[TCML] Caps in parallel

David Speck dave at davidspeckmd.org
Mon Sep 7 06:33:24 MDT 2009


Chuck,

The best arrangement is to have your caps in full series strings between 
the voltage sources. 

If you have mixed series/parallel groups, then you get into the 
following problem.

If one of the 5 caps in one column of one of your subassemblies develops 
a short, the remaining 4 caps in that column will now see the applied 
voltage of that subassembly, while the other 3 columns will sorta go 
along for the ride.  This will further stress the remaining 4 caps in 
that column. Further, the charge stored in the other 3 parallel columns 
of the subassembly will further stress the failed column, hastening its 
failure. 

Therefore, though it mechanically would be easier to do series 
connections of large sets of parallel caps, and electrically, you get 
the same total capacitance either way, from a fault tolerance 
standpoint, the best arrangement is a set of series columns of caps 
spanning the entire input voltage range without any horizontal parallel 
links, except at the very ends of the columns. 

You would do best with 4 columns of 15 caps bussed only at their ends, 
and without any other cross connections.

HTH,

Dave

chuck wrote:
> I have mounted several circuit boards with capacitors (WIMA FKP-1s) is
> series, 20 to each circuit board. The configuration is four columns of five
> caps each. Now I wish to connect the circuit boards such that the caps on
> each circuit board will be in parallel. Assume the connection of each group
> of caps is at the upper left corner (UL) and upper right corner (UR) of the
> circuit board. Now, by connecting each board UL-UL and UR-UR, I know the
> caps will be in parallel. What if I connect them UL (board 1) to UR (board
> 2) and UR (board 1) to UL (board 2)? This configuration is much simpler and
> uses less wire. Any hazards I am overlooking?
> Your opinions will be, as always, highly valued.
>  
> chuck
> aka asimov13647


More information about the Tesla mailing list