help-rapid cap bank charging

[Mag Man]

I need to charge cap bank(40x100uf 500vdc electrolytic parallel) in under 50mS using
120vac 60hz house power on 30amp breaker. Experimental system- cap bank fires for 10-20mS
at a 30-50mS(optimum) duty cycle. Up to 90mS duty cycle acceptable. I currently have a working
system using 6 power transformers (1:4) and bridges to charge caps, this is inefficient to say the least, looking to upgrade charging system, any ideas or solutions would be well received. Thank you

 

[alec_t]

Welcome to ETO!
Show us the schematic of your present set-up and perhaps we can go from there.
Be aware that those caps when charged can KILL you!

 

[ronsimpson]

Maybe you should tell us what you are really doing.

Every charge time, are the capacitors starting out at zero volts or do they have some charge in them?

 

[tcmtech]

Unfortunately, a bit of basic math should have shown you that charging a 4000 uF 500 volt capacitor bank from 0 to 500 VDC in 1 cycle of a 120 volt 60 HZ AC power system with a 30 amp capability is impossible.

My rough math says your 30 amp line capacity to do so under ideal conditions is off by a factor of ~150+.

See here.

http://mustcalculate.com/electronic...?vfrom=0&vto=500&vs=500&c=4000u&r=.5&time=.05

 

[dr pepper]

Its a bit late in the day for me but if I got this correct.
To charge 4000uF to 500v you need 500 joules (assuming a start voltage = 0).
In one second that would be a 500 watt load for that second.
If you want to do this continuously 50 times a second (20mS) then that would draw from the supply 500 x 50 = 25 KiloWatts = a load of juice, domestic sockets here in the uk can provide 3 KW.
The actual power would be diffrent due to parasitics, thats just a ballpark.
I agree that this would be exceptionally dangerous, both voltage wise and having body parts nuked by plasma discharge.

 

[ronsimpson]

To charge 4000uF to 500v you need 500 joules

That is why I want to know if the cap gets fully discharged. In many of these applications the cap gets 1/2 discharged. This helps the math.

 

[schmitt trigger]

I*t = C*V
Solving for I:
I =C*V/t plugging the requested values provides 40 amp charging pulses @ 500 volt, it would require 20 kilowatt.

Impossible on a common household circuit.

 

[AnalogKid]

The total energy stored in a capacitor is 1/2 x C x V^2.
0.5 x 4000 uF x 500 x 500 = 500 Ws (watt-seconds). So if the charging time is 1 second, you need to supply 500 W. If the charging time is 30 ms, you need to supply 16.667 kW. At 120 Vac house voltage, that's 139 amps, plus conversion losses.

Note that these numbers assume the caps have an ESR of zero ohms. Real-world caps might not be able to do what you want no matter how much power is available.

ak

 

[dr pepper]

In the Uk domestic supplies tend to be 15Kw or 24Kw.

 

[spec]

Hy Mag Man,

I see from your profile that you are from the United Stares. Care to tell us which state and put it in your identity window on the left.

Not only are the above observations prescient but you would have to analyze the stress put on the capacitors and establish if they are up to the job- my feeling is that ordinary reservoir capacitors would be struggling.

spec

 

[Mag Man]

Welcome to ETO!
Show us the schematic of your present set-up and perhaps we can go from there.
Be aware that those caps when charged can KILL you!

here is the current set up

 

[Mag Man]

That is why I want to know if the cap gets fully discharged. In many of these applications the cap gets 1/2 discharged. This helps the math.

