Charging Circuit

[mechdesign2k4]

Please forgive my ignorance in the relm of electrical engineering as I am a mechanical designer. I would like help with a project that I am working on. The description is as follows. I have a flash bulb, that I would like to flash with 1100 joules of power. The charge bank consists of 3 capacitors 1200u, 500V, and 3 caps 1600u, 500v. I would like to charge these capacitors after flashing in under 10 seconds. Can anyone help me with this circuit?

Sincerely,
Mark

 

[MacGyverS2000]

Specifically what part do you need help with? It sounds like you already have a circuit, so why not attach the power source designed for it and move on to a more important part of the project?

Dan - Owner

http://www.Hi-TecDesigns.com

 

[mechdesign2k4]

Thank you for your reply. I need help with the design of the charging circuit. I did not include the power available, it is 12v 18 amps.

 

[treez]

first...what voltage do you want to charge the caps up to?

So...1100 joules of energy to return to the caps in 10 seconds.

Q = C V......
Q = 8.4mF * V

Cap energy is 0.5 * C * V^2.

Sounds like you want a controlled boost converter to pour charge back into the caps.

 

[MacGyverS2000]

12V/18A is not a simplistic design to work on, particularly if you are not electrically inclined. I would suggest asking the company's EE dept. to take a look at this one. That kind of power can melt metal if not handled correctly.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[Comcokid]

There are IC's out there made to charge a photo-flash capacitor circuit. However, it sounds like your flash may be larger given the value of your capacitors you indicated.

A quick search (Digikey) uncovered the Linear Technology LT3420, and the data sheet even has examples of paralleling multiple chips to charge larger capacitors quicker - but still not as large as you indicate. (Digikey also has a Allegro Micorsystems A8439 as well, but given it's small surface-mount package, it might be hard for you to use).

Check out Linear Tech, Micrel, Maxim, National Semi, Fairchild Semi, and TI. Look for 'power management' ICs, boost-type switcher, PWM chips, and/or for photo flash chips, application notes, or circuit ideas. If you get lucky, you might find one of these companies has a demo board available.

 

[DHambley]

Let's see....

First of all, 1100J out of a total of 8400uF. That requires 512V. The capacitors are being used above their 500V rating.

512V in 10 seconds takes 430mA average, probably in the neigborhood of 1A peak. That's a 500W peak switching circuit....not trivial.

If someone asked me to design that, I'd give them an estimate of 80 hours for analysis and a crude breadboard, then several more weeks to fine tune and get boards layed out.

It just doesn't sound like a simple circuit that someone can "help you with" on a chat room.

Darrell Hambley PE
SENTEK Engineering

 

[mechdesign2k4]

Thanks All! I appreciate the input!

 

[Comcokid]

More application notes on DC/DC converters for high voltage:
Linear Tech Application note 118, some of which appeared in an article this month - Electronic Design News, Aug 7, 2008, "High-voltage, low noise dc/dc converters" . These were in the pile of papers on my desk when I read your post.

 

[HeyDoc]

Is this going to be a line operated device? If so, a modified Cockroft-Walton multiplier is likely to be what you require.
If you're using a regulating circuit, even a bang-bang regulator, you can safely charge the caps to their full 500 volt operating voltage. You will get up to 10% more energy during a pulse discharge such as that through a strobe lamp than the rated capacity of the caps would indicate. The discrepancy exists because the standard capacitance measurements are performed using a sine wave at 120Hz, and the discharge 'wavefront' is much faster than a 120 Hz sine wave.
A correctly designed line operated multiplier should recharge your cap bank in about 2.5 seconds, with an initial surge current of around 20 amperes lasting for the first few ac half-cycles, then tailing off to less than 2 amperes by the end of the cycle. Be sure to use zero crossing control on any thyristor you might use to control the multiplier.