[ARC5] Re: [The WS No.19 Group] Variable voltage HT PSU?

[J. Forster]

Thank you for confirming what I've been saying for a couple of years by actual
measurement. I've repeatedly been told that 'paper capacitors do NOT reform'

Well, Keith's numbers demonstrate my point. It may not be strictly reforming
(repair of the oxide film in electrolytics) but it looks similar from the
terminals.

I suspect that some papers go leaky because of water absorption. If you suddenly
apply full voltage after decades of storage, the sudden localized internal
heating can damage the caps. If you apply the voltage gradually, from a current
limited source, the absorbed water will either be driven off slowly and escape
due to internal heating or be electrolyzed into H2 and O2 and escape that way.

An interesting experiment, IMO, would be to take a leaky paper cap and bake it
at say 120 F overnight and see if the leakage goes down.

My personal policy is to ALWAYS power the HT of old gear slowly when removed
from extended storage.

FWIW,
-John


Keiths at Britishnotes.co.uk wrote:

> Hi, they are paper decoupling capacitors and the one I mentioned has been
> sitting at 90 volts DC for a couple of hours now and has dropped from
> leaking 1ma to 30ua and is still dropping. The DC across it is stable and
> not falling, so the capacitor must be doing something that is causing the
> fall in leakage. The capacitance is 0.09uf or there abouts.
>
> It's not uncommon at all for paper capacitors to self repair so long as the
> leakage current isn't allowed to be so high that it causes damage to the
> capacitor or am I wrong?
>
> Keith.