Re: Unknown type of HV capacitor

["Jim Lux" <jimlux-at-jpl.nasa.gov> (by way of Terry Fritz <twftesla-at-uswest-dot-net>)]

No... all you really want is to have the source be high enough impedance
that it doesn't kill the Q of the capacitor.  And, you want the resistance
low enough so that the RC time constant of the R and C is fairly short (so
you get a good step into the cap.)

You could use a capacitively coupled pulse generator, or something like a
transmission line pulse generator... It's a bit tricky because you want
good HF response (to put all the power into the RLC, so it rings) but you
also don't want to load it.  Go far enough with this and you wind up with a
RLC bridge...

Try it on some caps you have lying around first.... like those ceramic 0.1
bypass caps, for instance...

----------
> From: Tesla list <tesla-at-pupman-dot-com>
> To: tesla-at-pupman-dot-com
> Subject: Re: Unknown type of HV capacitor
> Date: Tuesday, August 29, 2000 5:39 PM
> 
> Original poster: "John Sanderson" <john.sanderson-at-eng.monash.edu.au> 
> 
> Thanks for the test suggestion, Jim.  When I get the chance to carry 
> out this procedure, I'll post the results.  (May take a while, 
> though, because some travel is involved.)
> Regarding the resonant frequency test, would it perhaps be better to 
> drive the "suspect cct" through a small capacitor (say 1nF) than a 
> resistor?  
> 
> John S.
> 
> > drive the "suspect circuit" through a reasonably large resistor (so the
> > generator's low output impedance doesn't perturb the circuit).  Hook
the
> > scope probe across the cap.  When you look at the square wave, you'll
see a
> > damped sinusoid on each of the edges.  The frequency of the sinusoid
gives
> > you the resonant frequency  (approximately f = 1/(2*pi*sqrt(LC), but
this
> > doesn't allow for the R, which, if the Q is low will "pull" the
frequency a
> > bit) and the decrement ratio lets you calculate the Q and hence the R
(since
> > you know the C, and the fRes, you can calculate Xc... Q is Xc/R)
> > 
> > This is good for a quick and dirty evaluation, but not so hot for
precision
> > measurements, where some sort of LCR bridge would be a better bet.
> > 
> > 
> > >
> > ><< If you can somehow measure the L and R... A square wave generator
of a
> > few
> > > kHz (say a TTL pulse generator or a 555), a series resistor and an
> > > oscilloscope will tell you... Measure the ringing frequency and the
> > > decrement ratio...  If the L is down in the nanohenries, and fRes is
in
> > the
> > > hundreds of kHz, you might be in luck. >>
> > >
> > >
> > >Would you please elaborate on this procedure for us uneducated 'scope
> > owners.
> > >
> > >Thank you,
> > >Marc S.
>  
> 
> 
>