LM317 tracking preregulator

Thread Starter

mike_the_begginer

Joined Dec 7, 2019
102
Hello, I built the attached schematic and I would like to know what tests should I do to make sure that the psu works correctly ?
I measured the output voltage without load (25.81 V) and with load (25.64 V). The difference was 0.17 V.
The output of the psu can be seen on the screenshot - bmp_42_007.
To test the output I used a circuit with 555 which generates pulses to short the output of the psu at a set frequency.
 

Attachments

Thread Starter

mike_the_begginer

Joined Dec 7, 2019
102
Q4, 5, 6, 7 are part of the preregulator. I use a 30Vac transformer, which rectified and filtered is about 40-42Vdc and I need to reduce the voltage at the input of the LM317. Then using the transistors and the preregulation circuit will make the input voltage of the LM317 to be about a few volts above the output. Also, I know that the transistors are more than I need, probably it should be sufficient 2 transistors, but I have them already mounted on the heatsink and for tests I will use them as they are.
 

Lo_volt

Joined Apr 3, 2014
199
I think what Dodgydave means is that there are plenty of linear regulators out there, including the LM317HV which can handle 57 volts at its input. Why not find one that can handle the high input voltage and skip the pre-regulator?
 

sagor

Joined Mar 10, 2019
469
Since the 317 is limited as to current capacity, you would only need one pre-regulator transistor, with a current rating slightly better than the 317 itself. Those 4 2N3055 are a typical design for a 10-20A pass regulator circuit, not as a pre-regulator, one would never use a 20A capacity pre-regulator to feed a 1A to 1.5A 317 regulator
 

Thread Starter

mike_the_begginer

Joined Dec 7, 2019
102
I understand that there are better solutions for building psu's.
But I like this project and I think that it is a good project for learning. This is why I am asking some questions here.
 

crutschow

Joined Mar 14, 2008
27,438
I think that it is a good project for learning.
One thing to learn is that your circuit is way overkill (It could be done with one resistor one Zener, and one or two transistors), but if it works for you, (and it appears to) than that's okay.

There are always complex ways to solve a problem.
The good engineering is to do the task as simply as possible.
 
Top