Just out of curiosity, I would be interested in simple balancing circuits for 1000 V.
They do, but besides, the leakage depends a lot on how long the capacitor has been stored with no voltage across it, and how long the working voltage has been applied (usually two minutes for the datasheet values).
Electrolytic capacitors were not initially developed as such: at the time (in the twenties I think), they were used as electrolytic surge arrestors: they were metal blades plunged into a saline solution and conducted strongly when a certain voltage was reached.
It didn't work that well, and someone noticed that a significant amount of charge was retained after use, and decided to change the liquid transil into a capacitor.
Ecaps have largely evolved since then, but they retain some of their initial properties: above a certain voltage, they conduct, hydrolyse water converting it into oxygen which combines with the Al foil to thicken the insulating layer, and hydrogen which escapes through the rubber seal thanks to the small size of the H2 molecule.
Thus, in principle Ecaps in series should balance themselves, but Ecaps have very little margins in order to minimize their volume, and consuming the electrolyte eventually comes at a cost.
Well balanced strings should pose no problem, but if one of the members is at +80% of the tolerance and another one at -20%, the smaller cap will be subjected to heavy stress each time the device is powered.
The weaker cap will eventually weaken further and fail, which is why matching and balancing is strongly recommended
It didn't work that well, and someone noticed that a significant amount of charge was retained after use, and decided to change the liquid transil into a capacitor.
Ecaps have largely evolved since then, but they retain some of their initial properties: above a certain voltage, they conduct, hydrolyse water converting it into oxygen which combines with the Al foil to thicken the insulating layer, and hydrogen which escapes through the rubber seal thanks to the small size of the H2 molecule.
Thus, in principle Ecaps in series should balance themselves, but Ecaps have very little margins in order to minimize their volume, and consuming the electrolyte eventually comes at a cost.
Well balanced strings should pose no problem, but if one of the members is at +80% of the tolerance and another one at -20%, the smaller cap will be subjected to heavy stress each time the device is powered.
The weaker cap will eventually weaken further and fail, which is why matching and balancing is strongly recommended
Leakage current is a function of temperature. Chemicon says the leakage at the capacitor's rated max temperature is 5 to 10 times higher than at 20C. Leakage specs are at 20C. See pages 4 and 8 in the attached file.
No, but the leakage current of one capacitor can be the charging current of the other.
Suppose one capacitor has 600 V across it and the other 400 V. If the capacitor with 600 V across it then has more leakage current than the one with 400 V across it, the difference between the leakage currents will discharge the capacitor with 600 V across it and charge the one with 400 V across it until the leakage currents have become equal. It's the same process as when you have bleeder resistors in parallel with the capacitors, except that the capacitors themselves act as the capacitors as well as the bleeders.
It doesn't sound unreasonable to me, but like I wrote, I never dared to try it. There is nothing guaranteed regarding reproducibility and voltage-dependence of the leakage currents, only that they are smaller than some datasheet value after so many minutes at their nominal working voltage.
Suppose one capacitor has 600 V across it and the other 400 V. If the capacitor with 600 V across it then has more leakage current than the one with 400 V across it, the difference between the leakage currents will discharge the capacitor with 600 V across it and charge the one with 400 V across it until the leakage currents have become equal. It's the same process as when you have bleeder resistors in parallel with the capacitors, except that the capacitors themselves act as the capacitors as well as the bleeders.
It doesn't sound unreasonable to me, but like I wrote, I never dared to try it. There is nothing guaranteed regarding reproducibility and voltage-dependence of the leakage currents, only that they are smaller than some datasheet value after so many minutes at their nominal working voltage.
Cvanac. did you need a big high voltage cap if so I may have something for you.
I have an old pair of Bosch MP 40uf 2.5K caps They measure a full 4" x 6" and
have a 1980 date code. If you could use them I would make a sweet deal on them
and even ship them to you fully charged ! ( the part about sending them charged
was a joke )
I have an old pair of Bosch MP 40uf 2.5K caps They measure a full 4" x 6" and
have a 1980 date code. If you could use them I would make a sweet deal on them
and even ship them to you fully charged ! ( the part about sending them charged
was a joke )
The below circuit is tried and true. It's essentially a stacked class B 'amplifier'. Clearly not as simple as 4 resistors, but it dissipates less power and controls any imbalance more effectively.
But with a true class B amplifier doing the “balancing”, won’t the “capacitor’s“ impedance become nonlinear with large AC signal? I’d hate to be adding crossover distortion where it would not otherwise be occurring….
This circuit draws mere 10s of milliamps to do it's job and is not able to modulate the bus voltage at any appreciable level, so no effect on the audio at all. I mention that it is a class B circuit to help anyone familiar with circuits in understanding its behavior, but it cannot do more than trim the cap leakage currents - that's all that is needed of it after all.