Bad Cat Black Cat 30R Low B+

[vintagekiki]

About leakage current at capacitor

During prolonged non-use electrolytic capacitors, the internal resistance decreases and the capacitor begins to leak.
The leaking capacitor does not perform its filtering function and the amplifier starts to buzz.
The solution is to replace the capacitor or to start the process of recovery the capacitor. In the absence of capacitor testers, testing or forming a capacitor be performed in the home variant.

In series with a capacitor, connect resistor 100k / 2W and this circuit connect to the highest voltage source (+ HV) in power supply. The formatting (recovery) process can takes between 4 - 12 hours.
Since the leaking capacitor has low internal resistance, the voltage on the DVM will be low. Over time under voltage the capacitor dielectric recovers, its internal resistance increases, at the same time the voltage on the DVM increases.

When over time the capacitor is recovered, voltage on the DVM on it reaches a value of 90-95% of the voltage (+ HV) power supply.
If the DVM connect in parallel with a 100k / 2W resistor, on the DVM the capacitor leakage current can be accurately measured.

The 100k / 2W resistor is a current limiter. Since the limiter resistor has a round value (100k), if it is connected directly to + HV 500V, the DVM will show 500V, which corresponds to leakage current of 5mA (Ohm's law), that capacitor is short-circuited.
If instead of 100k used 1M resistor, leakage current is 500uA, and the formation time much longer.

[bobloblaws]

Correct.
The leakage resistance of the cap (Rleak) and the internal resistance of your meter (Rm) form a voltage divider and the voltmeter measures the voltage drop across its internal resistance. For calibration you need to know Rm (aka Zin or Rin), should be specified in the manual. (If you can't find it, I will show you a way to find out..)

My main meter is a Fluke 179. In the manual it shows a spec for Input Impedance for DC volts as >10M <100pF. Fluke 179-manual.pdf

Edit: Actually, I measured it across the two probes with another meter and it is right around 11M, assuming that is a valid way to measure it. You said there was another way to do it as well, or is this what you meant?

The leakage resistance of the cap (Rleak) and the internal resistance of your meter (Rm) form a voltage divider and the voltmeter measures the voltage drop across its internal resistance. For calibration you need to know Rm (aka Zin or Rin), should be specified in the manual. (If you can't find it, I will show you a way to find out..)
If Rm is known, the leakage resistance can be calculated from the ratio of the voltage across the cap (Vc) and the meter voltage (Vm) as Rleak = Rm*Vc/Vm.
It follows the lower the meter voltage, the higher Rleak and the better the cap.

In my example I assumed Rm = 1M and a supply voltage of 350V. So if the meter reads 175V (=50% of 350V), Vm = Vc and Rleak equals Rm (1M).
If the meter reads lower than 50% of supply voltage, Rleak is greater than 1M.

OK, I re-familiarized myself with the voltage divider formula. So given that my meter apparently has an internal resistance of 10M I'm looking for a reading of less than 35V, correct?

[bobloblaws]

About leakage current at capacitor

During prolonged non-use electrolytic capacitors, the internal resistance decreases and the capacitor begins to leak.
The leaking capacitor does not perform its filtering function and the amplifier starts to buzz.
The solution is to replace the capacitor or to start the process of recovery the capacitor. In the absence of capacitor testers, testing or forming a capacitor be performed in the home variant.

In series with a capacitor, connect resistor 100k / 2W and this circuit connect to the highest voltage source (+ HV) in power supply. The formatting (recovery) process can takes between 4 - 12 hours.
Since the leaking capacitor has low internal resistance, the voltage on the DVM will be low. Over time under voltage the capacitor dielectric recovers, its internal resistance increases, at the same time the voltage on the DVM increases.

When over time the capacitor is recovered, voltage on the DVM on it reaches a value of 90-95% of the voltage (+ HV) power supply.
If the DVM connect in parallel with a 100k / 2W resistor, on the DVM the capacitor leakage current can be accurately measured.

The 100k / 2W resistor is a current limiter. Since the limiter resistor has a round value (100k), if it is connected directly to + HV 500V, the DVM will show 500V, which corresponds to leakage current of 5mA (Ohm's law), that capacitor is short-circuited.
If instead of 100k used 1M resistor, leakage current is 500uA, and the formation time much longer.

Thanks for the info VK. Another school of thought is that since caps are relatively inexpensive these days and reformed caps are not as reliable as new ones, it's better to just replace them.

