Hejsan

I'm trying to make a simple variable LM317 based power supply. It works like a charm except current limiting starts at ~2.7A where after it keeps the voltage but clamps current to a steady ~150mA. I didn't actually implement any current limiting, at least not intended, so I'm surprised that it does that and I wish it wouldn't.

The simulation does not show any current limiting. All components are Farnell stock and there is a 2200uF output cap. V1 is capable of delivering 10A.



Any input would be appreciated.

 

Q1, R7, R3 are causing the problem.

Your circuit would be better using a PNP pass transistor for Q3, current limit is set by the series resistor R4, for the pass transistor, or using a Series resistor in the Ground Terminal.

 

Something like this -

Regards, Dana.

 

Q1, R7, R3 are causing the problem.

You're circuit would be better using a PNP pass transistor for Q3, current limit is set by the series resistor R4, for the pass transistor, or using a Series resistor in the Ground Terminal.

Thank you, I will try the PNP solution.

But for my education only, how exactly are Q1, R7, R3 limiting the current? They were intended to give voltage feedback to the regulator.

 

Something like this -
Regards, Dana.

Thank you, I will try that.

 

Thank you, I will try that.

Q1 is turned on as sooon as the output goes over 1V, which immeadiately pulls the LM317 regulator down to 1.25v.

 

Q1 is turned on as sooon as the output goes over 1V, which immeadiately pulls the LM317 regulator down to 1.25v.

I would believe that would would happen if base of Q1 was directly connected to emitter of Q3 (and not through the R7/R3 voltage divider as shown). And that is not happening as the voltage remains at the set level while the current drops from 2.7A to 0.15A.

 

I followed the advice given, but the mystery persists:

In BOTH below cases when the current exceeds 2.7A it drops to 0.15A while output voltage is unchanged.



Then I removed Q1, R7, R3

 

You probably need to adjust your current sense resistor (R6) to turn on the transistor sooner. (assuming the models are correct)

In general the sooner you turn on the pass transistor the higher current you can get, but with less output regulation, and the later you turn on the pass transistor the less current you can get but with better regulation.

I would try the circuit in post #3 as that is the simplest and the one I have used in all my series pass supplies. Be extra sure the pass transistor can handle the wattage, and the model is correct.

Also you may want to add an emitter resistor (small) on the pass transistor to give some thermal runaway protection.

 

From an LM317 data sheet:

"When an overload occurs the device shuts down Darlington NPN output stage or reduces the output current to
prevent device damage. The device will automatically reset from the overload. The output may be reduced or
alternate between on and off until the overload is removed."

The drop in current may be this protection behavior. It may be a type of foldback current limiting. Other data sheets may have more information on the protection behavior. Summary: normal.

 

OK, I solved the mystery, it was my own oversight.

I have a 200 Ohm resistor in parallel with a string of series connected 1 Ohm 10W resistors as 'variable' load, and one of those resistors were defect. I wrongfully interpreted the result as current limiting when in fact it was just falling back on the 200 ohm resistor and hence showing 0.15A, silly me.

I'm going with the second design in my post #8, that one has the least voltage drop about 2.5V when going 0.15-10A

Thank you all for your help!