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

MMC cap failure!

Hi guys,

Last Sunday, I met up with Mike Tucknott and Steve Crawshaw (and a friend of
his) to do a bit of disturbing the peace with some pig power!
I took my coil round to Mike's parents' garage. We ran it in its standard
configuration, i.e. using my 1kW NST, to check it was working and in tune
before connecting it up to Mike's pig. Between the two of them, Mike and his
coiling buddy Brian Le Page, they have a well sorted system for controlling
the pig; a big variac, a welder ( with variable current settings) and a load
of neatly wired up switches and connectors.
We knew that the stand for my coil makes the coil a little too tall for the
average sized garage, and this one had metal beams right above the toroid,
so we decided to remove the bottom board (and feet) and rest it on the legs.
This gave us approx 4 inches more clearance. The gap, the new blower box
(built for the occation, the day before) and the cap were now resting on
some carpet on the concrete floor. The NST was replaced with the power feed
wires from the pig, which was sitting out of strike distance. Everything was
checked out OK and ready for action. With Mike at the controls we let her
rip!

With the welder set on maximum current limiting the output was noticeably
better than with the NST. We then cranked up the current a bit and tried
again. It got better, for a while, until Mike shut it down a bit sharpish,
saying something about the cap's burning! 'Oh no! Not my MMC!'.
The MMC was not on fire, fortunately, but what we had neglected to think
about was that the ends of the individual caps were now very close to a
concrete floor, which tends to conduct very well at high frequencies.
Judging by the black singe marks on the carpet there must have been a
terrific amount of corona leaking off the cap wires! We placed the cap up on
a sheet of 1/4 inch plastic, then turned up the current some more, and
continued to have fun.
About 2 minutes later something popped in the cap, with a flash and a nasty
smell.

The corner cap had failed in spectacular fashion! It burst in the middle of
one side and burnt two neighbouring caps, and the internal pressure also
blew the contents out one end and into another cap!
We can estimate the voltage output of the pig as 12kV, and each MMC string
had 22 caps (0.047uF -at- 1500VDC/450VAC) in series. I used the 450VAC rating,
with my NST at 10kV (RMS) as a guide to the string size. After looking at
other coilers' MMCs I realised that mine was rated a little bit on the
conservative side; it should have been bullet-proof!
I suspect that when the cap was resting close to the concrete and blasting
corona from its ends, it either overheated the cap or weakened it in some
other way. I'll replace all four damaged caps (and any suspect resistors
too) and keep in on a higher support next time.

(This is a copy of the write-up on my web site.  Check it out for the
associated pictures.  Look in the gallery and the testing sections.)

Steve Bell (Swindon, UK)
steve-at-breakfix.demon.co.uk
http://www.breakfix.demon.co.uk