RE: Cap life testing - Re: MMC cap bank - ion inception testing

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

Original poster: "boris petkovic by way of Terry Fritz <teslalist-at-qwest-dot-net>" <petkovic7-at-yahoo-dot-com>

 > At 10:22 PM 7/6/2003 -0400, you wrote:
 > >Hi Terry:
 > >
 > >A couple of times, you suggested that a streamer
 > strike to an MMC cap
 > >might be the cause of  trouble.  I'm having trouble
 > imagining a scenario
 > >where an MMC might be (re)charged to a higher
 > potential than Vgap.  Am I
 > >missing something?  I just hate to propagate
 > imaginary failure modes.
 > >
 > >Regards, Gary Lau
 > >MA, USA
 >
 > I was thinking more that a streamer might hit the
 > body of a cap and tear
 > through the insulation and into the cap plates.  At
 > that point in time, the
 > cap should be pretty well discharged.  However, the
 > streamer has enough
 > voltage to easily pass through the cap covering and
 > such.  Mostly physical
 > damage to the cap is what I was worried about.  If a
 > streamer could tear
 > say 1/8th inch into the body of the cap, that would
 > probably kill it.  But
 > the outer plates "should" guide the streamer away.
 >
 > One may wonder if say 30pF at 200000V (top load
 > energy) has the energy to
 > recharge the cap.  I.e., if a full discharge from
 > the secondary into a cap
 > could blow it.  I guess we have to assume the
 > streamer loss would be low
 > and the energy would go straight into the cap.
 >
 > E = 1/2 x C x V^2 = 1/2 x 30e-12 x 300000^2 = 1.35J
 >
 > 1/2 x 150nF x V^2 = 1.35J   V == 4243V
 >
 > So, it may be possible that a "perfect" streamer hit
 > could pop a single MMC
 > cap, if everything were just right, just from over
 > voltage.  Of course, the
 > cap could easily self heal from that odd event.
 >
 > I guess there is also a chance you could get a
 > primary to secondary power
 > arc going with the MMC cap in the middle.  That
 > would do more damage.  but
 > just a new single cap or two and it would be all
 > fixed in any case.
 >
 > A streamer hit might cause much more havoc if it
 > attached to a bleeder
 > resistor.  But I would think the streamer would be
 > guided to one of the end
 > caps rather than stay on the fragile film.
 >
 > I can't remember anyone reporting MMC damage from a
 > streamer hit.  I know
 > they certainly have been hit, but no damage was
 > seen. I don't know if any
 > commercial caps have been hit from the side into the
 > cap.  I can imagine
 > that may do serious harm.
 >
 > Cheers,
 >
 >          Terry

Hi Terry,
"Streamer" that hits primary circuit from secondary
side is actually TC power arc.It's frequency is
usually in range 10-30 Mhz for most of TC circuits .
Several possibilities may happen (at least in theory).
Depending on cap bank position with respect to spark
gap and cap bank geometry connection in MMC string.

1)
Single cap in string ,due to fast transients,might get
higher voltage than usual.
2)
Also,caps may be more sensitive to such a high
frequency (and permitivity drops for higher
frequencies)?
3)
Leaving some energy in cap bank after the "streamer"
hit ,if high speed rotary employed,could allow psu to
charge cap in next bang to a little higher voltage
than before.This isn't very likely to happen/or be the
problem and options 1) and 2) should be priority for
investigation for various cap and gap positions in a
primary circuit.

Just couple of ideas to throw in.

cheers,
Boris