78xx series regulators - Page 2 - UK Vintage Radio Repair and Restoration Discussion Forum

Radio Wrangler · 1st Feb 2015, 9:33 am

For a better quality regulator.... more current, better accuracy, less noise and low-ish dropout, have a look at Linear Technology's LT1085 and family. They're a bit pricy but I've been using them since the mid 1980s and never had a problem. Farnell and RS do LT stuff.

David

WaveyDipole · 2nd Feb 2015, 12:29 pm

I have a variable power supply that I made several years ago in the 80's that is based on an LM317 regulator that drives a couple of 3055 transistors and gives me anything from 1.5v to 30v at up to 3amps. It wasn't made to the same standard that I would have applied if I had made it today, but it is still going strong and I still use it as a bench PSU from time to time.

On the subject of regulator pinout, I was curious as to why the difference in pinout arrangement for 78xx and 79xx?

jimmc101 · 2nd Feb 2015, 12:48 pm

Quote:

Originally Posted by WaveyDipole

On the subject of regulator pinout, I was curious as to why the difference in pinout arrangement for 78xx and 79xx?

The substrate of the IC is directly connected to the case for heatsinking, it must be the most negative pin of the IC, for the 78xx this is GND for the 79xx it is the input, hence the difference.
(The substrate is always p-type, there is no opposite polarity version in normal use.)

Jim

Ian - G4JQT · 2nd Feb 2015, 9:52 pm

I've been watching this interesting thread and have a question. What is the normal failure mode of an LM317K?

To experiment I have made up a small board with the regulator mounted on a piece of 130mmx70mmx1.5mm aluminium. The smoothed input voltage is 26V and output set to 12.8V.

The test load is a 21W car bulb, so current about 2A - very near the maximum rating for the LM317, so I reckon I'm pushing it in this test.

As the regulator heats up, a scope across the load shows the ripple increasing from a few millivolts, to volts. The aluminium heatsink is obviously inadequate (very hot to touch), but this is only a test. Nevertheless the voltage seems to hold steady.

I'm kind of curious to see what happens if I just leave it on. I assumed it would just shut down - does that mean go open circuit at the output until it cools? I can't quite work out from the data sheet what happens when the junction temp exceeds 150C!

Normal max load will be about 1.5A when I build it properly and use it.

Thanks

Ian

Mike. Watterson · 2nd Feb 2015, 10:58 pm

The other test is to feed 60V in and generate regulated 48V. Then short it.

Or 5V in and 1.35V out and short it.

Or 5V in and 1.35V out and put +60V on output.

kalee20 · 2nd Feb 2015, 11:44 pm

That last test (5V in and then whack 60V on the output) is likely to damage the regulator and is why the anti-parallel diode is advised across regulators.

I don't know about the normal failure mode of an LM317, but it's safest to assume that any semiconductor device, when it fails, fails short-circuit. So in the case of a regulator, add a thyristor overvoltage crowbar circuit if your load is likely to be damaged by a shorted regulator.

Diabolical Artificer · 3rd Feb 2015, 7:24 am

The few LM317's I've had fail still have had an OP, but unadjustable and unstable. Failure was due to a short between pins.

Andy.

Mike. Watterson · 3rd Feb 2015, 11:13 am

Quote:

Originally Posted by kalee20

it's safest to assume that any semiconductor device, when it fails, fails short-circuit.

You *know* what sort of valuable items I want to protect.

WaveyDipole · 4th Feb 2015, 9:56 am

Quote:

Originally Posted by jimmc101

The substrate of the IC is directly connected to the case for heatsinking, it must be the most negative pin of the IC, for the 78xx this is GND for the 79xx it is the input, hence the difference.
(The substrate is always p-type, there is no opposite polarity version in normal use.)

Jim, thanks. That seems to make sense regarding the substrate, but presumably that would also mean that you can't mount a 7812 and 7912 on the same heatsink block?

julie_m · 4th Feb 2015, 10:21 am

Not without at least one mica washer, no.

1st Feb 2015, 9:33 am	#21
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,901	Re: 78xx series regulators For a better quality regulator.... more current, better accuracy, less noise and low-ish dropout, have a look at Linear Technology's LT1085 and family. They're a bit pricy but I've been using them since the mid 1980s and never had a problem. Farnell and RS do LT stuff. David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

2nd Feb 2015, 12:29 pm	#22
WaveyDipole Nonode Join Date: Mar 2013 Location: Leicester, UK Posts: 2,535	Re: 78xx series regulators I have a variable power supply that I made several years ago in the 80's that is based on an LM317 regulator that drives a couple of 3055 transistors and gives me anything from 1.5v to 30v at up to 3amps. It wasn't made to the same standard that I would have applied if I had made it today, but it is still going strong and I still use it as a bench PSU from time to time. On the subject of regulator pinout, I was curious as to why the difference in pinout arrangement for 78xx and 79xx?

2nd Feb 2015, 9:52 pm	#24
Ian - G4JQT Octode Join Date: Jun 2004 Location: Reading/Fakenham, UK. Posts: 1,324	Re: 78xx series regulators I've been watching this interesting thread and have a question. What is the normal failure mode of an LM317K? To experiment I have made up a small board with the regulator mounted on a piece of 130mmx70mmx1.5mm aluminium. The smoothed input voltage is 26V and output set to 12.8V. The test load is a 21W car bulb, so current about 2A - very near the maximum rating for the LM317, so I reckon I'm pushing it in this test. As the regulator heats up, a scope across the load shows the ripple increasing from a few millivolts, to volts. The aluminium heatsink is obviously inadequate (very hot to touch), but this is only a test. Nevertheless the voltage seems to hold steady. I'm kind of curious to see what happens if I just leave it on. I assumed it would just shut down - does that mean go open circuit at the output until it cools? I can't quite work out from the data sheet what happens when the junction temp exceeds 150C! Normal max load will be about 1.5A when I build it properly and use it. Thanks Ian

2nd Feb 2015, 10:58 pm	#25
Mike. Watterson Heptode Join Date: Dec 2014 Location: Limerick, Ireland. Posts: 901	Re: 78xx series regulators The other test is to feed 60V in and generate regulated 48V. Then short it. Or 5V in and 1.35V out and short it. Or 5V in and 1.35V out and put +60V on output.

2nd Feb 2015, 11:44 pm	#26
kalee20 Dekatron Join Date: Feb 2007 Location: Lynton, N. Devon, UK. Posts: 7,087	Re: 78xx series regulators That last test (5V in and then whack 60V on the output) is likely to damage the regulator and is why the anti-parallel diode is advised across regulators. I don't know about the normal failure mode of an LM317, but it's safest to assume that any semiconductor device, when it fails, fails short-circuit. So in the case of a regulator, add a thyristor overvoltage crowbar circuit if your load is likely to be damaged by a shorted regulator.

3rd Feb 2015, 7:24 am	#27
Diabolical Artificer Dekatron Join Date: Jul 2009 Location: Sleaford, Lincs. UK. Posts: 7,667	Re: 78xx series regulators The few LM317's I've had fail still have had an OP, but unadjustable and unstable. Failure was due to a short between pins. Andy. __________________ Curiosity hasn't killed this cat...so far.

4th Feb 2015, 10:21 am	#30
julie_m Dekatron Join Date: May 2008 Location: Derby, UK. Posts: 7,735	Re: 78xx series regulators Not without at least one mica washer, no. __________________ If I have seen further than others, it is because I was standing on a pile of failed experiments.