Can Current Mirrors Drive Multiple LEDs Efficiently?

[Jet Black]

Hi all,

I am trying to build a circuit for driving multiple LEDs under test at a given current, and the LEDs are to be individually tested with multimeter. Now I know I could just run them all in series and they would get the same current, but if one LED breaks, then the whole circuit goes down. I was thinking of using a current mirror, but I have not really worked with them before and am not sure if I can chain them together i.e. a single reference current to drive them all, with a common voltage source driving the rest, and a common ground to all of them, with the LEDs on the 'copy' arm of the current source. the component I am thinking of using is this:

http://docs-europe.electrocomponents.com/webdocs/0c5c/0900766b80c5c6ff.pdf

if it would help for me to draw a diagram, I can do that!

thanks!

 

[HHOboy]

Are you using the same LED's. If so then couldn't you put them all in parallel? They would all get the same voltage and thus draw the same amount of current then you could put your multimeter in series with the parallel array and divide the current by however many LED's you have and you have the current through each LED. You could even do it with different LED's but there would be some additional calculations that may have to be involved nothing a good spreadsheet can't solve.

I guess I don't exactly understand what you are trying to do so my suggestion may be completely useless. I don't have any experience with current mirrors.

If you want another suggestion would be to use a npn transistor for each led then give each base the same low voltage reference signal. You would also need the same large resistance for each base. This takes advantage of the amplification factor of the transistor (Hfe). For example the Hfe for a PN2222A is 100 soooo if you a reference signal of 3 volts for each transistor and a 1 MΩ resistor you should theoretically end up with a 30mA which is in the ballpark for LED's. Also it doesn't matter the voltage you give each LED.

Hopefully this was some help even though I couldn't help you with the current mirror.

 

[Baluncore]

Current mirrors are great when properly engineered, but I would avoid them when driving LEDs as the currents and differential collector voltages are high enough to give differential temperature errors. Using emitter resistors is the only way to linearise a mirror in that situation. The mirror packages you link to, when used in parallel, do no more than behave as an expensive single transistor.

So reduce the requirement to discrete NPN transistors. As a suggestion, use a +5V / -5V power supply. Connect all the NPN bases to the zero volt rail. Then an accurate resistor to set the current, say 1k → 5.5ma, from their emitters to the -5V rail. Now put your LEDs between the collectors and the +5V rail. Any LED from 0V to 5.5Vfwd can be driven since operation of the NPN in saturation is possible.

Variation of LED current should be less than about 2%. A shorted LED will not kill the system. An open circuit LED will require the transistor base provide the entire emitter current, so pick a transistor that can handle that base current.