Re: MMC's vs CW and Duty cycle

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

Hi David,

	You hit on a good point here.  The ratings given in the charts are for 24
hour operation or until the temperature reaches a steady state value.  In
my cap testing, that was about 30 minutes.  I ran the numbers for your caps
under the conditions you mention and 5.6 amps should drive them up 10C.
With double that current, they should go to 40C.  However, you have plenty
of time until they get that hot.  They even may last hours at 40C but that
would be pushing it...

Current destroys the caps by overheating and melting the polypropylene
dielectric inside the caps.  But if you start with cold caps, you have a
considerable time that they can be run over current before the centers
start to meltdown.  Designing for such things is really hard because the
numbers are all fairly "fuzzy".  One would really have to actually try and
test the caps to see if they can run so close to destruction in a
particular application.  It sounds like you system is running hard but happy...

When MMCs die from too much current it is usually not a 2X or even 3X
overload.  It's those 5X to 100X+ overloads that kill them...  Unlike
voltage, they can take a very high sort term current load as long as they
are cool to begin with.  The fact that these caps are designed to last
millions of hours and we use them for only a few hours really works to our
advantage...

Cheers,

	Terry


At 08:45 AM 8/25/00 -0700, you wrote:
>Hey all, Bert, Terry,
>
>The recent posts by Bert on the VTTC Tank thread has got me wondering. Now
>as Bert has pointed out, the fact that I am running a single string of
>MMC's in my VTTC, I should have had failiar due to over current. I am
>stumped. The theory and reality do not mesh...
>
>Last night I did some tests:
>
>First I hooked up the Vac coil to run in filtered DC mode, two 833C's
>powered by a 125# plate transformer. Measured Tank capacitance (7 series
>0.015 uF Panasonics): 2.12 pF. Measured Primary Inductance: 81 uH. 9K in
>grid leak. 16 uF filter cap.
>
>
>1: Brought up plate voltage to 2300V, 6-7 in flame discharge, but would
>trip 20 amp breaker after 10-15 seconds of run time, barely perceptable
>color on plates. No detectable heating of Caps. That should be 11.8 amps
>on the caps.
>
>2: Rewired Plate Transformer for lower output. Brought Plate voltage up to
>1825V, 4+ inch flame discharge. Ran (5) 40 second tests, with 15 second
>breaks to test temp of caps. After all runs there were some notcable
>heating of the leads, but no real heating of the caps them selves. This
>should have put 9.34 amps on the caps.
>
>Do these results make sense? Do I need to go for longer run times?
>
>Sugestions, anyone?
>
>Regards,
>
>David Trimmell
>www.ChaoticUniverse-dot-com 
>