Capacitor Charge time / power

[mobbarley]

Hi all!

Im a novice electronics enthusiast.

Im charging a capacitor bank from a 350V supply. The total capacitance is 940uF.

I want to be able to approximately calculate how long this will take with different resistors, then juggle the time/resistance to reach a reasonable power figure (disapation) for the resistor.

If someone can please explain how this is done so that I can substitute differrent capacitor/resistor values.

Thanks,
John.

 

[VEBill]

"...novice....capacitor bank....350V..."

Yikes! This is not the best way to start out in electronics.

Google 'RC Time Constants'.

 

[BobM3]

The time constant (time to reach about 70% of the end value) is: t = RC.

 

[mobbarley]

Okay, not that much of a novice. I am building a strobe but this is the first of my own design and I'd rather not cook resistors & caps / my variac in trial and error

 

[mobbarley]

Thanks Bob, I have also found a calculator using VE1BLL's suggestion.

Is there an easy way to approximate a series resistor's dissapation? P=IV will only work when the capacitor is fully discharged & the AC is at full wave (rectified mains).

 

[Skogsgurra]

The sad truth is that the capacitor takes half and series resistor takes half. The well known formula W=0.5xU^2xC reflects this.

So, if you are charging your 940 uF capacitor to 350 V, you will dissipate .5x.00094x350^2 which computes to near 58 Ws each time you charge it. If you charge it once every second, your dissipation will be 58 watts. If you do it ten times a second, the dissipation will be 580 W.

Simple as that.

Take care. These voltage levels are no playthings.

Gunnar Englund

http://www.gke.org

 

[itsmoked]

Quite the energy waste there..

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[logbook]

I like the analysis Skogs did. It cuts right through to the heart of the problem. The "time constant" is something else though. By the sound of it you are going to half wave rectify the mains into the capacitor bank, ie use one diode and a resistor as the charging circuit.

Rate the diode for the peak current, that is the peak mains voltage divided by the resistance.

If you want to charge the bank in a few mains cycles there would be no simple formula. If you charge it over a period of tens of seconds then I think the mains pulsations will all tend to even out and the time constant formula will become applicable again, albeit with a scaling factor. Just guessing at this stage. I would run a SPICE simulation because there is less chance of making a large calculation error.

Tell us what time constant you want and whether or not you are doing half wave rectification of the mains and then we can offer more specific answers.

 

[VEBill]

"...going to half wave rectify the mains..." [to make 350 volts].

Depends on his local line voltage. In North America, with 115 Vac input, he wouldn't reach 350 Vdc with a half-wave.

There's lots of AC-powered strobes around. Must be plenty of schematics available on the WWW. Of course, a found circuit plan can be easily adapted and extended.

 

[logbook]

... but 240*sqrt(2)= 339V so I figured he might be UK or similar.

 

[waross]

I believe that the time constant is closer to 63%. The precise number is not a guess, or something determined by experiment but is derived from "N", the base of of the system of natural logarithms. (1/(N-1)??
respectfully

 

[IRstuff]

EE001: RC circuits charge according to 1-e^-t/RC

So one time constant is 1-1/e=63.21%

TTFN

 

[mobbarley]

Australia. 240V.

Thanks for everyones help.

 

[logbook]

Yes IRstuff, but this equation assumes a steady DC charging voltage. If we are now charging through a diode from a sinuosoid you have to include some sort of correction term to allow for that. It can't be as simple as taking the half-cycle mean voltage because the capacitor will only charge when the sinusoid is above the capacitor voltage. A closed form solution is probably not possible. That's why I suggested using a SPICE simulation.

 

[IRstuff]

No problem, but the "RC time constant" that was bandied about is a constant, defined by the DC charging condition. Whatever the OP is doing, you would apply some other nomenclature to that.

TTFN

 

[logbook]

http://www.logbook.freeserve.co.uk/image/mains.gif

The RC time-constant is 4x what you might expect, and even then the curve isn't exactly an exponential. However it is pretty close.

http://www.logbook.freeserve.co.uk/image/mainsgr.gif

 

[mobbarley]

Thanks Logbook!