I am using LM350 for regulating 12V for 1.7A load. As per datasheet it can drive max 3A load.

But in my case, it's stops regulating after keeping this circuit for burn test for 3-4 days of testing. During burn test,randomly it stops regulating even though current is well below 3A.

I have stepped down output from bridge rectifier by using resistor (0.5Ohm,2W) and voltage I have marked below.

Any suggestion without much modification in existing hardware.

Thanks in advance.

LM350 Heat sink

  • \$\begingroup\$ Can you add a picture of your circuit, showing the regulator and other components? \$\endgroup\$
    – jippie
    Oct 1 '15 at 8:50
  • \$\begingroup\$ A quick calculation: For the LM350: 19.6 V in - 12 V out = 7.6 V. The current is 1.7 A so the LM350 needs to dissipate 1.7 A * 7.6 V = 12.9 w so you need a decent size heatsink like 6 x 6 cm. Does the LM350 have a heatsink ? If not it will get far too hot and it will then limit the voltage and current to protect itself. I know the datasheet says 3 A but that is only if conditions allow that (meaning sufficient cooling is in place). \$\endgroup\$ Oct 1 '15 at 8:51
  • \$\begingroup\$ How big is the heatsink? \$\endgroup\$
    – Andy aka
    Oct 1 '15 at 8:52
  • \$\begingroup\$ Dear all, 9*6 cm size heat sink ( refer image attached ) I think it's sufficient. Right? \$\endgroup\$ Oct 1 '15 at 9:05
  • 2
    \$\begingroup\$ "Quick" experiment : re-run that same test with a fan blowing air over the heatsink. Do the dropouts still happen? If not, you know which direction to take the design. NB if you really have omitted the second mounting screw as in the photo, you aren't coupling much heat into the heatsink. Thermometers are useful here, to correlate your thermal calcs with reality. \$\endgroup\$ Oct 1 '15 at 11:04

There are two limits to how much power a regulator can handle

a) Its data sheet max current and max power handling
b) Its heatsink capacity. You'll notice on that diagram, it says 'with Heatsink' next to the regulator.

If you read the data sheet, you will see that the max power handling is dependent on you keeping the case temperature low enough. If you don't have an adequate heatsink, the case temperature will rise. Fortunately for you, the LM350 is internally limited, so that before the temperature becomes dangerously high, it stops regulating to protect itself.

If you want 1.7A at 12v out, that's around 13 watts being lost in the regulator (voltage drop * current). That's going to need a significant heatsink. Look in the data sheet to see what the max case temp is allowed to be for 15 watts, then pick a heatsink with a thermal resistance that will keep it to this temperature in your ambient. You will need to do some sums, but fear not, thermal resistances add up like resistance. If you take heat flow as current, and temperature difference as voltage, you can apply a thermal 'ohms law' to the sums.

  • \$\begingroup\$ Thanks for your reply. I am using 9*6 cm size heat sink ( refer image attached ) I think it's sufficient. Right? \$\endgroup\$ Oct 1 '15 at 9:05
  • 1
    \$\begingroup\$ You think it's sufficient. However your circuit stops regulating from time to time. Do you still think it's sufficient? My feeling (I'm not going to look up the degC/watt for that heatsink, you can do that) is that that heatsink is not enough if naturally convected. In the words of my answer, I don't think it's 'significant' enough. I suspect it would be OK if blown with a fan. If you want to limit the case to (say) 80C in a 30 ambient at 13 watts, that's about 4C/w. Go to a catalogue, and look at the sizes of heatsinks of that thermal resistance. Still think yours is sufficient? \$\endgroup\$
    – Neil_UK
    Oct 1 '15 at 9:13
  • \$\begingroup\$ I am verifying my heatsink and as per heatsink calculation. Tj=Pd(Rjc+Rcs+Rsa) + Ta, Here if I use Tj=100 C, Rjc=1.2 c/w (from datasheet),Ta=30 C, Pd=12.9W(as per above calc)...I need to calculate Rsa value but from where to get Rcs value,Is it mentioned somewhere in datasheet or typ value according to package used ? \$\endgroup\$ Oct 1 '15 at 10:27
  • \$\begingroup\$ Assuming 's' is for sink, then Rcs is given by your layer of grease, or insulating washer. There are rough figures for either of those available. Calulating Rsa is either a difficult job for computational fluid dynamics, or an innacurate one with rules of thumb and vertical heatsink dimensions. Much better to just measure in practice, or go with the data sheet value. It can vary by an order of magnitude depending on forcded airflow and orientation \$\endgroup\$
    – Neil_UK
    Oct 1 '15 at 13:17
  • \$\begingroup\$ Is it good idea to connect 1Ohm, 5W resistor instead 0.5Ohm,2W? \$\endgroup\$ Oct 5 '15 at 4:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.