Re: Measurements of a cap's ESL, ESR (fwd)

---------- Forwarded message ----------
Date: Mon, 25 May 1998 09:03:15 +1200
From: Malcolm Watts <MALCOLM-at-directorate.wnp.ac.nz>
To: Tesla List <tesla-at-pupman-dot-com>
Subject: Re: Measurements of a cap's ESL, ESR (fwd)

Hi Bill,

> Date: Sat, 23 May 1998 05:47:53 PDT
> From: Bill the arcstarter <arcstarter-at-hotmail-dot-com>
> To: lau-at-hdecad.ENET.dec-dot-com, tesla-at-pupman-dot-com
> Subject: Re: Measurements of a cap's ESL, ESR
> Coilers,
>  Gary Lau wrote:
> >I have done an experiment to determine the ESR and ESL parameters of 
> >my tank capacitor.  This is a surplus unit, .01 uF, 100KVDC, 
> >2.5"x5"x10", marked F-C-I, KM14-1000-10, .01 MFD-100KVDC, DEC 1983.
> ...
> No kidding!  I was doing something like this last night too!  
> Synchronicity? :)
> My work last night served as a "reality check" using lumped 
> low-frequency parts before I go measure some real (possibly 
> distributed?!) coiling components.  Gotta crawl before I can run...
> Here's what I did:
> Found a generic green polyester cap - measured 0.47 uf (not a pulse cap)
> Found a plastic spool of 20 gauge wire - used this as my air-core 
> inductor.  This measured as 0.293 mH. with a DC resistance of 0.51 ohms.
> The above measurements were performed using my Alfa LCR-24 meter which 
> is within about 2% (according to the calibration shop at work).
> I placed the cap and inductor in parallel, and drove the cap via my 
> 600-ohm output freq generator (with an additional 330 ohm in series for 
> current monitoring purposes).  I monitored the cap voltage using my Tek 
> 543B oscilliboatanchoroscope.
> The computed vs. measured resonant frequency were mostly in agreement 
> (at about 13700 Hz)
> At resonance, this LC network acted as a pure resistance of about 38 
> ohms (measured input current and voltage etc).  This did NOT initially 
> match what theory would indicate - ie - net resistance is L/(RC)

The problem is that you are mostly measuring the ESR of your signal 
generator in series with the 330 Ohm resistor. A parallel circuit 
*must* be driven by an infinitely high impedance (current source).
Try driving the circuit with 1 MOhm in series with the generator and 
wee what you get. A current transformer can be used in series with 
the tank with a low value of burden resistor for monitoring 
circulating currents (presents a very low ESR in the tank).

 > (See my derivation of that result at:
> http://www.geocities-dot-com/CapeCanaveral/Hangar/6160/resonan1.html )
> To make a long story short - this particular cap has a resistive 
> component measured as 15 ohms at 13.7KHz.  Shockingly high.
> The resistive component of the inductor was measured (at 13700 Hz) as 
> 1.5 ohms - so there was an additional 1 ohm due to (?) skin effect or 
> eddy current in the wire... or something???
> Once I'd characterized these additional resistive components - I 
> recalc'd the L/(RC) impedance value - which then fell into reasonable 
> agreement with the observed results.
> Today I'll go measure my homemade poly plate cap - I'll report my 
> findings.
> >I'm thinking that ESL is the more significant parameter in terms of 
> >being a predictor of a cap's usefulness in Tesla coil service.  I was 
> >wondering if anyone else had made similar measurements of home-made 
> >or commercial caps that I could compare my figures to?
> Really?  Why wouldn't ESR be a better metric?  FYI - Our Maxwell pulse 
> cap has a stated nameplate ESL of 0.060 uH...

Quite agree :)