Sujet : Re: discharging caps
De : bill.sloman (at) *nospam* ieee.org (Bill Sloman)
Groupes : sci.electronics.designDate : 13. Sep 2026, 07:37:12
Autres entêtes
Organisation : A noiseless patient Spider
Message-ID : <1185gap$csfm$3@dont-email.me>
References : 1 2 3 4 5 6 7 8 9 10 11 12
User-Agent : Mozilla Thunderbird
On 13/09/2026 5:43 am, john larkin wrote:
On Sun, 13 Sep 2026 02:53:12 +1000, Bill Sloman <bill.sloman@ieee.org>
wrote:
On 13/09/2026 2:12 am, john larkin wrote:
On Sun, 13 Sep 2026 01:20:46 +1000, Bill Sloman <bill.sloman@ieee.org>
wrote:
>
On 13/09/2026 12:54 am, john larkin wrote:
On Sat, 12 Sep 2026 16:24:19 +1000, Bill Sloman <bill.sloman@ieee.org>
wrote:
>
On 12/09/2026 7:28 am, john larkin wrote:
On Fri, 11 Sep 2026 17:08:00 +0100, Simon Simple
<nothanks@nottoday.co.uk> wrote:
>
On 10/09/2026 15:20, john larkin wrote:
>
<xxxx>>
2.4 KJ is about 16 firecrackers or five .45 cal bullets worth of bang.
>
If I've done the sums right, that's a capacitor bank bigger than six
beer cans storing about the same energy as one AAA cell!
>
People would notice if that were shorted.
>
Yes, I used to operate capacitor banks of maybe 2-4 times that physical
size charged to 1kV and discharged via a spark gap through a marine
sparker for seismic work. Long time ago before H&S was a thing.
>
I got one of the giant caps. It says that it's 30,000 uF but it
measures 50,000. So we'll have 4000 joules with four in parallel at
200v.
>
I think it's normal for elecs to have more initial capacitance than
the nameplate value.
>
https://www.dropbox.com/scl/fi/apargzu733abcq6ww7w3b/30mF_250v.jpg?rlkey=00osgeyqpxceowgqf7l32mlnr&raw=1
>
One of the more depressing things about electrolytic capacitors is the
need to reform them by keeping the rated voltage across them for a day
or so if they have been unused for a while.
>
The actual electrolyte is a layer of aluminium oxide on the electrode
surface, and that diffuses away slowly if the capacitor is stored
uncharged. In extreme cases you have apply the reforming voltage through
a big resistor because the thinned down layer can no longer stand off
the rated voltage. With the big resistor you can measure the reforming
current as the process proceeds.
>
Obviously the initial capacitance is higher than the rated capacitance.
That may have contributed to your initial 50mF - a +50% tolerance only
get you up to 45mF
>
I reformed a couple of electroltics many years ago when I bought some
military surplus electrolytics for my home built hi-fi. I didn't bother
to monitor the current.
>
I'm thinking that they overdo the capacitance to account for later-on
drying out or some electro-chemical effects, to ensure the guaranteed
minimum c over the (short) specified lifetime.
>
Not that I know of. Note that the tolerance printed on your can was -10%
+50%.
>
Hey, I saw that. It measured +66%.
>
Which does suggest that it had been left uncharged for long enough to
need reforming.
>
It charged to 100 volts with a power supply that was limited to 20 mA.
The rampup was very linear.
The tolerances are just arbitrary cut-offs on what's probably a Gaussian
distribution. Presumably the manufacturer samples individual capacitors
often enough to be confident that the distribution is stable, but don't
measure every capacitor that they ship. That +66% might just have been
an outlier.
>
The need to reform long unused electrolytic capacitors was well known
everywhere I worked. Nobody ever warned me about any "drying out", and
mythical electro-chemical effects weren't mentioned either. Back when I
was a graduate student in chemistry the electrochemists did get a hard
time, but back then they didn't have good tools for looking at the
microscope detail of the very thin layers where the electrochemistry
took place.
>
Caps are tricky parts.
>
And electrolytic capacitors are even trickier than regular capacitors.
Charge soak isn't nice.
>
Charged to 100v and discharged for 2 seconds, that big cap went to
about 0.25 volts and then only recharged itself to 2 volts.
>
Not bad.
>
It's more of a worry in capacitor-based A/D converters. Dual ramp and
quad ramp schemes come to mind. Not all that clearly - that was back in
1979. Going from a Mylar to polycarbonate dielectric got the soak down
far enough to make it tolerable.
I remember dual-slope ADCs. I designed a few. Very long ago. IC adc's
are dirt cheap now and super accurate. 24 bits for $2.
Sigma delta parts. Back when I last used one, in 1993, it spat out 24-bits - three bytes - but the accuracy spec was 20-bit. There were more precise versions around but they weren't dirt cheap.
Sloman A.W., Buggs P., Molloy J., and Stewart D. “A microcontroller-based driver to stabilise the temperature of an optical stage to 1mK in the range 4C to 38C, using a Peltier heat pump and a thermistor sensor” Measurement Science and Technology, 7 1653-64 (1996)
The caps that people can make inside ICs are about the best caps that
are possible.
If a little too small for you application.
-- Bill Sloman. Sydney
Haut de la page
Les messages affichés proviennent d'usenet.
NewsPortal