Old capacitors present several types of problems for vintage radio restorers:
- their identification - nature and characteristic values
- their replacement with current components
- their eventual reforming (electrolytic capacitors)
Can an old electrolytic capacitor be revived?

Electrolytic capacitors consist of 2 metal armatures between which a liquid electrolyte bathes. Upon initial startup, an insulating layer "forms" on one of the plates. This extremely thin layer easily gives the capacitor a capacitance in multiples of 10µF. This type of capacitor is polarized. After a long period of non-operation, the oxide layer deteriorates. To reconstitute it, it must be "reformed". That is to say, let a direct current pass until the insulating layer is reconstituted. The value of the repolarization current will be limited to less than 10 mA to avoid heating that could damage the component. The depolarized capacitor behaves like a resistor varying from a few Ω at the beginning of the operation to a few MΩ at the end of the operation. Therefore, on the one hand, the current will be limited at the beginning of the operation and on the other hand, this current will tend towards a non-zero limit at the end of the operation. If after several hours the current in the capacitor does not decrease significantly, discard it.
Principle diagram
The reforming voltage will be drawn from the 220V mains behind an isolation transformer. In half-wave rectification, this gives 310V DC.
Warning! the voltages involved
can be lethal

Reforming capacitors
(H-L Jeanmonod) I reformed capacitors that
were short-circuited, on a Philips 510A, successfully, more than
five years ago and the radio still works, even without being used often. The
method only applies to Philips capacitors with liquid electrolyte,
recognizable by the sound of liquid when shaken.
It must be ensured that the capacitor is sealed and has no leaks,
otherwise, the method of emptying and replacing the inside
with a new capacitor must be used. I start by applying a low voltage,
from a laboratory power supply that goes up to 24 Volts. If the
capacitor is short-circuited, the current rises and then after a certain
time decreases to become almost zero.
At this stage there are two methods,
either you reform it with the power supply of the radio it is in, or
you can reform it with an adjustable HT power supply (example fig1). If
you want to reform it in the radio itself, you need to make an assembly of two
1N4007 diodes and several 10k Ω 2 Watt resistors. You will connect
the anodes of the two diodes in place of the anodes of the HT rectifier,
removing it from the chassis, as well as the other tubes. The 10k Ω resistors
in series to form a tapped resistor of 10, 20
,30, 40, 50 k Ω. One end of these resistors will be connected to the
cathodes of the two diodes. Using a movable wire, the positive pole will be connected
of the capacitor via a milliampere meter to the last resistor
of 10 k, i.e., 50 k in total. The device is switched on and the
current is monitored, which must not exceed 5 mA. Then we will move to the next
tap by setting 40 k in total, if the current is greater than 10 mA, it
is necessary to return to 50 k and let the capacitor form, the current decreases
little by little, while the voltage rises across its terminals. Then the resistance is decreased
in successive steps of 10 k, as the
capacitor forms, while monitoring the mA-meter which must not exceed
10mA. A voltmeter can be simultaneously connected to the capacitor terminals
and observe its forming voltage rising. If 10mA is exceeded, the
capacitor heats up, and if it boils, a
pressure cooker
effect occurs, the overpressure electrolyte spurts out through the safety valve,
spraying everything in the vicinity. Therefore, it is not necessary to leave
a capacitor reforming unattended, and interrupt the operation if one
is absent to resume it later.
The same must be done if excessive heating is observed; generally,
once cooled, a few volts have been gained when the operation is restarted.
The operation can last from less than an hour to several hours depending on the degree
of deterioration of the alumina oxide layer in the capacitor. When
the service voltage indicated on the capacitor has been reached and the
leakage current is less than 2 mA, the operation has been successful. The best way to
automate the operation would be a constant limited current power supply,
adjustable from 0 to 20 mA and whose
voltage would vary proportionally up to 350 volts.
Acceptable leakage current
(Joseph-Henri Lévy)I found in the user manual of a SPAGUE TO-05 capacitor tester, a table of maximum leakage current values for electrolytic capacitors based on capacitance and service voltage. A small calculator can be found further down this page.
3 - 100VDC
C(µF) | ImA |
1 | 0,31 |
2 | 0,32 |
5 | 0,35 |
10 | 0,40 |
20 | 0,50 |
30 | 0,60 |
40 | 0,70 |
50 | 0,80 |
70 | 1,00 |
80 | 1,10 |
100 | 1,30 |
125 | 1,55 |
130 | 1,60 |
150 | 1,80 |
200 | 2,30 |
250 | 2,30 |
500 | 5,30 |
1000 | 10.00 |
1500 | 10,00 |
2000 | 10,00 |
3000 | 10,00 |
101 - 250VDC
C(µF) | ImA |
4 | 0,38 |
8 | 0,46 |
10 | 0,54 |
12 | 0,54 |
15 | 0,60 |
16 | 0,62 |
20 | 0,70 |
30 | 0,90 |
40 | 1,10 |
50 | 1,30 |
60 | 1,50 |
70 | 1,70 |
80 | 1,90 |
100 | 2,30 |
120 | 2,70 |
125 | 2,80 |
140 | 3,10 |
150 | 3,30 |
200 | 4,30 |
300 | 6,30 |
251 - 350VDC
C(µF) | ImA |
4 | 0,30 |
8 | 0,50 |
10 | 0,55 |
12 | 0,60 |
15 | 0,68 |
16 | 0,70 |
20 | 0,80 |
30 | 1,05 |
35 | 1,18 |
40 | 1,30 |
50 | 1,55 |
60 | 1,80 |
80 | 2,30 |
100 | 2,80 |
120 | 3,30 |
125 | 5,43 |
150 | 4,05 |
200 | 5,30 |
351 - 500VDC
C(µF) | ImA |
2 | 0,38 |
4 | 0,46 |
5 | 0,50 |
8 | 0,62 |
10 | 0,70 |
12 | 0,78 |
15 | 0,90 |
16 | 0,94 |
20 | 1,10 |
25 | 1,30 |
30 | 1,50 |
40 | 1,90 |
50 | 2,30 |
60 | 2,70 |
80 | 3,50 |
90 | 3,90 |
125 | 5,30 |
A tool for calculating max. leakage current is available on this site.
Sources and references
[1] "Technologie des composants électroniques, Tome 1, R. Besson, Edition Radio, Paris, 1977.
[2] "Operating manual, Model TO-5, Capacitor analyzer and turns ratio bridge", Sprague products company, 1960