[vintagekiki]

https://circuitdigest.com/tutorial/what-is-capacitor-leakage-current-and-how-to-reduce-it

https://passive-components.eu/leakage-current-characteristics-of-capacitors/

https://www.youtube.com/results?search_query=Reforming+Old+Electrolytic+Capacito rs
Reforming Old Electrolytic Capacitors

https://www.youtube.com/watch?v=0pySErvzUIY
Capacitor leakage tester design & build

https://www.youtube.com/watch?v=ifWfvaIbWZY
How to Reform Electrolytic Filter Capacitors on tube radio or amp.

https://www.youtube.com/watch?v=FuanfhJTWvs
Reforming Old Electrolytic Capacitors

https://www.youtube.com/watch?v=qDABYKoVO4Q ¹⁾
3 Ways to Check Capacitors in Circuit with Meters & Testers

[Helmholtz]

My main meter is a Fluke 179. In the manual it shows a spec for Input Impedance for DC volts as >10M <100pF. Fluke 179-manual.pdf

Edit: Actually, I measured it across the two probes with another meter and it is right around 11M, assuming that is a valid way to measure it. You said there was another way to do it as well, or is this what you meant?

Fine, so your Rm is 11M. This makes the method much more sensitive. Now at 50% voltage, the leakage resistance would be 11M, an excellent value.

So given that my meter apparently has an internal resistance of 10M I'm looking for a reading of less than 35V, correct?

Sorry, no.
If the meter voltage is only 35V, the cap's leakage resistance would be almost 10 times higher than than meter resistance (meaning around 100M). To get a more realistic measuring range, you could shunt your meter with an 1M resistor, effectively making Rm = 1M. Use the formula I provided.

(BTW, "high leakage" means high leakage current and low leakage resistance.)

My method allows to determine very high leakage resistance under high voltage. It is not meant for re-forming.
If an ecap hasn't been used (seen voltage) for longer periods, re-forming as described by vintagekiki may help to bring leakage down. But if the suspect cap is out of an amp that gets some use, re-forming makes no sense as the cap would already have been re-formed by its operating voltage in the amp.

[bobloblaws]

Sorry, no.
If the meter voltage is only 35V, the cap's leakage resistance would be almost 10 times higher than than meter resistance (meaning around 100M). To get a more realistic measuring range, you could shunt your meter with an 1M resistor, effectively making Rm = 1M. Use the formula I provided.

Oops, I found my mistake. I flipped the R1 and R2 values in the divider formula. The result should have been 315V. (10M/(1M +10M) * 350)

[bobloblaws]

So I started testing caps for leakage, beginning with the two 250uF cathode bypass caps. I had already identified one of them as bad as it gave me no capacitance reading at all. I've already installed brand new replacements into the amp. I figured might as well do both while I'm at it even if one of them still has some life in it.

Now, you mentioned something earlier about it taking a long time. Is the idea that the meter should start at or near the source voltage and then count down to calculated voltage? So in the example we used where we had a 350V source and a nominal reading of 175V the meter should start at or near 350V and slowly count down to somewhere around 175V but not lower?

My source is in fact 360V and if I'm doing the calculation correctly correct I'm looking for a threshold of 324V. I started with the "bad" cap and the meter showed around 280V right off the bat. I didn't leave it connected for any length of time to see if it would change. Then I tried the "good" (better?) cap and when I connected it the meter it read roughly 360V and slowly starting going down. It settled at around 300V. So if I'm doing this correctly I believe it shows this cap has an unacceptable amount of leakage. But where does the nominal 1M of leakage resistance come from? Is that an industry standard?

[Helmholtz]

So I started testing caps for leakage, beginning with the two 250uF cathode bypass caps. I had already identified one of them as bad as it gave me no capacitance reading at all. I've already installed brand new replacements into the amp. I figured might as well do both while I'm at it even if one of them still has some life in it.

Now, you mentioned something earlier about it taking a long time. Is the idea that the meter should start at or near the source voltage and then count down to calculated voltage? So in the example we used where we had a 350V source and a nominal reading of 175V the meter should start at or near 350V and slowly count down to somewhere around 175V but not lower?

My source is in fact 360V and if I'm doing the calculation correctly correct I'm looking for a threshold of 324V. I started with the "bad" cap and the meter showed around 280V right off the bat. I didn't leave it connected for any length of time to see if it would change. Then I tried the "good" (better?) cap and when I connected it the meter it read roughly 360V and slowly starting going down. It settled at around 300V. So if I'm doing this correctly I believe it shows this cap has an unacceptable amount of leakage. But where does the nominal 1M of leakage resistance come from? Is that an industry standard?

Caution: You must never connect a cap to a voltage above its rated voltage. The cathode cap might be only rated for 25V or so. Higher than rated voltage will destroy the cap, can even lead to explosion.

[bobloblaws]

Caution: You must never connect a cap to a voltage above its rated voltage. The cathode cap might be only rated for 25V or so. Higher than rated voltage will destroy the cap, can even lead to explosion.

Hmmm, OK. The caps in question are rated at 100V, but point taken.

I'm left a bit confused though. Should my source voltage for the test be 100V then?
Edit: I realized intended operating range for the cap in question means the rated voltage and not the actual voltage so I changed the above number from 10V to 100V.


Maybe I should stick to the DCR method. :-)

Edit: obviously I took your 350V example too literally. Looking back at it now I see you said "e.g. 350V". My bad.

