strange power supply output voltage behavior

Thread Starter


Joined Jan 13, 2018
I am in the process of designing a linear power supply.
I am using circuit board relays to switch in 3 different feedback voltage dividers for fixed 5V, fixed 12V and Vadj which will be an external potentiometer.
I have not designed the current limiting or short circuit protection circuitry yet, but i will.
Here are my issues.
1. Lets say i set fixed 12V for 12.05V. It will hold the voltage at that point as long as the circuit is energized. If i turn the circuit off for 5 minutes and turn it back on the output seems to have climbed a few hundred mV and then drops down slowly back towards the set voltage. This problem only seems to happen with the 741 opamp, i was using a 358 before and it would not do it. The 741 provides my circuit better load regulation than the 358, that is why i am experimenting with the 741 as my pass transistor driver.
2. When i connect a 7 ohm test load to my power supply with 12.05V the output drops by 10mV which is not a big deal. However when i disconnect the load the output voltage will eventually start to drop slowly. Somewhere around 10mV per minute

any insight or suggestion would be awesome!



Joined Feb 8, 2018
The 741 is unsuitable for use with a single positive supply. The output cannot swing close enough to its negative supply rail. Some carefully selected FETs might work OK because of high gate threshold voltage.

There is no frequency compensation in the op amp-FET circuit. It may well be unstable.

Relay switching of the feedback is likely to lead to the feedback going open-circuit in a transient fashion, applying the full input voltage to the output. There is no reason whatever to switch the high sides of the feedback pots.

There is no reason to use multiple reference voltages.

Ordinary zeners make poor voltage references.

I recommend using an LM317 with current-boosting bipolar transistors. The circuits are published, simple and reliable.

Thread Starter


Joined Jan 13, 2018
Thanks for the reply.
The relay switching of the feedback works without issue when the circuit potentiometers have been set for my desired outputs. (5V and 12V) I can switch back and forth between them without any issues. But i suppose this is an over complicated and redundant approach to selectable output voltages on one positive binding post.

The 12V zener and potentiometer are for a comparator reference that switches my transformer windings depending on how low my ouput voltage is. It does not have anything to do with my 12V output.

I will try adding some frequency compensation capacitance on the opamp-fet circuit.
Could providing a negative rail for the 741 make it perform better with my described transient conditions? So that it could swing down to 0V

Thread Starter


Joined Jan 13, 2018
I used a 358 in the same configuration earlier in my design process. It is strange, the 741 provides significantly better load regulation in my configuration.

12.05V with 7 ohm load drops to 12.04V with 741. Non inverting input reference is set less than 1.8 volts which is how low the 741 swings when connected to ground.

12.05V with 7 ohm load drops to 11.6V with lm358.

But i have no compensation provisions for this portion of my circuit.