[Boatanchors] Q of electrolytic capacitors

[Brian Clarke]

Charge it up.
Disconnect everything.
Measure the unloaded Voltage with a DMM or a VTVM.
Disconnect the Voltmeter.
Connect a 100 k resistor across the capacitor. Start a stopwatch at the moment of 
connecting the load.
Dab the Voltmeter on from time to time to see when the Voltage has fallen to 37% of 
the initial Voltage. Note exactly how long this has taken.
Now, use t=CR to calculate the leakage resistance. Remember that the leakage 
resistance is in parallel with the 100 k.
Bear in mind that the tolerance on capacitors of yesteryear was -50% to +100%.

Anyway, let's assume that you charged the capacitor to 350 V. 37% of this is 130 V.
The time constant of 100 uF and 100k is 10 sec. If the Voltage falls to 130 V in 2 
sec, the cap is gone; in 5 sec, it's on the edge. So, if the lightly loaded Voltage has 
fallen dramatically before one time constant, the capacitor is probably a dud anyway.

But, why are you concerned with the leakage resistance? It has almost no effect
as a smoothing capacitor. Measure the ESR.

73 de Brian, VK2GCE.
  Yong Surk asked:

  Thanks for your reply. What I want to know is how to test old electrolytic
  caps to see if they leak or not. I thought a leaking cap is equivalent to a
  resistor connected in parallel with the cap. How do I find out if an old cap
  is leaking or not? Also how much it is leaking.