Richie - Garry's MMC cap failure explained.

["Tesla list" <tesla-at-xxxxxxxxxx>]

Original poster: Terry Fritz <twftesla-at-uswest-dot-net>

Hi Richie,

You may have hit on the exact problem.  These caps have ALL vacuum
deposited metal film layers.  There is no solid metal foil in them.  This
seems to be causing two problems.  The metal film does not seem to attach
to the end caps well.  The caps that are getting hot on only one side have
the foil breaking up at the end cap attachment.  This picture shows the end
cap area of a cap that had a "hot side".

http://hot-streamer-dot-com/TeslaCoils/Misc/blownMMC/Pc020019.jpg

The low pulse rating of 57 amps is probably taking into account that the
poor film to end cap attachment will simply fail under high current.  To
make things worse, when these caps get hot, they expand greatly further
tearing up the layers and end attachments.  So the high pulse currents of a
TC probably blow these things right out.

What is odd is that these caps are rated at 3.8 amps RMS!... but they don't
say at what frequency.  However, the WIMA graph at:

http://www.wima-dot-com/polypropylenkurven.htm

tells a terrible story!!

At my 350kHz test frequency, the foil types have a dissipation of 0.009 to
0.028.  However, the metalized film types have a dissipation of 0.050 to
"right off the graph".  It is easily conceivable by eyeing the graph that
they have 20X the thermal dissipation of a good metal foil cap which would
explain the very high heating I see and match my testing.  The metal on
these caps is not very thick (and surprisingly inconsistent).  I bet some
manufactures get a thicker metal film if they use metalized layers but
still want good RF performance.

So I think you are right in that the deposited metal film in these caps are
the problem here.  Poor end attachment, poor high frequency dissipation,
and low pulse ratings are the cause of this dramatic MMC failure.  The 3.8
amps RMS rating is probably fine for 1000Hz, but at 350 kHz the thin
deposited metal film has terrible loss. 

Hopefully, my known good/bad cap list will help to identify caps better
since the data sheets can be misleading.  However, I would think we should
avoid the all vacuum deposited metallized film caps with very low pulse and
dV/dt ratings like the plague...

Cheers,

	Terry


At 09:40 PM 12/2/2000 +0000, you wrote:
>
>Terry,
>
>Maybe these capacitors make use of Metallised Polypropylene instead of
>using actual aluminium foil ???  I don't know if this matters that much,
>but I might expect foil to withstand high peak currents better than a thin
>layer of metal deposited over the dielectric.
>
>Our Yellow Arcotronics caps (marked RS) here in the UK,  use Aluminium
>foil for the end connections,  and an intermediate layer of Metallised
>Polypropylene.  The internal construction is like two caps in series.
>
>I would be interested in hearing your thoughts on this.
>
>							-Richie,
>
>> Original poster: Terry Fritz <twftesla-at-uswest-dot-net>
>> 
>> Hi Stan,
>> 
>> 	I just keep looking and looking at that data sheet...  3.8 amps RMS...
>> These caps I have here would be lucky to sustain 0.5 Arms...  I wonder if
>> the peak amps of only 57 is the key.  Perhaps the high pulses simply ripped
>> them apart and destroyed their current handling ability.
>
>.... snip ....
>