[vintagekiki]

DIY Simple ESR Meter

https://en.wikipedia.org/wiki/ESR_meter

https://www.homemade-circuits.com/esr-meter-circuit/
Simple ESR Meter Circuit
... simple ESR meter circuit which can be used for identifying bad capacitors in an electronic circuit without removing them ...

http://danyk.cz/esr_en.html
ESR meter - Electrolytic capacitor tester

https://ludens.cl/Electron/esr/esr.html
An Equivalent Series Resistance Meter

https://www.elettroamici.org/en/misuratore-esr/
ESR meter

[Helmholtz]

Hmmm, OK. The caps in question are rated at 100V, but point taken.

I'm left a bit confused though. Should my source voltage for the test be 100V then?
Edit: I realized intended operating range for the cap in question means the rated voltage and not the actual voltage so I changed the above number from 10V to 100V.


Maybe I should stick to the DCR method. :-)

Edit: obviously I took your 350V example too literally. Looking back at it now I see you said "e.g. 350V". My bad.

With low voltage applications you could use a 9V battery or the DCR method. Or try both and compare results.
But high voltage ecaps have been reported to sometimes show increased leakage (only) at elevated voltages.

You should wait until the meter voltage has settled (voltage doesn't drop anymore).

The 1M threshold I proposed is a value that just shouldn't disturb amp voltages. Depending on function/position in the amp a somewhat lower value may be acceptable.

[bobloblaws]

The 1M threshold I proposed is a value that just shouldn't disturb amp voltages. Depending on function/position in the amp a somewhat lower value may be acceptable.

Arghhh, this is a tough one.

I get 380VDC when I disconnect the rectifier from the reservoir cap. When connected B+ drops to 330V. I'm looking for about 360V. Rectifier tube tests good on my tester and I also tried a different tube for good measure. Running without the rectifier diodes for now.

Plate voltage on the EL84's is just a couple volts less than B+, what I would expect. Cathode resistors test at 121 ohms and bypass caps are brand new. Still getting high voltage readings on the cathodes (16V / 121R = 132mA ). Same result running all four tubes or just the inside pair or just the outside pair. I have tried different sets of tubes with same result, and all tubes verified with my tester.

I dropped the input voltage on my variac to where the B+ was 300V instead of 330V while monitoring cathode voltage. Cathode voltage went from 16V to 14V. This suggests to me the problem is on the tube end as opposed to the rectifier end.

Thoughts on logical way to proceed from here?

[bobloblaws]

Arghhh, this is a tough one.

I get 380VDC when I disconnect the rectifier from the reservoir cap. When connected B+ drops to 330V. I'm looking for about 360V. Rectifier tube tests good on my tester and I also tried a different tube for good measure. Running without the rectifier diodes for now.

Plate voltage on the EL84's is just a couple volts less than B+, what I would expect. Cathode resistors test at 121 ohms and bypass caps are brand new. Still getting high voltage readings on the cathodes (16V / 121R = 132mA ). Same result running all four tubes or just the inside pair or just the outside pair. I have tried different sets of tubes with same result, and all tubes verified with my tester.

I dropped the input voltage on my variac to where the B+ was 300V instead of 330V while monitoring cathode voltage. Cathode voltage went from 16V to 14V. This suggests to me the problem is on the tube end as opposed to the rectifier end.

Thoughts on logical way to proceed from here?

I forgot to mention, I also replaced the standby switch.

[Helmholtz]

Arghhh, this is a tough one.

I get 380VDC when I disconnect the rectifier from the reservoir cap. When connected B+ drops to 330V. I'm looking for about 360V. Rectifier tube tests good on my tester and I also tried a different tube for good measure. Running without the rectifier diodes for now.

Plate voltage on the EL84's is just a couple volts less than B+, what I would expect. Cathode resistors test at 121 ohms and bypass caps are brand new. Still getting high voltage readings on the cathodes (16V / 121R = 132mA ). Same result running all four tubes or just the inside pair or just the outside pair. I have tried different sets of tubes with same result, and all tubes verified with my tester.

I dropped the input voltage on my variac to where the B+ was 300V instead of 330V while monitoring cathode voltage. Cathode voltage went from 16V to 14V. This suggests to me the problem is on the tube end as opposed to the rectifier end.

Thoughts on logical way to proceed from here?

Well, I answered your question about testing ecaps for leakage. Doesn't mean I actually suspected your supply caps.
I already told you that an idle current of almost 70mA per tube is way too much and will pull B+ down. Do both EL84 pairs have identical cathode voltages (I already asked)?
What are the EL84 grid voltages (pin 2)?

[bobloblaws]

Well, I answered your question about testing ecaps for leakage. Doesn't mean I actually suspected your supply caps.

I'm not sure what your point is. Anyway, everything I just wrote is just my summary of what I am presently observing.

Do both EL84 pairs have identical cathode voltages (I already asked)?

Yes they do. In full power mode B+ is 312 and both cathode resistors have 13.8V. In half power mode B+ goes to 330V and cathodes to 16V.

Edit: to be precise I should have said "In half power mode B+ goes to 330V and the cathode resistor to 16V".