Reasons LT regulator fails under load

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi All,

I have built 2 identical LM317 regulators, circuit is straight from Valve Amplifeirs 4. Both regs use the same transformer and rectifier, reservoir and common mode choke. Reason for 2 is that I need to power more than 2A heater current, so have split the load into 2 regs. This was the easiest solution with the parts I had to hand.

Both regs test OK reading at 6.5V unloaded, but under load one reg drops out to 0.6V when the other is fine. And when disconnecting the good reg the problem remains. The load is not the issue as it is under 1A, so I am looking for other solutions.

Can anyone suggest reasons why it is failing?

I did search here but could not find the answer, even though I am sure this issue has come up. So maybe this will also help any future searchers too.

Cheers

Charlie
 
It was tested with valve heaters. Both circuits are identical so am unsure where to look next, other than checking and double checking wiring.

It did regulate at 6.5V for a short time under load, but quickly dropped to 0.6-0.8V. I havn't tested with a resistive load.
 
It could be power dissipation.

What is the supply to the regulator?

Either way, one solution could be a current booster for the regulator.

see jpg
 

Attachments

  • LM317-BOOST.JPG
    LM317-BOOST.JPG
    21.3 KB · Views: 257
They are not isolated pre regulator, but they only share a common ground post regulator. The bad reg fails when it is the only device connected so I am assuming it is not interference from the other reg.

The supply is 2 x 9V 1.6A transformer in parrallel, bridge, 10,000uF, common mode choke. So current shouldn't be a problem, and as I said the other reg is working fine under any load up to 1.5A. I will see if I have a PNP to try that out, thanks for the suggestion DUG.
 
Is the failing one getting too hot? It could be that both are running near their limits, but manufacturing tolerances mean that one is just OK and the other is just over the edge. Have you checked that the initial start up current is within the regulator spec, and that it won't overheat during these first few seconds? The chip can overheat very quickly, as it is small and has small thermal capacity.
 
Are you sure it's not a thermal issue?

You are feeding the regulator with 12.73VDC (9* 2^0.5) so a voltage drop of 6.23V. For a current of 2A you are dissipating 12.46W

If we assume a room temp of 25C and a Rth of 5C/W we get a max temp of 25 + (12.46*5) = 88C. The shutdown limit is 125C.

I guess it's not a thermal issue :D


Still quite toasty so I would add a small heatsink.
 
Last edited:
There is definitely something wrong, the good reg will not work at all if the bad reg is connected. I guess I should of not shared the recitifier etc at all.

The markings on the bad 317 have vanished so that might suggest over heating, so I will replace it and test again. It may be as DF96 suggests as the regs are pushed to well over 1.5A on start up.

Currently I have a 10uF on the Adj to ground, is that big enough? I have some voltage to play with so can also try a resistor on the output.
 
In my experience LM317s tend to be fairly rugged things. I've abused a fair few leaving some shorted for extended periods of time, with them getting ridiculously hot and then working just fine with the short removed. I've also used a few in situations with less then adequate heatsinking (fail the finger test but not beyond spec) and I've never had a problem. I've also never blown one either despite the abuse.

I'd be looking for an overlooked implementation error, like a dead resistor or a small solder bridge or something. Either that or one of the regs is faulty. If connecting the unloaded, second regulator, up to the shared power supply causes the functioning regulator to stop working, then this implies that the second regulator is drawing way too much current and is causing the unregulated voltage to drop significantly.
 
There is a total of 6 valves split into 1 reg feeding 4 and the other feeding 2. I tested the working reg with the other section of valves and it works fine, so I am assuming that there is an issue with the reg itself.

OK, if the good reg powers either circuit just fine, the fault must ly with the bad regulator.
Even though these are very rugged devices that are protected against thermal and electric overload, they can fail.

There is definitely something wrong, the good reg will not work at all if the bad reg is connected. I guess I should of not shared the recitifier etc at all.

If the outputs of the regs are not interconnected (which I assume based on your answer of separated loads), then it must be the voltage on the pre reg side that gets pulled down, perhaps by an overload. Measure the voltage on the 10,000 uF cap in the setup quoted above and compare it to the voltage when only the good reg is connected (and powering its load), there shouldn't be too much difference. If there is, it looks like the faulty reg draws much more current than it should.

The markings on the bad 317 have vanished so that might suggest over heating, so I will replace it and test again.

That's a sign. Another distinct sign of overheating is when the metal tab has gotton a blueish discouloration.

It may be as DF96 suggests as the regs are pushed to well over 1.5A on start up.

That may be the case and we're indeed dealing with tolerance issues between the regulators. Replace it first and see what happens.

Currently I have a 10uF on the Adj to ground, is that big enough? I have some voltage to play with so can also try a resistor on the output.

The cap between ADJ and GND mainly improves ripple rejection, but in the datasheet of the LM117/317 there's a "slow turn on" circuit in the application notes (page 16). This might be interesting for preventing those high current peaks at power on by (relatively) slowly letting the voltage on the output rise to desired voltage. This would "preheat" the heaters and lower the initial current draw.
 
Last edited:
Thanks for all the comments so far, all very useful.

I tried a few different setups and seperated both supplies completely, and there was still instability now in both regs. So after some more trial and error I found that actually the common mode chokes were causing the issue. I wired out both chokes and now both regs are working fine, and have been stable for the past hour. They are running at about 60 degrees C so within their spec.

I will keep trying to get back to the original schematic as there must be a reason for the instability.

Can anyone offer suggestions on this?

Cheers

Charlie
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.