Re: Cap Question

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

Original poster: "Dave Larkin by way of Terry Fritz <twftesla-at-qwest-dot-net>" <teslaman15-at-hotmail-dot-com>

>Three factors affect performance, ESR, dissipation factor, and series
>inductance.  Pulse caps do have a technical advantage in all these, but the
>losses in either are so low that they are negligible.  In a Tesla coil, the
>other primary circuit losses are far greater than the losses in a good poly
>cap.  Modern cap losses simply are too low to matter.
>
>I would say that if an MMC and a pulse cap are both equally derated that
>they will have the same reliability in that they will both last "forever".

Very true.  However if you do the math an mmc, for anything more than a few 
kVA, derated to the same extent as a commercial pulse cap should be is more 
expensive.

Take an example - a large pig coil tank cap, 80kVDC 0.07µF.

MMC with geekgroup caps - 19 strings of 40 would be required.  760 caps, at 
$3 each that works out at - $2280. Excluding shipping.

Commercial - Brand new, EFO polyprop. $980.  Excluding shipping.

No contest.  Although you are right specialty capacitor manufacturers have 
less in the way of production controls that Philips, Evox-Rifa etc. in most 
cases they have had 30-40 years of practice building essentially the same 
product, and have gotten quite good at it :-) Certainly Maxwell and NWL run 
real _cleanroom_ operations, although PCI and CP are a little more 
'homespun'.  However their engineers over-design to compensate for that, so 
in the end it all evens out.  Pulse cap manufacturers also offer warentees, 
I think you'd have a hard time taking a little $3 cap back to radio shack 
and saying 'it broke' ;-)

>Statistically, MMCs should be "better" since they have all the big mass
>production controls behind their manufacture where pulse caps tend to be
>made "custom".  Unusual failures of pulse caps have been traced to
>variation in quality that is expected with a very low volume product with
>"hands on" assembly.  So if you make a million Tesla coils and run a big
>study.  MMCs will win due to very good manufacturing controls.  The low
>volume caps will show all kinds of odd problems.  Of the 5000 Panasonic
>caps I sold, only one "seems" to have failed for no good reason.  It was
>cheerfully replaced for free :-))
>
>MMCs can be very easily adapted to one's coil requirements and you can keep
>them on the shelf and quickly use them for most any coil depending on how
>your string them up.  They can also be made adjustable.  If one does not
>have a lot of money, you can overdrive them with a fairly predictable
>reduction in reliability which is often acceptable to coilers.  For small
>coils, MMCs are far cheaper.  There is a point at the mid size pig level
>that pulse caps start to be cheaper than giant arrays of MMCs.  Then, big
>pulse caps rule!  MMCs did inspire some nice equations and such for
>predicting RMS currents which can be used to specify the right pulse cap
>too.  If your building a coil like Bill's giant model 13, an MMC the size
>of a truck is totally impractical,  Pulse caps are the only option "up 
>there".

Exactly.  I agree that the mmc's vulnerabilty to over current, and the math 
needed to deal with it, has certainly 'advanced the art'.  When I started 
coiling no-one even considered rms pri. current.

> >>The MMC cap is the current state of the art in coiling. Stable value,
> >>tolerates repeated overvoltage, and completely variable to be as >unique 
>as
> >>your coil (for such a simple things, I have yet to see two >duplicate
> >>coils, they're like a fingerprint). They also have the nice >benifit of
> >>being inexpensive. $3 get's you .15uF at 2kV, and we sell >them to 
>coilers
> >>by the thousands.
> >
> >The EFO polypropolene pulse cap is the state of the art, and has been for
> >the past 30 years.  The mmc, thanks to its self healing capabilities,
> >achieves a useful cost/runtime compromise unachieveable with conventional
> >pulse caps.  In layman's terms - they cost less and die sooner.  So for 
>the
> >casual coiler they are the best choice, however that doesn't mean they're
> >the best cap.
>
>It may not be terribly hard to perhaps build self healing into pulse
>caps...  It may increase cost and increase ESR/loss some.  Maybe need a
>special oil too.  Probably not a big deal since pulse caps should be
>operated "properly" :-))  One big killer of caps is over voltage.

Already been done.  Most of the big pulse cap manufacturers use 'self 
healing' fuse link type technologies for their high energy density low rep 
rate caps.  It lets them use thinner dielectric, and hence get a better 
energy density, because they don't have to worry so much about a few punch 
throughs.

RF type caps are tricky candidates for self healing, because the fuse links 
increase the esr quite considerably, which matters far more in a CW or high 
rep. rate application (eg. Tesla coil) than in the shot-a-second thumpers 
from the likes of Aerovox and NCL.

Don't get me wrong, MMCs are a great innovation, but they are not the be all 
and end all of Tesla coil tank caps.  For big coils, or high run time coils, 
the good 'ol pulse cap has a good few years left in it yet!

-Dave-