[TCML] cap strings

[bartb]

Hi Marko,

The smaller value capacitance will see the highest voltage. Because the 
current in a series string is the same for all capacitors, the charge 
(Q) at any point in time is the same on all capacitors. The voltage drop 
across each cap can be found by V = Q/C. Q is the charge in Coulombs. As 
such, the capacitance itself can be viewed as the charge stored per volt 
(farads = coulombs/volts).

In any event, since the charge is the same across each cap you need to 
find it (C for each cap is already known). To find Q you need to pick a 
point in time as the current will decrease exponentially as 1/e. It gets 
hairy to account for current in time in a series string of different 
capacitors when considering the time allowed before breakdown and 
current limited source.

But, for example, say Q is .0003 Coulombs (I x t) (small current limited 
source and controlled time to discharge).

The 0.5uF caps will see Q/C = .0003/0.0000005 = 600V

The 0.05uF cap will see Q/C = .0003/0.00000005 = 6000V

So, yes, the smaller cap will see about 10x the voltage not accounting 
for the impedance of each cap which might vary considerably. Typically, 
the source voltage will not exceed the lowest value voltage in the 
series string in a design like this. In your case, your probably 
thinking you can put something like 10kV across the string and hopefully 
the 0.5uF caps will never reach their breakdown voltage of 1500 volts 
before the discharge event. It does work in theory, but I think it runs 
a real risk of blowing the caps if the charging current is large enough 
and time between breaks was not controlled well.

A good experiment as long as you don't mind loosing a few caps along the 
way. But if venting caps (poof!) is not what you had in mind, best to 
work with identical capacitances and breakdown voltage values. If you do 
try it, you may want to rectify the hv output.

Well, gotta go (NFC & AFC Championship day!).

Take care,
Bart





mark olson wrote:
> At 11:18 AM 1/19/08 -0600, you wrote:
>> Hi group,
>>
>> If a string of caps in series have different voltage ratings, do the 
>> ratings add, or does the voltage add evenly
>> across each cap?  In other words, if the string has one 20kv cap and 
>> four 1.5kv caps with 25kv across the
>> string, does each cap see 5kv and the 1.5kv's blow, or does the 
>> rating of the string equal 26kv and everything
>> is fine?
> If the 20kv is 0.05mfd and the four 1.5kv are 0.5mfd, then the 20kv 
> will see 10x the volts the 1.5kv?  Will this work?
> Thanks again
> Marko
>
>> Thanks,
>> Marko