On Jan 26, 10:28*pm, "Phil Allison" wrote:
"George Herold is a real Pain "



In other words... If the driving frequency is high enough that the
capacitive reactance is essentially zero, then the net impedance is
essentially the ESR. Right?

** Errr - yep.

But it ain't that simple.

One has to examine the actual impedance curves for typical electros to see
what the game is - the curve is like no other kind of cap.

Think of Q factors of circa 0.05 and ESRs that way exceed the calculated
impedance at 100kHz.

Nice tip thanks Phil. *The 100uF electro's I have came out with ~0.15
ohms of ESR, using your method.
I've never seen an impedance curve for an electro cap. *Do you have a
link or know which manufacturer's website I might check?

** Found this PDF on the net *- seems to have lots good info on the humble
electro.

http://www.epcos.com/web/generator/W...Catalog/Capaci...

See page 15, *figs 13 and 14.

One can easily see the dramatic effect temperature has on the impedance
minima or ESR *- *electros work better when they are HOT !!

Also, the minimum impedance value ( same as the ESR ) occurs around 50 -
100kHz and is quite broad - the *higher the ESR and the lower the temp the
broader.

At 20C the 47uF, 350 volt electro in fig 14 exhibits a deep impedance minima
( essentially 0.4 ohms resistive ) from 10kHz to 2 MHz.

.... *Phil

That's great! Thanks Phil, I found some graphs by Vishay, but they
weren't nearly as nice. (No temperature dependence and frequencies
only to 100kHz) Of course at high temperatures the electro's leak
more.

I used your techinque to look at a 1000uF cap. I used a DSO with a
bit of signal averaging. ('scope triggered from a sync output from the
generator) I got something like 0.016 ohms at 100kHz. Our SRS RCL
meter measured 0.014 ohms at 10kHz for the same cap. and then 'lost
it's mind' at 100kHz. (reported a negative capacitance.)

George H.