Constant current source can't light an LED

[nathanpc]

I've designed a very simple constant current source (just to learn more about them). While I was testing it on a breadboard by shorting it using a multimeter to measure the current everything was fine, so I designed a PCB and sent it out for manufacture. The PCB came back and I populated it.

When I tested it again with my multimeter everything was fine, but when I decided to test it with a simple red LED (which I didn't remember to test even when I had it on a breadboard) it didn't light up (with the current set to 20mA). So I decided to measure the current through it: 0uA and the voltage 1.245V.

I'm really intrigued by why the LED isn't lighting up and why there's no current through it. Could anyone explain to me why it doesn't work?

Here's my schematic, you can also check it in my GitHub repo:

[Psi]

What battery voltage are you running the circuit from?
VCC and V-

[w2aew]

What battery voltage are you running the circuit from?
VCC and V-

Yes, it is most likely a voltage headroom problem.  The LM358 can only swing to within about 1.5 or 2V from the positive supply.  Then, you're loosing another 1.3V or more across the Vbe junctions of the transistors.  If you were running on a 5V supply, this only about 2.7V available to put across the LED (at low current levels), and even less at high current levels where the drop across the sense resistor is larger.  This is probably not enough to turn the LED on.

[nathanpc]

Thanks very much, that was exactly it! turned the voltage up and it now works perfectly.

[nathanpc]

Thanks very much for the tip. I'll definitely test (and choose the best values of R and C) and include the filter in the next revision of the board.