I am using an LM324 and three TIP122 transistors to drive three 40V 40W led's. The problem I am having is the LED's don't shut of all the way (a very dim glow remains). If I use only one op amp (out of the four on the chip) and one transistor, the LED turns off properly. If I use three of the four op amps, the LED's do not turn off.

I have tried placing a resistor (20K) in line with the negative feedback to see if I was having a current bias problem, but that didn't help. I also substituted power MOSFETs for the TIP122 without luck. I also tried buffering the input using one of the op amps as a voltage follower. This actually made the problem worse. It seems there is a biasing problem, but I don't know where or how to test this.

One thought was my issue comes from the fact that the LM324 isn't rail to rail, but I am not sure about that since I can get it working with one op amp. If the input voltage is zero, I am seeing about 1.5V on the output of all three op amps.

Is there leakage between op amps on an LM324? Why would one LED work but not all three?

Any help would be greatly appreciated. Thanks.

I can't post an image of the circuit because I don't have a "reputation". However, you can find it here.

EDIT: Schematic

  • 1
    \$\begingroup\$ You need to link to the image, not another post. I can't see the image there anyway. BTW: I've run into this problem when my pull-up resistor values were too close to resistor value between my logic pin and the base of the TIP122. I ended up with ~1V on the base of the TIP122 and some current was allowed to flow. \$\endgroup\$ – Hair_of_the_Dog Aug 26 '13 at 22:21
  • \$\begingroup\$ I updated the circuit diagram link. \$\endgroup\$ – c f Aug 26 '13 at 23:22
  • \$\begingroup\$ Why the caps in parallel with the LEDs? I didn't check the transistor specs, but they may not like the inrush current. \$\endgroup\$ – jippie Aug 27 '13 at 6:39
  • \$\begingroup\$ The caps are a mistake. They are suppose to be bypass caps to help with the long leads from LED to driver. Thanks for pointing that out. \$\endgroup\$ – c f Aug 27 '13 at 15:12

So you're seeing about 1.5v on the output, but the LM324N datasheet says that the output can drive down to at least 20 mV (into a 10K load). It also seems unlikely that something is pulling the output high(er). This would imply that the problem is on the input side of things.

The datasheets says that the inputs can go to 0V, but I suspect that this is not entirely true. It might be that they only go to "almost" 0v. Even a couple of mV higher than 0v might be enough to mess up your circuit.

I suggest changing your circuit. It's a rather major change, unfortunately. The easiest (from a testing standpoint) would be to switch to a different opamp that can run off of a +/- power rail. If the opamp has a negative rail then it will be fine if the input is close to 0v (and not V-). You might be able to run the existing opamp that way, but the datasheets are ambiguous about that (and I'm too tired to look into it closely).

A different change, which would be much more complex, unfortunately, would be to change the circuit so that the input to the opamp never gets close to 0v. There are several ways to do this, but none of them are easy. One way is to put the current sense resistors on the high side, between the power rail and your LED's. Then use a resistor-divider to bring that voltage down within the range of the opamp. This might require changing some polarities and such (swapping the + and - inputs), but don't take my word for that-- actually figure it out first.

Another possibility where this circuit is going wrong is that it is actually unstable. You might need a cap between the opamp output and the negative input, and a resistor between the transistors and the negative input. I usually figure that stuff out in a simulator, but if you wanted to guess at some values I would go with 10K resistor and a 100 pF cap.

  • \$\begingroup\$ I will have to think about the high side current source. I am using a micro controller as the input so I really like the low side better. I also didn't look at this on my scope which is a bit dumb. I'll look to see if there is any instability. \$\endgroup\$ – c f Aug 27 '13 at 4:51
  • \$\begingroup\$ @cf It should be possible to have low-side drive (Transistors on the low side), but high-side current sense. You could also use low-side-current-sense with some sort of level shift (an upside down voltage divider to V+, or another opamp). \$\endgroup\$ – user3624 Aug 27 '13 at 5:24
  • \$\begingroup\$ @DavidKessner Isn't it possible to drive the base of each TIP122 from X3-3 directly? In other words can you leave off the opamps? \$\endgroup\$ – Hair_of_the_Dog Aug 27 '13 at 6:37
  • \$\begingroup\$ With the darlington transistors direct driving may be an option, but your input voltage will need to be very accurate as you no longer have the opamp control loop to control the current. \$\endgroup\$ – jippie Aug 27 '13 at 6:47
  • \$\begingroup\$ @jippie My thought was that if X3-3 is a uC output pin and a pull-down resistor, with what I assume is PWM, you would have pretty good control of the bias voltage, no? \$\endgroup\$ – Hair_of_the_Dog Aug 27 '13 at 17:26

Not the answer you're looking for? Browse other questions tagged or ask your own question.