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

Re: [TCML] homemade pulse cap Now MMC cap



Hi Bert,
Sorry for my greatly delayed response to this, been busy out of my mind... Anyway for my next big project I did order MMC caps tonight, cost more than a pretty penny (40kV 50nF worth.) I bought the CD940 series, the 2kV .1uf kind. I know these are on the naughty list, due to being just film, as opposed to foil/film, but considering the cost difference to the CD942 series I think it should be good because it was, I think, very over engineered. I guess my question is this, as long as I am well below the pulse and RMS current ratings (which I very much am) per string I should be Ok yes? I plugged my values into Java TC and got "good" on standoff voltage (using a 14400 piggie.) In the eyes of the MMC people, do these values look Ok?

11 strings of 21 caps, 231 caps in total

per cap
V=2kV
per cap C=.1uF
DV/DT=1712
Peak amps=171
RMS Amps=8.3

So the whole array
V=42kV
DV/DT=I have no idea, as I'm not sure what this is or why it is important...
Peak amps=1881
RMS Amps=91.3

Now knowing this and using a properly sized cap, my expected transformer and break rate specs, and approximate values for primary inductances I get the following requirements for my cap bank...

V 14400/.707=a shade over 20kV (so I have just over double what I need)
Peak amps 394 (so I have 4.7  times what I need)
RMS current 41.64 amps (so double on this one.)
It looks to me like it should be good, the only thing I worry about is I don't know is the specs account for the high frequency at which these will be used (about 60 kHz)... Anyway just want to know if this looks Ok, or if I should expect flame-age... Thanks!

Scott Bogard.

On 7/23/2012 11:19 PM, Bert Hickman wrote:
Hi Scott,

I didn't see any replies to your question, so I'll take a stab at it...

The strategy of using two (or more) thinner layers of dielectric between plates has historically been used for decades by commercial and homemade capacitor builders for many of the reasons you mentioned.

Multiple dielectric layers are not inherently self-healing. A defect in one dielectric layer is protected by the remaining good layer(s), making the cap considerably more robust than one made using a single layer. However, if the remaining layer(s) are punctured, the entire capacitor catastrophically fails.

Problems of corona in voids between dielectric layers, or between metal and dielectric layers, were initially resolved by inserting a thin Kraft paper layer in film-foil capacitors. The Kraft paper layer was saturated with a dielectric fluid having a comparatively high dielectric constant (~3 - 7). Capacitors using this approach were sometimes called "soggy film" caps.

Most commercial capacitor manufacturers eventually migrated to a "hazy film" dielectric where one or both surfaces were mechanically roughened. The roughened surfaces were more easily wetted by dielectric fluid, and hazy film eventually eliminated the need for Kraft paper in all but the toughest energy-discharge applications.

Gaps between the foil plates and dielectric were eliminated by evaporating a thin layer of metal directly onto one or both surfaces of the dielectric. And, by carefully controlling metallization thickness (and sometimes adding thin-film fusible isolating links), dielectric failures could be isolated hundreds, or even thousands of times, resulting in graceful degradation rather than sudden death. "Self-healing" technology also eliminated the need for multiple dielectric layers per capacitor element, and permitted manufacturers to more highly stress the dielectric system. This revolutionized HV film capacitor technology, allowing manufacturers to build caps with much higher energy-density.

It turns out that your idea of stacking a series of "floating" conductive and dielectric layers is actually being used (albeit in a slightly different configuration) in the very snubber caps that most folks use for MMC's. CDC effectively stacks two (or more) identical single-layer self-healing capacitor elements in series (using a clever metallization pattern and proprietary construction techniques) to create reliable "dry" film HV capacitors.

Conceptually, your idea could work. However, almost nothing "sticks" very well to polyethylene (PE) or polypropylene (PP) film... unless the film has previously been corona-treated. Simply painting an adherent conductive layer onto untreated PP or PE film will not work. Your approach might work if you can locate some metalized PP capacitor film for a reasonable price. Metalized PP film does appear every now and then on the surplus market, but it's usually not cheap. You would also need to make an appropriate multilayer winder and may also need to construct a filtered clean air work environment. End termination could also be a problem unless you use foil for inner and outermost layers. I also suspect you'll find that MMC caps are really not all that expensive in comparison... :^)

Good luck,

Bert

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla