Alan Ibbetson wrote:
>I have some old capacitors I want to use in a simple bridge plus
>capacitor PSU. They are marked "Dubilier Duconol 50uF 2500v Photoflash
>Made in England TF".
>
>I have searched the amps archives and note that folks suggest that
>derating this type of capacitor to 50-66% of its marked voltage makes
>it OK for mains PSU use. Why? Electrolytic capacitors are rated for
>ripple current, which obviously translates to heat via ESR. What is the
>physics behind the suggested voltage derating for oil filled paper
>block capacitors?
>
>Oh, and once derated, how do I figure out how much ripple current they
>can withstand, short of giving them some stick and risking a burst can
>and spilled PCBs?
>
>Should I just bin them and go buy ten new 450v electrolytics?
Yes, bin them - now.
At 2.5kV they will work well for maybe two years of intermittent use,
and then suddenly fail dead-short. There's no drama, nothing bursts and
AFAIK they don't contain PCBs anyway... but they *will* eventually fail.
A few batches of these came on to the UK surplus market at the height of
the VHF DX boom, and many went into power supplies for 4CX250 amps. Many
of those failed, and probably the rest of them are destined to. When
mine failed, I was lucky to find a proper PSU-rated paper cap and that
has been just fine for more than 10 years.
Overheating due to ripple current is the root cause of failure in
power-supply use. For photoflash use the internal construction aims to
minimize inductance, but the designer didn't even have to think about
heat transfer from the inside to the metal can. Even though the metal
can stays cool to the touch, I suspect that there is significant heating
deep inside, which eventually weakens the dielectric until it fails.
Voltage de-rating does seem to help, but we cannot know whether it will
completely prevent failure, or merely delay it.
--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek
|