Electronic – Regulator/heatsink heating up even though load is only 50% of regulator Imax

heatsinkpower supplyvoltage-regulator

aI have made a small +/- 24VDC liner supply for the use of powering an audio preamp PCBA. LM317 for plus/ LM337 for the neg (both TO-220), 10uf output capacitance added. Even though this circuit doesn't draw more than about 700mA (inly 50% of what these regulators are spec'd to), my heatsinks are heating up pretty seriously after about a half hour. The output of the regulator looks good, but i am trying to manage my heat dissipation issue. The heatsink's I've chosen are larger than what you might expect (google "SK 104 63,5 STS" you'll see the ficsher electronik heatsink I've selected) transformer secondary voltage is 28vac converted after the bridge to 28 DC via a 4700uf electrolytic. I have only two questions:

  1. Both of my heatsinks are mounted to perf board, and not touching one another, not touching any other component nor ground. If i have not efficiently isolated the output to the heatsink and due to the 317/architecture the base plate is allowing the voltage to flow to the heatsink itself, even though it's really not going anywhere – could the energy just being in the heatsink cause the regulator to work harder, thus generating that heat?

  2. If I add two to-220 package Darlington's to increase the output load capabilities of both regulators, i understand i would have to heat sink them as well, but will that decrease the overall heat dissipation already being generated without them (i suspect not, i suspect it will just radiate out of the heatisnks for the darlingtons)?

Best Answer

The key things to figure out are:

  • Regulator current
  • Voltage drop

The first you've stated as 700mA.

The second we'll figure out. 24V AC will regulate to 24V/0.707 = 34V DC. With 24V output, we have a drop of as much as 10V (34V in - 24V out.)

10V at 700mA will dissipate... 7W in your part, worst-case. That's kind of a lot for the LM317 / 337 in TO-220.