Update: Small PSU with LM317 and LM337

Small PSU with LM317 / LM337

Small PSU with LM317 / LM337

Lately I used my small variable PSU with LM317 / LM337 as fixed PSU build in a Module. In an earlier post I wrote: “Because it is small the heatsinks are not that big and close together. Don’t drive it to hard”. This only applies when you want the output to be variable. If used for fixed voltages you can drive it harder when you choose an appropriate AC input voltage. The AC input voltage should match the DC output voltage in this case. This lowers the voltage drop at the voltage regulators LM317 / LM337. It is about 4V then. With 1A load the power dissipation is down to 4W. I have tested the small PSU for several hours loaded with 1A at 15V without any problem getting to hot. The test was done on the bench in free air. If you use an enclosure check for the air flow.

Small PSU with LM317 / LM337 as module

Small PSU with LM317 / LM337 as module

 

4 thoughts on “Update: Small PSU with LM317 and LM337

  1. Hallo Harald! this is nice, I was thinking to build something similar but with the lm350 or lm338 instead to have enough room without getting too much hot (I don’t want to use fans in my vsynth)
    So, reading this maybe I can build something like your psu, you have temp figures? how have you tested it? with an electronic load? I bought one recently and discovered how much varies some psu when put it under different loads.
    Love your project and how detailed you are documenting all your stuff. Keep going!

    IM

    • Hallo Isidoro,

      thank you for your interest and question. Using an LM350 or LM338 will not change the power dissipation for the heat sink. Power dissipation only depends on the voltage drop across the regulator and the current. And do not forget that the rectifier diodes and the load capacitors heats up as well.
      Usually i use an electronic load for testing. And i can confirm the variability of PSU. It is a lot crap out there. (Look for ripple at heavy load, reaction time, switching noise, …) Here i have used 20 high current resistors 150Ohm/5W in parallel 10 for each channel. That’s because my electronic load has only one channel and i wanted to load both channels equally.
      The temperature measured was about 65deg Celsius at the body of the LMxxx and 45deg Celsius for the rectifier diodes and load capacitors under full load (1A). You can use bigger heat sinks. That will lower the temperature but the over all power input (heat) in your system will stay the same. I’ll put the temp figures on my website asap.

      • wow! Harald, thank you for your quick reply.
        I’m following all your synth and vocoder news, please count with me if you ever want to sell the leftovers PCBs especially the vocoder ones, I never though in build a vocoder but I find your design pretty interesting and it has pretty unique features.

Comments are closed.