Archived from groups: rec.audio.pro (
More info?)
In article <p5ednUH06ZnXfbvcRVn-ow@comcast.com> arnyk@hotpop.com writes:
> Ahem, I studied this at length and found that ESR is THE standard
> measurement for evaluating the life of capacitors, these days. Given that
> ESR testers go back to the vacuum tube days, it looks like this is nothing
> new!
Can you point me to a good piece of reading matter on the subject,
particularly relating ESR to failure modes? I thought I had a general
idea of what ESR is and that it's desirable that a capacitor be
designed and built so that it is as low as possible. From the bit of
reading I've done, my sense is that this is a calculated value rather
than one that's directly measured, and that what's measured is
actually leakage (as well, of course, as capacitance) when evaluating
the design and health of a capacitor.
I wouldn't be surprised if one reason, perhaps the primary reason, for
rising ESR over the life of a capacitor is deterioration of the
electrolyte or deterioration of the plates, either of which could
cause a reduction in capacitance. While ESR might be a good production
line QA tool, if you wanted to know if a capacitor was still
functional in situ, wouldn't it make more sense to measure the leakage
(if its function is isolation) and capacitance (if its function is to
create a frequency-dependent circuit)?
I don't have to know why a power supply filter capacitor marked
5,000 uF is, 27 years later, only 2,000 uF, but if it is, I can
predict that there will be more ripple coming out of the power supply
than the designer originally intended to live with. Therefore, the
capacitor should be replaced.
Similarly, if a 100 uF coupling capacitor is now 5 uF, I can expect
that unless the value was much greater than required by the design,
low frequency response will be affected and the capacitor should be
replaced.
If, with or without a change in capacitance, the ESR has risen so that
it acts like a resistor in series with the capacitor, I could also
predict a change in frequency response. If the actual leakage has
increased so that DC can flow through the capacitor, I could expect a
bias problem if it was a coupling capacitor or a fuse blowing if it
was a power supply filter.
With all of those good diagnostic tools available, why worry about
ESR? What else is there that a capacitor should do that would be
predicted by an ESR measurement and not be apparent by some other
bench diagnostic means?
I'm talking low frequencies here, of course, and I recognize that even
in audio circuits, performance outside the audio band is important, so
I won't say I have the complete picture here, I just don't know what
else I'm missing that I might care about.
> Since ESR is so widely used as *the* criteria for end-of-life I thought
> that somewhere I'd find a chart of capacitance versus ESR, but no dice.
I wouldn't expect them to be directly related, other than perhaps for
a specific capacitor construction and value.
> On
> balance, capacitance seems like a poor parameter to track, because initial
> capacitance can vary so much.
That's why designers pay attention to tolerances put on the capacitors
by their manufacturers. It's also why electrolytic capacitors, which
generally have pretty wild tolerance, aren't used for precision
frequency determining components. So it's reasonable to design based
on the low end of the tolerance and say that when the capacitor no
longer meets that tolerance (for whatever reason), it's time to
replace it.
> It's possible that virtually every cap that has loss significant amounts of
> capacitance, also has ESR problems.
If ESR manifests itself in the functional world as I think it does,
the same thing that causes loss of capacitance is likely to cause a
change in ESR. However, there may be applications for capacitors where
ESR is not all that important as long as capacitance doesn't change
along with it.
Bottom line - I think I can tell whether a capacitor is doing what
it's supposed to do in the circuit in which it's installed without
knowing anything about ESR. If I was designing a circuit, I might want
to know about the ESR of the capacitor I intend to use becuase there
may be advantages to choosing a capacitor with a low-ESR design.
(Since this seems to be an advertised feature, I assume it increases
cost, if for no other reason than to cover the cost of documentation
and advertising.) Or it may be that for the function I intend for that
capacitor, ESR (particularly over the product's expected lifetime) may
not matter a bit, so I can make my marketing department happier by
choosing a cheaper component.
Let's not confuse theory of capacitor design with practical circuit
design or bench troubleshooting.
--
I'm really Mike Rivers (mrivers@d-and-d.com)
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo
/bkprecision.com/www/np_searchmodel7.asp?lf=ESR+Meters>
Facebook
Twitter