I'm wondering about making a flashlight with 21-49 bright white LED's powered by 4-6 NiMH batteries. I would like to make a PSU for driving the LEDs by constant current and I'm looking at the possibility of using the boost converter LM2623.
In order to increase efficiency of the circuit I was planning in connecting 3 LED's in series and paralleling 7-17 of these series. Then I hope to incorporate a "small" current sensing resistor to minimize the efficiency loss in it. Will the following circuit work? Or do you have other suggestions?
I'm not very experienced designing circuits and especially layouts (operating freq. 300kHz-2MHz) for this, and considering the physical size of the LM2623 is an 8-pin 3x3 mm IC I will probably have some real challenges....but what's a fight without a challenge?
space
In order to increase efficiency of the circuit I was planning in connecting 3 LED's in series and paralleling 7-17 of these series. Then I hope to incorporate a "small" current sensing resistor to minimize the efficiency loss in it. Will the following circuit work? Or do you have other suggestions?
An externally hosted image should be here but it was not working when we last tested it.
I'm not very experienced designing circuits and especially layouts (operating freq. 300kHz-2MHz) for this, and considering the physical size of the LM2623 is an 8-pin 3x3 mm IC I will probably have some real challenges....but what's a fight without a challenge?
space
You need some resistors in series with LEDs, some ten ohms to allow for tolerances and temps of every led chain or the one with the smallest voltage drop will eat itself (the more current the led takes, the more it heats up, the less the voltage drop, the more current will flow, puff)
Thx for your reply VEC7OR. I'm aware of the problem you are describing but it is not my main concern atm.
I've done an experiment driving 10*3 LEDs without series resistors from a 300 mA (=100% load) CCS for 24h+ continues to check for the problem you describe. No problems observed. (Drive voltage dropped from 9.75 to 9.30). The reason none (which ofc would have meant all) did die in this relatively short time span I expect is due to the fact that these LEDs don't have very sharp turn-on characteristic.
Though in the finished design I will be matching LEDs for more equal current sharing. I will also be running them at lower than 100% allowing for some uneven current sharing over a much longer timespan.
space
I've done an experiment driving 10*3 LEDs without series resistors from a 300 mA (=100% load) CCS for 24h+ continues to check for the problem you describe. No problems observed. (Drive voltage dropped from 9.75 to 9.30). The reason none (which ofc would have meant all) did die in this relatively short time span I expect is due to the fact that these LEDs don't have very sharp turn-on characteristic.
Though in the finished design I will be matching LEDs for more equal current sharing. I will also be running them at lower than 100% allowing for some uneven current sharing over a much longer timespan.
space
- Status
- Not open for further replies.