Using metallized film caps instead of foil caps in MMC

[NickyLee]

Hello everyone, this is my first time visiting this forum so please excuse and correct any etiquette oversights on my part.

I would like to collect some opinions on using metalized film capacitors in MMCs.

I have heard using them is a no-no for reliability reasons, but I have also seen builds that use them. I would love to be able to use them for my build for budget reasons, but I know that choice will probably come back and bite my a** later. The bank would be about .02uf and around 25-30kv rated. JavaTC says the coil will resonate at about 115kHz if that's important. I can share other coil info if needed.

If people have experience or stories of using metalized film caps instead of foil caps please share them. Any input is appreciated

[Twospoons]

The problem with met-film is that high peak currents can burn away the metalization from the end connection. If you have an I²t rating for your caps, and you know the voltage and frequency of operation, you can check your peak current will be within limits.  Be aware that supplied I²t generally assumes a half-sine pulse at 50 or 60 Hz, which may not be valid for high frequencies.  I would recommend derating.

Some years back I was working designing electric fence energisers, using 1200V met-film caps. Based off the I²t from the manufacturer, everything should have been within spec, but after a week of running my prototypes died - because the end termination had burnt away inside the caps. After talking to the manufacturer, describing the issue and our application, they revised the I²t down by almost 50%.   The difference was our circuit was draining the caps in a 100us pulse, not the 10ms half sine assumed by the original rating.

[NickyLee]

Thank you for your input.

The caps I'm looking at (R76TN31505050J) have no mention of an I²t rating but they do have a non-repeated max pulse current of 450A. I'm planning on running 3 strings in parallel in my MMC for a combined max pulse rating of 1350A. JavaTC estimated the peak pulse current to be around 230A, so hopefully that is enough margin for some amount of fun before failure.

[davekni]

I would like to collect some opinions on using metalized film capacitors in MMCs.

Depends greatly on the particular capacitor.  As Twospoons recommends, check capacitor specifications.  RMS current capability is key, and peak current (peak CdV/dt).  Some capacitors are combinations of foil and film.  And, of course, polypropylene dielectric.
If you have the luxury of access to a thermal imager, test a capacitor with high frequency AC current.  Good ones will be relatively cool overall, with warmest part in the middle, since leads provide some heat sinking.  Bad film parts (with poor connections to end of foil roll) will be hot at ends by leads.

The caps I'm looking at (R76TN31505050J) have no mention of an I2t rating

Instead they have a V^2/t rating, of 9.6E12 V^2/second for that part.  Multiply by C^2 (by 0.15E-6 ^ 2) to get I2t.
RMS current is listed as 16A, down to 6.3A at 85C.  That may be most limiting specification.  Depends on your planned duty cycle and cooling.  And on desired life time.  Specifications are generally for commercial use where caps are expected to have good reliability for long term product use.  TCs are often used for only a few hours during their entire existence.

[NickyLee]

I plugged everything as best I could into the Kaizer Power MMC calculator and everything looks good. All the power parameters have a SF of 2.1 or (much) greater, and the single cap temperature rise is 4.4 C.