 

[Mag Man]

RON,
THE DISCHARGE IS~50VDC DURING THE FIRING TIME OF 10-20MS,
I HAVE SEEN CONVERTERS (DC/AC) 12VDC 10AMP(AUTO CIGARETTE LIGHTERS) TO 120VAC 33.3AMP. I WOULD LIKE TO USE 50VDC 20AMP TO
500VAC 30+AMP CIRCUITRY OR SOME SUCH CONVERTER, SINGLE PULSE
TESTING PROVIDES DESIRED RESULTS. DUTY CYCLE 60-70MS WILL NOT
ALLOW FULL CHARGE WITH PRESENT CURRENT PARAMETERS. THIS
IS THE OBSTACLE I SEEK TO OVERCOME, THANK YOU FOR YOUR REPLY

 

[spec]

here is the current set up

After seeing your schematic, I can't imagine a better way to achieve your objective. It could possibly be realized by using switch mode techniques but, in view of the high voltages and currents that would involve a massive design and development exercise, and the components would be expensive.

As the problem stated on the schematic is that the transformers are overheating, you could look at ways of improving the rectification efficiency, but that would require six large inductors, one between each bridge rectifier and capacitor bank or one very large inductor in series with the capacitor bank

The other approach would be to implement an efficient cooling system for the transformers. If forced air cooling was not sufficient, oil cooling may be possible but it would be important to use transformer oil compatible with your transformers. Can you post a picture of the physical set up and show one of the transformers in detail.

While your schematic is pretty clever and well drawn, it may be better to connect all six transformer secondaries in parallel and use just one high current high voltage bridge rectifier. This should ensure that the transformers current shared better.

spec

 

[kubeek]

If you have 460Vac at the output of the tranformer, then the peak voltage is 650V and your 500V caps will charge to that level, unless you somehow stop the charging process in time. This basically means a lot of boiling electrolyte and other stuff flying around.

You have available 120Vac at 30A, which si 3600W of power available to you. You ask for 500 joules delivered every 70ms, which is 7142W. Those 7kW are only if everything is 100% effective.
You need better supply of power, like a three phase circuit.

 

[alec_t]

I WOULD LIKE TO USE 50VDC 20AMP TO
500VAC 30+AMP CIRCUITRY OR SOME SUCH CONVERTER

No such converter exists (or if it does, please let me know where I can buy one ). The output power would be 15 times the input power!

 

[AnalogKid]

I HAVE SEEN CONVERTERS (DC/AC) 12VDC 10AMP(AUTO CIGARETTE LIGHTERS) TO 120VAC 33.3AMP.

No, you haven't. Run the numbers.

12 V @ 10 A is 120 W. 120 Vac at 33.3 A is 4000 W. No car on earth can supply 4 kW from the cigarette jack. A 4 kW output inverter requires 416 A @ 12 V after conversion losses, over 40 times what the car's wiring can support. Plus there's that whole alternator rating thing.

ak

 

[AnalogKid]

it may be better to connect all six transformer secondaries in parallel and use just one high current high voltage bridge rectifier.

Don't think so. You could have significant current circulating among the mismatched transformer secondaries, current that further increases transformer heating while diverting energy away from the load. The bridge diodes act as droop-share resistors. Co-locating them on a common heatsink will improve load balancing.

ak

 

[Mag Man]

After seeing your schematic, I can't imagine a better way to achieve your objective. It could possibly be realized by using switch mode techniques but, in view of the high voltages and currents that would involve a massive design and development exercise, and the components would be expensive.

As the problem stated on the schematic is that the transformers are overheating, you could look at ways of improving the rectification efficiency, but that would require six large inductors, one between each bridge rectifier and capacitor bank or one very large inductor in series with the capacitor bank

The other approach would be to implement an efficient cooling system for the transformers. If forced air cooling was not sufficient, oil cooling may be possible but it would be important to use transformer oil compatible with your transformers. Can you post a picture of the physical set up and show one of the transformers in detail.

While your schematic is pretty clever and well drawn, it may be better to connect all six transformer secondaries in parallel and use just one high current high voltage bridge rectifier. This should ensure that the transformers current shared better.

spec

 

[Mag Man]

I was considering dc/ac converter circuitry such as used for automobile cigarette
lighters, ie; 10-20amp 12vdc input = 120vac 33amp output(4000watt),adapting this
technology to achieve 500vdc(bridges) 30+amps. No market for such a thing therefore nothing manufactured