2.5 - 6 VDC regulated power supply with overcurrent indicator problem

Thread Starter


Joined Sep 6, 2015
I recently began work on a adjustable power supply using an LM350 adjustable voltage regulator. I am aware that the LM350 has short circuit protection but no visible means of indicated that it is experiencing overcurrent due to a shorted load. So, a comparator circuit was added with a red and green LED to indicate this but to also short the adjust pin of the regulator to ground which results in the voltage dropping to ~1.8VDC. When I built the circuit to test it, I shorted it to test the protection. It worked, some what but not in a way I expected. The voltage did drop however, with the short circuit in place of the load, high amps were still being drawn from the circuits power source. I am under the impression that the circuit needs to be redesigned to remove current but if I remove it, there will be no power being fed through the short circuit to feed the comparator to indicate a short. Is there anyone who can help with improvement of this design? The circuit is currently under space constraints so I'm trying to minimize adding too many components and I didn't want to build it using any 'exotic' parts.

Originally, the circuit used an 741 instead of a 321 OpAmp, but for some reason, I am having better results with the 741. A 2n5060 SCR is shown without a reset switch in the case of a trigger event for clarity. I tried using a BC547 but it didn't work for the regulator, the power settings were always stuck. :/

Is there a way to change it to a self resetting design? Whereas , it would pull power from the load, indicate a short but would return to normal operation once the load was removed or corrected. I was given an idea from someone to use a 1M ohm shunt to bypass the regulator to continue feeding the comparator but I didn't want to add a relay and use a second mosfet instead. Is this appropriate to do this or is there a better or more correct way to do this?

I have posted the current schematic below. Any help in this matter is greatly appreciated.

3-6v_adj_ps_w_short_protector14.sch - gschem.png


Joined Aug 1, 2013
First, what you have discovered is the difference between overcurrent protection and current limiting. The 350 will protect itself, but that's not the same as protecting the load. Your schematic looks like it has the right pieces; it's just a matter of configuring them for the function you want.

Generally speaking there are two kinds of current limiting, constant current (CC) and foldback (FB). Constant current is just like it sounds, if the load attempts to draw more current than the circuit allows, the circuit adjusts the output voltage such that the output current is a constant at the set amount. In a short circuit condition this results in the output voltage decreasing to its minimum adjustable value and sitting there. When the short is removed, the output pops up to its original set value.

Foldback limiting is more like an electronic circuit breaker. When overcurrent occurs, the output drops to wither its minimum adjustable value or to zero, and sits there. When the short is removed, it still sits there, latched off until it is reset somehow, usually by cycling external power to the circuit. Linear Technology has some switching supply controllers that have a hiccup mode: While latched off they periodically try to increase the output voltage and see what happens. If the short is removed they return to the set output.

There are several threads on this forum, many schematic on the innergoogle, and basic circuits on the 350 datasheet showing how to add CC limiting to the output. A drawback of the standard approach is that if you want an adjustable CC setpoint, all of the output current flows through the adjustment pot. Of course there are ways around this, and since you already have an external current measuring shunt and opamp in your design you are about 75% of the way there. So, do you really want the latch-off operation of an SCR, or one of the other operational options?
The 741 will not work in your application because its input common mode range does not extend down to the negative power supply rail (0 V in your case). At 3 A output current you have only 0.3 V at the opamp input, but the 741 needs a minimum of 2 V or so (working from memory) above its negative rail. The 321 input stage was designed specifically for this kind of application.


Thread Starter


Joined Sep 6, 2015
Well, when I first began operating the circuit, I had the notion that it would fold back the voltage to the lowest setting but found that during the event of a short circuit, the circuit would brown out the power source feeding the circuit. I was under the impression at first that the circuit would remove the current from the load and indicate that the load resistor was too much for the circuit. For instance, since the load resistor is 2.2 ohms, I had adjusted the comparator's input adjust to activate when the load reached say, 1 ohm. The SCR latches and remains latched, even after power has been removed from the circuit until the cathode and anode is shorted and it will reset. I had been searching the interwebz for some time for an alternative by providing current foldback but with self recovery after a set amount of time. Unless I'm going about this the wrong way.
I guess that I'm thinking that during the event of an overcurrent, the power is cut from the regulator but the comparator is still powered on so that it can indicate the event via the red led. During this time, current foldback occurs to minimize current to the load. Another circuit added will let the comparator run for say 10 seconds. Enough time for the indicator to be viewed then all power is pulled completely to reset the foldback portion of the circuit. The cycle starts all over again when the command button is pressed.

For the case of the 741, I've had good results running it. It works best in dual power mode but can run in a single supply mode. But the 321 is a more modern and more stable opamp and runs well on single supply.

Perhaps I wasn't search the correct set of terminologies as I'm certain that this has already been achieved in the past somehow.