[ARC5] Electrolytics- How Much is Too Much

[Richard Knoppow]

----- Original Message ----- 
From: "Kenneth G. Gordon" <kgordon2006 at frontier.com>
To: "ARC-5 Mail List" <arc5 at mailman.qth.net>
Sent: Wednesday, January 30, 2013 9:07 AM
Subject: Re: [ARC5] Electrolytics- How Much is Too Much


> On 30 Jan 2013 at 5:59, David Stinson wrote:
>
>> High Voltage electrolytics- how much current leakage
>> at rated voltage is "too much?"
>
> There have recently been published some details on 
> allowable leakage in
> NEW electrolytics. I'll try to dig that out and publish it 
> here soon. It varies with
> capacitance and voltage rating.
>
> However, the permissable leakage for OLD electrolytics was 
> different,
> higher, since those weren't as good as modern ones.
>
> The apples and oranges thing...
>
>> Say- I have an original can capacitor,
>> rated 40-40-40 at 450 volts.
>> After sitting with all three tied together and at
>> 400 volts, total leakage was about 1.5. mA.
>
> So to my way of thinking, that is too much. Think of it 
> this way: P=EI: in this
> case, that is 0.6 watts of leakage, and the heat resulting 
> from it.
>
> Since the can is sealed up, that heat just sits there. 
> After a while, the can
> should feel noticeably warm. I don't believe electrolytic 
> caps should feel
> warm at all....but maybe that is just my own prejudices.
>
>> only in what level of leakage is, in your opinion,
>> acceptable in order to maintain a stock part in a set.
>
> If that were my unit, I would replace it. Tom, at 
> www.hayseedhamfest.com
> has FP can-type electrolytics of new manufacture, if you 
> want to retain stock
> appearance. Otherwise, I would disconnect it, and install 
> axial-leaded new
> ones under the deck....if there is room.
>
> I believe the acceptable leakage would be that which would 
> preclude the
> capacitor from getting warm enough to feel it.
>
> Ken W7EKB

     I think the amount of leakage depends on the 
application.  Leakage is the same as _parallel_ resistance, 
that is, it acts like a resistor in parallel with the 
capacitor.  Since one purpose of a capacitor is to block DC 
a low parallel resistance may defeat the purpose even if the 
reactance and ESR seem to be OK.  For power supply filters 
leakage can result in higher than normal current, where the 
capacitor is used as a bypass capacitor for a cathode biased 
tube high leakage has the same effect as reducing the value 
of the resistor, that is, it reduces the value of the bias. 
While the increased cathode current will tend to counteract 
this it doesn't cure it and the result is high distortion 
and sometimes an overheated tube.
    The existence of parallel resistance is often forgotten: 
leakage is a measure of it but capacitors are also specified 
for equivalent insulation resistance. This is not so much a 
measure of the resistance of the case, although that is 
included, as the resistance of the dielectric. New paper 
dielectric has pretty high insulation resistance and modern 
film caps have extremely high resistance, hens low leakage. 
Electrolytic caps have rather low dielectric insulation 
resistance even when new but as they age it can go down. 
One result is heating of the capacitor which accelerates the 
degradation of the dielectric and speeds the end of life. 
Another is that more DC gets through which can upset the 
circuit the cap is used in.
    While the cap reforms to some extend each time its 
energized I have a suspicion that very old caps, with lots 
of leakage, don't ever reform properly and should be 
replaced.
    While electrolytics may fail by going open I have and 
them short. This can be a disaster if it takes out a 
transformer or choke and can damage tubes where the cap is a 
cathode bypass one.
    The values given in previous posts for leakage seem 
reasonable.  Leakage can be measured with a sensitive VOM 
and small power supply.  I have an old General Radio 
variable power supply that goes up to about 300 volts and 
can get decent measurements with a Triplet 630 type VOM. 
Its also possible to see the charging behaviour of the cap 
with this arrangement.  This is often a clue as to the 
condition of the cap. A good cap will charge up and reach a 
minimum current, the time depending on its value. Caps in 
poor condition continue to draw high current even after a 
reasonable charge time.  This is also true for paper caps.
     BTW I have found several leaky paper caps that tested 
good for value and dissipation factor (series resistance) on 
an impedance bridge.  They did not work in circuit.
     BTW, its possible to measure resistance with any 
voltmeter but if you have a VTVM with very high input 
resistance like a Hewlett-Packard 410 A/B/C you can measure 
extremely high resistances with the aid of a source of 
voltage.  The technique is described in the 410-B manual at 
BAMA.  This technique is useful if you do not have a megger 
or if you need to measure very high resistance values with 
lower voltages than a megger generally uses.


--
Richard Knoppow
Los Angeles
WB6KBL
dickburk at ix.netcom.com