Serial capacitor bank / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Serial capacitor bank

Move Thread

LAN_403

Fusion

Wed Mar 21 2007, 08:09AM

Registered Member #354 Joined: Thu Mar 30 2006, 10:12AM
Location:
Posts: 55

I bought 400 electrolitics capacitors (400V) at ebay: overall is 10 kjoule

Then I finished soldering first capacitor bank with lots capacitors in parallel and then in series.
I loaded extreme to extreme with low voltage: 10 Volts, that corresponds to 1.25V to every capacitor bar. But I found some of them loaded to 2.5V and others to 0.5V.
It is due different leakage currents?
I measured voltage drop from 10.0 to 9..0 in 90 seconds.

Initial capacitance was about 10% less than specifications.
I measured capacitance before and after the capacitance of every capacitor bar (fully discharged) and they recovered 2.5% of capacitance.

I will load them with 330V every bar for 5 minutes (to recover oxide coating) and see if I recover more capacitance, then test loading extreme to extreme bars with 330V and see what happens.
I would use 33kohms in parallel with every capacitor bar in order to compensate leakage currents, but I do not know if it is enought.

Sulaiman

Wed Mar 21 2007, 10:14AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

I have a few recommendations for such a situation;

Across each parallel row of capacitors put-

1) 22 kOhm 10W resistor in series with an led.
This will help balance the voltage across each row and indicate the state of charge,
and slowly discharge the capacitors once charging current is removed, allowing safer working.

2) A diode with cathode/band to +. Get the highest current rating you can find/afford (eBay?)
This is to prevent reverse-volting the capacitors, especially when discharging into an inductive circuit.

3) Optionally add TVS or Zener diodes to limit the maximum voltage per row.

As I'm sure you're aware, the capacitor bank will be scarily lethal
so concentrate on how you will monitor and discharge the capacitor bank BEFORE you consider about how to charge it.

When you reform the oxide layer the capacitance value will if anything decrease slightly, not increase.

Fusion

Wed Mar 21 2007, 02:00PM

Registered Member #354 Joined: Thu Mar 30 2006, 10:12AM
Location:
Posts: 55

Thanks Sulaiman

I have made a cocrof-walton to charge the bank. Did anibody connected ground to "-" end?

When finished I planned close the capacitor bank in order to avoid access to inner parts.
I will not use led serial with resistors due safety: if one led is broken, a capacitor would explode.

My idea is set 3 rows of 15x100 ohms in parallel so if one is broken, it will exist another 2 rows. I used 2W resistors that widstand 15x600=9000V.

It exist an alarm that advises when capacitor bank is not in full discharge mode.
Also a voltimeter measures the capacitor banc voltage.

I spent half of the electronic cost into safety operation. Is like building a commercial equipment instead only an experimental setup.

It is very interesting building and testing the great capacitor bank.

THANKS VERY MUCH Sulaiman, I did not realised that 22kohm must be 10W until reading your post.
I charged yesterday the bank to 1600V but had 2 explosions: one due using wrong limiting resistor that now does not exist and other due unwanted shorcircuit after falling a banana between two capacitors rows.
Now I use a 100W lamp to charge and discharge in a safe way during tests.

After that I loaded with 330V extreme to extreme and I had only 10% voltage difference between rows. My only problem now is how can I measure >2mF in order have same capacitance in every capacitor row.

[Edit: Fixed double post]

Energy

Fri Jun 29 2007, 04:51PM

Registered Member #868 Joined: Fri Jun 29 2007, 04:48PM
Location:
Posts: 2

I was wondering do you or do you expect to have any problems with putting 2 banks in serie? I wonder this because both banks are never identical, this could give you the problem that the stronger bank wil damage or even destroy te weaker one.

Sulaiman

Fri Jun 29 2007, 05:43PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

I have a bank of 12 series x 8 parallel photoflash electrolytics, 80 uF, 4kV total (quite small but still scary)
Once the electrolytics have been charged a few times the leakage current drops dramatically,

With a batch of capacitors, once you have several in series/parallel the variations sort of balance out naturally,
to compare capacitor rows/columns a capacitance meter isn't much use,
charge up to a set voltage, discharge via a resistor, measure time to 1/2 voltage or some other fixed level,
comparing times compares capacitance V=Vo (!-exp(-t/RC)) etc. i.e. C is proportional to discharge time.

I suggested a 10W resistor where in theory a 5W is required SQRT(300x20k)
because using resistors at full rated dissipation means running them pretty hot, which I don't like.
I almost always use twice the wattage rating actually required for resistors in my hobby designs;
much safer (one less thing to worry about) and relative to a complete project cost it's not significant.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob