RE: DTSSTC electrolytics for performance?

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Mccauley, Daniel H" <daniel.h.mccauley@xxxxxxxx>



This is all good in theory, but in practice this type of equation
doesn't quite work as there is a finite
impedance between the charging voltage (AC line) and the capacitor.  The
AC line usually will not be able to
recharge the capacitor to full capacity during each cycle.

What happens instead, is that during operation, the voltage(t) on the
cap will reach some steady-state condition and the voltage will be much
lower than what your calculated bus voltage will be.  So you have to
look at several things to determine exactly what is happening at the
capacitor.

Also, the voltage seen at the bridge will be reduced due to the ESR of
the capacitor as well.  As soon as a pulse initiates, the voltage at the
capacitor will drop at a rate of Ipeak * CapESR and can be considerable
especially if peak current is very high.  This is why its important to
use high quality electrolytics such as Powerlytics etc...

Dan


>I will make a theoretical example of your coil, assuming it draws say
>1kw at say 200bps, thats 5j a bang.
>
>Energy in a 1500uF 350V cap : .5(.0015F)(350v^2)=~92J
>Energy left in cap after bang : 92J - 5J = 87J
>Voltage in cap at 87J : sqrt(87/.5C) =340.6V
>
>But since you hit the cap twice between recharging on the mains cycle,
>you lose that energy again (and repeating the above calculation, but
>assuming 340.6V now) you are down to about 330V on the caps.
>
>I personally dont think losing 10V on each cycle is too bad, but 20V
>would be a bit much in my opinion, since the second bang will be that
>much lower in energy.  Doubling the cap to 3000uF is a perfectly fine
>idea in my opinion ;-).  Lets the IGBTs go with more of a bang too!
>
>Steve Ward