[ARC5] Use of a Variac

Thu Apr 23 17:58:26 EDT 2020

Thanks for the info, Tom.  I might add for the group's benefit that the
goal is not zero leakage.  Unlike a non-electrolyic cap, the dielectric is
breaking down and reforming while in a dynamic equilibrium; it's
manifestation is leakage current.  If the current is too high then the
quality of the reformed oxide is low and the subsequent heat buildup would
only serve to hasten its demise.

For an important piece of gear I will reform in situ with a current probe
(if physically possible) or else remove and reform.  We go through so many
pieces of donated gear at the California Historical Radio Society that we
often have to take shortcuts like simply ramping up with a variac using a
current meter on the variac side and an NBS calibrated nose on the chassis
side.

Cheap thrills alert!
If one puts an analog current meter in series with the cap it is
interesting to watch the forming process in real time.  It's not uniform as
one might imagine.  The current will start high then descend, then suddenly
increase, then drop down and then jump up again in an erratic manner.  It's
a bumpy ride as new oxide layers form upon the old low quality oxide.
Eventually it settles down and that's when the forming voltage is bumped up
to the next level.  It will be non-uniform again but not nearly as much as
when those first new layers were formed.  The subsequent forming voltage
increases usually go fairly smoothly and that's a good sign the ultimate
result will be a good one.

Dennis AE6C

On Thu, Apr 23, 2020 at 2:20 PM Tom Lee <tomlee at ee.stanford.edu> wrote:

> I'm a big believer in re-forming electrolytics if you don't have
> replacement caps handy. I have saved many pieces of kit that way. Not
> all caps are salvageable, but many are. The key is to do exactly what
> you describe: Free their terminals, then apply voltage with a
> current-limited suppy. The open-loop variac method can work (and has,
> for many), but it has its risks of permanently damaging the cap. The
> current limit is key. Set it so that the max power dissipated by the cap
> is below a half-watt or so.
>
> How do you know when you're done? The standard pass/fail criterion is
> that the capacitor should have a self-leakage that corresponds to no
> more than 1% loss of the stored charge per second at the working
> voltage: Ileak < (0.01per second)*C*Vmax, where Ileak is in amps, C is
> in farads, V is in volts.
>
> If the leakage can be brought down below that limit, the cap is probably
> going to give good service. I most recently did this on a Tek 485 scope.
> I have new replacement caps waiting in a box, just in case, but the
> thing has been going strong for well over a year since re-forming.
>
> --Tom
>
> --
> Prof. Thomas H. Lee
> Allen Ctr., Rm. 205
> 350 Jane Stanford Way
> Stanford University
> Stanford, CA 94305-4070
> http://www-smirc.stanford.edu
>
> On 4/23/2020 13:39, Richard Knoppow wrote:
> >    I think the idea of the Variac to bring voltages up slowly is based
> > on not having filter caps short. I don't think it really works. What I
> > used at -hp- and have now is a metered Variac. If something isn't
> > right the current meter will come up too fast and I can hit the off
> > switch quickly. Still won't protect something that has a shorted power
> > transformer. I am also not a believer in re-forming electrolytic caps.
> > While there is a long military treatise on this I think there are some
> > assumptions made about how caps fail that are wrong. If you want to
> > reform the caps do it with a variable, current controlled, DC supply
> > with the caps out of the device. A metered Variac is very useful
> > because you can tell right away if something is drawing excessive
> > current, or maybe too little current. A dim light tester is better
> > because it will show immediately  if something is shorting.
> >
> > On 4/23/2020 1:18 PM, Robert Eleazer wrote:
> >> We frequently hear people suggest that you bring up the input voltage
> >> on an older piece of equipment with a variac to allow capacitors to
> >> fromate and so forth.  Of course this remains as a somewhat
> >> controversial approach.
> >> I have a URM-127 that I need to recalibrate the frequency dial and in
> >> the process of researching that procedure I found the following:
> >> /a. /If the signal generator has been stored in excess of two years,
> >> perform the following pre-test procedure to assure proper operation
> >> of the power supply filter capacitor.
> >>
> >> (1) Operate the signal generator for 30 minutes at 57.5 VAC, 60-400
> >> Hz, 1 phase input.
> >>
> >> (2) Adjust the input to the required value using a variable
> >> transformer between the power supply and the power cable. Check input
> >> voltage prior to connecting to the signal generator.
> >>
> >> (3) Remove the variable transformer from the power cable and connect
> >> the power cable to the electrical power source.
> >>
> >> Wayne
> >>
> >> WB5WSV
> >>
> >>
> >> <
> http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient>
>
> >>     Virus-free. www.avg.com
> >> <
> http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient>
>
> >>
> >>
> >> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> >>
> >> ______________________________________________________________
> >> ARC5 mailing list
> >> Home: http://mailman.qth.net/mailman/listinfo/arc5
> >> Help: http://mailman.qth.net/mmfaq.htm
> >> Post: mailto:ARC5 at mailman.qth.net
> >>
> >> This list hosted by: http://www.qsl.net
> >> Please help support this email list: https://www.qsl.net/donate.html
> >
>
> ______________________________________________________________
> ARC5 mailing list
> Home: http://mailman.qth.net/mailman/listinfo/arc5
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:ARC5 at mailman.qth.net
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: https://www.qsl.net/donate.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/arc5/attachments/20200423/f30277d7/attachment-0001.html>