LM317 Issues

[khendar]

I've been fiddling with some LM317 circuits, adjusting resistance values and generally figuring out how they work. I had it hooked up on a breadboard with a 180ohm resistor and a 10k pot and was feeding in 12VDC from a modified ATX power supply. When I switched the power on and adjusted the pot I suddenly had sparks and smoke coming from the pot. I replaced the regulator with another one and got the expected 1.25-10.5V from the output, so I can only assume that that particular LM317 was cactus. Out of curiosity I measured the resistance across the pins of the broken regulator out of circuit and found that they were almost dead shorts. When I measured the other one I found much higher resistances so I came to the conclusion that I must have somehow burnt out the other regulator.

So my question is what could have caused this failure ? There was no smoke or other obvious failure from the regulator. I had been working on a circuit based on the LM337 and so I might have accidentally swapped the Vin and Vout pins. Was measuring the resistance on the pins a reasonable method of determining whether the part was destroyed?

[Rick Law]

There are too many reasons to pin point particularly without seeing how you have it connected.  I blew a couple of 317's by mixing up pins or mishandling the output.  Another one I blew because the breadboard wire fell off my hands onto the board.  So check and double check before flipping the switch is a good thing.

ATX is not a good choice for newbies like us.  I am not an experienced ee-guy hence I said us.  I am just a few car lengths ahead of you but by no means an expert.  The reason why NOT the ATX:  First, some ATX design is unless you have a certain size load, it doesn't turn on.  Second, ATX pumps out too much current - enough to melt a breadboard (and I have one with an 8pin spot melted by the overheated OpAmp powered by 4AA (Eneloops).  The battery clip also melted).

Experiment with a power supply that has current limit as well as volt limit.  You are better off using a 9V battery.  9V battery pumps out only 200-300mA max so it limits the danger.  Another alternative is the CC/CV boards.  CC/CV typically are current-limit rather than constant current.  I use that as a low cost substitute.  Not great, not even good (high noise, not reliable...), but better than an ATX.

Oh, get the rechargeable kinds, then you can re-use and re-use.

[khendar]

I can't provide specifics as to how it might have been hooked up prior to blowing as I was swapping jumper leads around so there were probably numerous ways it could have blown. It is enough for me to know that hooking the pins up correctly has the potential to blow the device, and I know now to be careful about what I hook up to where and to always double check the data sheets.

As for the power supply, I've set it up with a load resistor so that there's no problems with it turning on. The amperage hasn't been a problem yet but when I work on anything sensitive or expensive I do use a battery. I'm working my way up to building an adjustable voltage/current power supply. My current design is based on Jim Rowe's PSU from Silicon Chip, after that I'm going to build an adjustable current version. In the meantime I'm fine with blowing up a couple of regulators or breadboards whilst I figure this stuff out. Magic smoke is kind of exciting