) the tank capacitor would
recharge _much_ faster.  If you increase the supply current and leave
everything else unchanged,  the break rate must increase in order to
process the extra power.  It is my bet that the higher break rate caused
excessive heating and weakened the dielectric in the MMC.

Putting two of these MMCs in parallel would roughly half the break rate,
and it would also share the tank current between two banks of caps.
Dissipation is proportional to I^2,  so they should run a lot cooler.

Peak output voltage from the pig could easilt be over 20kV, although I
doubt that overvoltage caused this failure.   I run a 70nF MMC (24
strings of 16 x 47nF caps in series,  rated 7200VAC, 24kv peak)  on
a 2.5KVA system with a supply that hits 20kV.  The MMC doesn't even get
warm because the break rate is only 200BPS.  I think you need more caps in
parallel not more in series.

On a lighter note, what spark length did you get before it all went
horribly wrong ?
                                                        Cheers,

                                                        -Richie,