Re: More MMC qustions

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

Original poster: "Terry Fritz" <teslalist-at-qwest-dot-net>

Hi Mike,

dV/dT is the voltage change in volts per microsecond that caps can take.

In our case, suppose we have 2000V across each cap and our coil runs at 
250kHz.  During a single cycle, the voltage goes to +2000 to -2000 volts in 
1/2 cycle which is 2uS in this case (1/250000 x 1/2).  So we have 4000V in 
2uS or 2000 V/uS.

This value determines the peak current the caps will see.  Thus, it is an 
indictor if the caps are going to get torn apart by current spikes (like 
full metalized caps do).  If we have a 50nF cap in this case, the peak 
current can be found.

Zc = 1 / (2 x pi x F x C) = 1 / (2 x pi x 250000 x 50e-9) = 12.732 Ohms 
reactive.

Simply using I = V / R

Ipeak = Vpeak / R  == 2000 / 12.7 = 157 peak amps.  So the caps have to 
handle 157 amp peaks without failure.  Metal foil caps have no problem with 
this.  However, full metalized caps that have super thin conductors will 
"snap".  For some reason, everyone uses dV/dT but dI/dT and other things 
are used too.  A nice table for Geek caps is at:

http://www.cornell-dubilier-dot-com/film/9422000.htm

BTW - If one uses the geek cap numbers and figures out the current if you 
directly short them at 2000V, you will see why directly shorting high 
current caps is a bad idea.  The only thing limiting the current is the 
series resistance and the inductance.  In this case we have 5mOhm and 32nH.

150nF and 32nH gives a resonant frequency of 2.3MHz.  32nH at 2.3MHz has a 
reactance of 462mOhms (the 5mOhm resistance can be forgotten).  So the 
current is:

2000 / 0.462 = 4330 amps peak!!  the caps are rated for 432 amps peak so a 
direct short is almost "exactly" (you can here the engineers thinking ;-)) 
10X the rating.

BTW - It really is an exact science ;-)))  It can get a little messy, but 
there is real science behind it all.

Cheers,

         Terry

At 07:59 AM 7/4/2003 +0100, you wrote:
>Hi All
>
>I helped to sort out a bulk buy a year or so ago in the UK of rolled poly caps
>they were x-MOD spec 1.6Kv -at- 0.068 uF. so far I pushed these baby well
>over the limit with no ill effects, one MMC board with 4 strings of 16 caps
>and powered by a 15KVa pole pig with a variable speed rotary gap even with
>long runs 5 to 6 minits at a time they do not get hot and so far none have
>failed in service.
>
>One thing I hope someone on the list could tell me is what is DV/DT rating
>and how does this effect cap performance if any...
>
>Since this is not an exact science you should try the caps and see what
>happens and report back any finding that you make...
>
>That way we all learn more...
>
>Cheers Mike Tucknott (abusing caps in the UK)