Hi Al, are your caps as close as possible to the 317's leads, if not you may have a power oscillator. These devices burst into oscillation at the drop of a hat. If you have more than about 1uF on the output, wire a diode backwards across the chip. If not the cap tries to discharge into the chip and it dies PDQ. 3 diodes and a fuse across the output is a good crowbar. I used similar circuits in an article for WW many years ago and it worked fine.
Is this the low dropout 317 you are using? If not it may need a higher input voltage to regulate properly.
Switch on surge on the valves should not be a problem as the 317 has internal current limiting.
These are nice devices but are full of traps for the unwary user.
Ed
Hi Ed. Transformer sec'y. is 6.3 v. CT. Full winding connected to a bi-phase rectifier (CT = 0v.) then to a 10,000 uF reservoir cap. Two wires, each about 4 inches long from that cap to the regulator circuitry, which was the traditional bird's nest! No cap. at all across the output. Voltmeters connected across reservoir cap and Vout. Dummy load for initial testing was 100 ohm R. At which point, the 2v N/L dropped to about 1.3v. A wander around with the 'scope did not reveal any signs of H.F. oscillation.
Al.
Strange indeed.
I've always found LM317s to be reliable and easy to use. I've only managed to blow one up and that was by sticking HT up it's output.
- Joe
Got the same issues with a battery set here, but this one has two HT voltages and requires a negative 0-12 volt bias. For the amount of negative current required probably a 7905 or 7912. Agree with the hefty diodes across the output, a couple of zeners should help.
Hi Al, What is the voltage on the resevoir cap, if you are rectifying 3.15 as bi-phase you may only get about 4.5 on the cap as the input to the 317. This could be a bit low.
Try a bridge rec across the whole of the 6.3 (disconnect the CT) and see if it works better. You then need to look at the dissapation on full load.
Ed
This crossed my mind too, but as he said he'd tried it on a 15 volt DC PSU, I discounted it. He also built the PSU, but I'm sure it's fine
Any progress to report today, Al?
Cheers,
Mark
Progress: yes, Mark. I went back to that cct. using the LM317 and had a good look at my build (checking for wiring errors and poor layout): no problems found. I had a thorough look at it with the 'scope: no signs of oscillation. The one thing that did puzzle me however was the voltage on the control pin: it is fed from a potential divider, 240 ohms and 150 ohms, from Vout to 0v. With Vout at 2 volts, the control pin was at about 0.3 volt. To me, that implies that the control pin was taking a lot of current: methinks too much, and thus implying a defective LM317. I tries two other LM317s: same result. The PSU source was a bridge rect. across the 6.3 v.a.c. winding. (I am unsure about the history of these particular LM317 devices.
I know that one of them I have used before. But . . . . )
At which point I stopped and changed tack, and began thinking about a simple linear regulator using discrete components. And that is what I now have working, albeit as the traditional 'lash-up', on the bench as I write this. It's a simple two-transistor affair and is good for 2v nominal @ > 1 amp with 20 mV peak-peak ripple. The regulation isn't heart-stopping: 2 volt at N/L falling to 1.9v @ 1 amp.: quite adequate for the purpose. Tomorrow I shall continue to get a crowbar cct. working, then put the whole issue (LT and HT sources) together and give an extended soak test. At that point I will then present the entire cct. here for a peer review: don't hold your breath!
Al. / April 8 //
Hmm... The ADJ current should be less than 100uA - from what you've written, these ICs are definitely unwell. I've found them to be pretty well behaved over the years, and TBH I'd probably persevere with them here. If I get some pare time later today (perhaps this evening) I might knock one up, just in case...
Still, look forward to seeing the transistor alternative
Mark
Likewise.
ICs may be simple, reliable &c but not as interesting.
- Joe
Update: 2.0 v.d.c. 1-amp regulated PSU: cct. diagram:
[attachment=4509]
Performance figures.
Load Vout Ripple (p/p)
N/L 2.0 -
750mA 2.0 10 mV
1.0 amp 2.0 20 mV Design full-load
1.3 amp 1.9 20 mV Design overload
Vout adjust range: N/L: 1.4 → 2.2 v.
F/L: 1.4 → 2.0 v.
Vout normal setting: 2.0 v.
Crowbar trip voltage: 2.25 v.
------------------------------------------------------------
Comments.
It's not the most elegant design I've seen, but then again, what were my design criteria?
1. Absolute minimum cost! It's not so much that I'm a miser: it's more a question of having accumulated a huge 'junk box' over the years, there comes a time when one has to do some pruning!
2. To produce a design that is reliable but not overtly complicated.
3. To specify parts of which I have ample spares (even the transformer!)
4. To meet a minimum performance spec.: 2.0v @ 1-amp. regulated & low-ripple being the 'design centre' spec. (i.e.: a traditional 2v accumulator, but one that never needs re-charging).
And with that . . . . your comments are cordially invited.
Aside: the HT 'battery eliminator' section will be discussed later.
Al. / April 10, 2012 //