[Amps] More Capacitors vs LC filters

[Jim Thomson]

Date: Wed, 18 Apr 2012 13:47:07 -0600
From: "John Lyles" <jtml at losalamos.com>
Subject: [Amps] More Capacitors vs LC filters

One more thing about big oil-filled capacitors: When the voltage gets above about 10-12 kV, manufacturers build their capacitors with series capacitor packs. Each one is wound paper or film with foil, connected by a jumper wire or strip. 
If you figure a 45 kV cap is made out of 3 series-connected 15 kV assemblies, and the total cap is, say 5 uF, then each cap must be 15 uF internally. All is fine and good until one of those packs shorts out. Lets say you are operating at 30 kV for convenience and this happens. 
Now you have two capacitors, of 15 uF each, operating at 15 kV each (approx), very close to their death zone. Doing the calculation, the original 5 uF capacitor could store up to 4 kJ at 40kV (1.3 kJ per internal cap), close as you would dare run to ratings. But is is derated to operating at 
2.25 KJ. Now, in the failure mode, which is, by the way, realistic, in a pulsed system over time, it is operating with two assemblies and the total stored energy went up to 3.37 KJ. This is part of the problem that will cause a 4 KJ capacitor to 'rupture' and expel oil and flames, in addition to slitting metal case. 
I don't know how those big 'Lytics are made, the 3900 uF 450 VDC jobbies, but if there is are any multiple-connected caps inside, then failure scenarios like this warrant being thought about. 
John
K5PRO

##  Points well taken.     The  3900uf @ 450 vdc caps  are just ONE cap inside, same deal with the even bigger caps
More Capacitors vs LC filters

.  Some where along the line, cap makers figured out how to stuff a lot of uf into a smaller package.   Reading through the tech notes,
what they did was dimple the foil, etc on the  electrolytics.   Much like dimples on a golf ball, except they dimple on both sides  of the foil, etc.    By doing this, they managed to increase the  total C by a huge amount.

##  The 3900 uf caps are not small either, but small for their total C.   The  2300uf caps  I have are   2.5 inch od  x 4 inch tall.      The 2500 uf caps are smaller,  at  2 inch OD  x 4 inch tall...and weigh even less, a mere 14 ozs each. 
The biggest ones I have are  10,000 uf each  @ 450 vdc.     These are 3 inch diam x 9 inch tall. 

##  cap makers  also terminated them internally  differently these days.    They now use wide, flat  connections, instead of small diam, round connections.   That reduced the  series  XL  by a whole bunch.  Then they can operate them on higher freqs.   If you look at old 1980 handbooks.. the  switching supplies always used like  nine  caps in parallel..since one big cap would not handle the ripple current.  

##  On modern day style lytics, the ripple current rating now increases  with increasing freq.   The old style lytics were the opposite.    I only operate these caps  at 70-75%  of the max V rating.   With 7650 vdc no load,,,and 24  caps.... no big deal.   7 caps could short out and then with 17 left, they are 100% maxed out.     Still, you have to use some common sense with the amount of energy involved.    I have never had a cap short out since 1972 when I got into this hobby.     The worse case is the uf  will drop off on  real old lytics, when they dry out.   Never seen one go open either.  

later... Jim  VE7RF