Discharging those big old capacitors

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Advice please. I want to make some mods to my valve amp. Clearly there are some big voltages kicking around and some capacitors that I wish to change include 100uF 450V. What are the safety precautions I need to take here. I have nice thick gloves goggles et al but am not clear on safe discharge and storage of capacitors.
 
First, disconnect the amplifier from the AC line. If you have a voltmeter, clip the negative lead to the chassis, or negative side of the capacitor, and using one hand only, check the voltage at the positive terminal of the capacitor. If the circuit uses bleeder resistors, the capacitor voltage will drop to nearly 0V in a couple minutes on it's own. If not, solder a couple leads with clips to a resistor with at least a 2 watt rating. The value of the resistor isn't critical, but 10K will discharge a 450V 100mfd cap in less than five minutes. Clip the resistor across the leads (again using one hand) to discharge it, and remove the capacitor from the circuit. Resist the temptation to discharge the cap by shorting the terminals with a screwdriver or something similar, as the high current "jolt" can permanently damage the cap.

Leave the resistor connected to the capacitor for a day or so after it is removed, since dielectric absorption will cause electrolytic caps to "charge up" again after being apparently fully discharged. Examine the removed cap for any signs of bulging or and leakage at all. If any is suspected, discard the cap.
 
Member
Joined 2009
Paid Member
Eh.... 10 kOhm across 450 V will cause 20 W to be dissipated in the resistor. With high voltages it's really easy to get multiple watts of dissipation in resistors that are considered "high value" in the "sand" world. Do the P = E^2/R math and use a safety factor of at least 2 (for prototypes) or 4~5 (for permanent installation).

As an estimate, you can calculate the amount of time needed to fully discharge the capacitor as t = 5*R*C (R in Ohm, C in Farad, t in seconds).

In addition, note that many tube rectifiers have max specs for the amount of capacitance they can drive. Check the spec sheets or use the same value as was in the original equipment.

~Tom
 
One variation on the 'resistor with two leads and clips' theme is to solder one (insulated) lead to the resistor. Then, use heat-shrink to fasten the resistor to the end of a wooden or plastic chopstick so that the resistor lead is protruding 1/4" or so. Clip the lead to chassis ground, and hand-hold the 'probe' against the + point on the capacitor.
I don't like sticking my fingers into a chassis which may have charged caps. 'Grabber' adapters on the voltmeter are handy as well - lets you monitor how the voltage is dropping.

John
 
As an estimate, you can calculate the amount of time needed to fully discharge the capacitor as t = 5*R*C (R in Ohm, C in Farad, t in seconds).
Thanks, Tom.
That formula shows that for common tube PS rectifier sizes you can increase the resistance quite a bit, thus lowering the watt-rating required on the resistor.
For instance, the 10k resistor would discharge a 40uF cap in 2 seconds, by my calculation. (And the voltage would be below dangerous levels before that..)
So, a higher ohm-value resistor and a bit of patience will do the trick....
 
The capacitor memory voltage only applies to caps that were in DC circuits for a period of time and then removed from the circuit. These caps will act in a battery like fashion to redevelop a DC voltage while still out of circuit.

A cap in an AC circuit, that is quickly removed from the circuit could have a voltage up to the peak AC voltage. But this voltage won't have the long term memory that a DC cap has.

A cap that remains in an AC circuit when power is turned off will dis-charge itself through the circuit.
 
A cap in an AC circuit, that is quickly removed from the circuit could have a voltage up to the peak AC voltage.

It takes a finite time to charge a capacitor, I thought. I guess my question was:
"Will a 'large' capacitor charge to 120 volts in 1/240 second, when a voltage rising to 120 volts max is applied?"

Also, will arcing across the switch contacts not discharge the cap to some extent?

It would be easy enough to answer this on the bench with a voltmeter and a switch, but I'm not at home just now.

John
 
Member
Joined 2005
Paid Member
The capacitor memory voltage only applies to caps that were in DC circuits for a period of time and then removed from the circuit. These caps will act in a battery like fashion to redevelop a DC voltage while still out of circuit.

A cap in an AC circuit, that is quickly removed from the circuit could have a voltage up to the peak AC voltage. But this voltage won't have the long term memory that a DC cap has.

A cap that remains in an AC circuit when power is turned off will dis-charge itself through the circuit.

Through the circuit might be an issue. I would feel better about dissipating it. What I am unsure of is how much current will discharge - which I can not measure.

My calculations come out that 400ohm would take .04 seconds and have a discharge around .3a (if full charged) at 120v. That is less than the rated wattage for one of the Mills 50w resistors and on top of that that low of a time would not even effect it with heat. It seems like that would work very well. I could increase the resistance but keeping time down would be best. I think switching to 1000ohm would be fine as it would be 1/10th of a second and much lower wattage. My only reason is to make sure that when the circuit is on that the resistor do as little as possible. That being said I could go with 5000ohm and have a half second discharge time. I am not sure what reason I have to pick one over another.
 
My calculations come out that 400ohm would take .04 seconds and have a discharge around .3a (if full charged) at 120v. That is less than the rated wattage for one of the Mills 50w resistors and on top of that that low of a time would not even effect it with heat. It seems like that would work very well. I could increase the resistance but keeping time down would be best. I think switching to 1000ohm would be fine as it would be 1/10th of a second and much lower wattage. My only reason is to make sure that when the circuit is on that the resistor do as little as possible. That being said I could go with 5000ohm and have a half second discharge time. I am not sure what reason I have to pick one over another.

Are you talking here about a bleeder resistor to be wired permanently into the circuit? And exactly where in the circuit is the cap located?

John
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.