From morga–(at)–peckle.ncsl.nist.gov Thu Sep 21 17:20:10 CDT 1995
From: Roy Morgan
Subject: Re: Forming caps
Date: 21 Sep 1995 16:41:36 GMT
This is a multi-part message in MIME format.
— Roy Morgan / Tech A-266 / NIST / Gaithersburg MD 20899
(National Institute of Standards and Technology, formerly NBS)
301-975-3254 Fax: 301-948-6213 Internet: morga–(at)–peckle.ncsl.nist.gov
To Re-form electrolytic capacitors:
With the “patient” set off, set the external supply at the rated voltage
of the cap(s), and feed the old set at the input to it’s B+ filter through
a 100K, 2W resistor. The old caps will slowly come up to voltage as
their elecrolytic layer re-forms after long storage. You may want to
unhook bleeders or screen voltage dividers if present in order to get no
dc load other than the caps. Once re-formed up to nearly the cap rating,
increase the external supply voltage to the point where increased voltage
only increases the current drawn (the electrolytics begin to “leak”.) You
can vary the series resistor depending on the voltage of the cap you’re
trying to reform.
If you have no external supply, remove all tubes except the rectifier, place the high resistance in the circuit between the rectifie=
r and the filter capacitors and use a variac to raise the voltage slowly.
If the final cap(s) voltage is high enough, it doesn’t need to be
replaced. If it’s too low, put new one(s) in (leave any original cans in
place for appearance, and substitute new axial lead ones under the
Some caps take only a few minutes to re-form. Some take a day or so! Be
patient. Your Adjusta-Volt or Variac can be well-used for this if your
external supply is solid state, or has a separate hv supply transformer.
I have one good for 900 volts no-load having 5R4’s and separate filament
transformers. This lets me re-form 500 volt electrolytics if I need to.
With a 500 volt supply, and a number of 100k or 200k resistors, you can
re-form a number of caps all at once. Measure the voltage on the caps as
time goes on with a high-input-resistance meter (VTVM or solid state
DVM). Allowing an electrolytic to idle with a small leakage current of 1
to 5 ma won’t hurt it, so if the thing re-forms to it’s limit during the
night after you’ve left it on the re-former, no harm is done.
Most electrolytics in good health will leak at a voltage from 125 to 200
percent of the continuous rating. If the leakage voltage is only a little
above the needed circuit voltage, or is below about 110 percent of the
cap’s rating, then you can excpect it to not live too long. New axial
lead caps are fairly cheap, and are good peace of mind in my opinion.
PAPER COUPLING CAPS:
Test interstage coupling caps (e.g. from an audio driver tube to the grid
of the output amp tube) by measuring the dc voltage at the grid (across
the grid resistor if it’s not going to ground). Use a high-impedance
voltmeter like a VTVM or DMM. If it’s above zero, you need a new cap!
The vast majority of paper caps from the 30’s through the 60’s are at
least moderately leaky now. Your tubes will thank you with long life for
replacing these caps. Most disk ceramic caps have indefinite life expectancy,
as do good quality modern film caps.
You can do this kind of testing while you are re-forming the filter caps
in-circuit. The tubes are off, and will not be harmed by excessive plate
current while you find all those leaky paper caps. The voltages across
them will be higher than normal running conditions, because the driving
stage is not drawing any plate current.
SCREEN BYPASS CAPS:
With B+ applied and the tube pulled or set off, the voltage at the screen,
again measured with a high-impedance voltmeter, should be the full B+ or
value at the other end of the screen dropping resistor. If not, the cap is
Set your high-impedance voltmeter to a high-enough range and clip one end
of the cap to the DC probe and connect (carefully) the other end to a B+
supply corresponding to the rating of the cap. The meter will jump up
briefly and then settle down toward zero. Analog meters (VTVM’s) are good
for this because you can watch the movement of the needle. Once the
reading settles, any indication much above zero indicates leakage. A
quick ohms-law estimate with the input resistance of you meter will give
you a value for the leakage. DVM’s are often 10 megohms, and so the
leakage will be indicated at about 10 volts per microampere.