LM317 strong enough to work with 67V down to 48V?

[777funk]

I have heard that the rating on regulators is on the differential but not sure here. I ask because I think I've had a couple regulators go bad in this build (regulating down to 48V then all the sudden it's up to the full voltage rail).

 

[ruffrecords]

You can get problems at turn on when the 317 has to sustain the full 67 volts. This could definitely damage it. I would suggest you use the TL783 which will operate with input voltages up to 125V.

Cheers

Ian

 

[Tekay]

Shouldn't be any problem using a LM317T. Have you added a diode between pin 1(anode) and 2?

 

[[silent:arts]]

All my smaller PSUs use LM317 for phantom without any problems.

 

[gyraf]

..the maximum voltage specified is the difference between input and output voltage..

The TL783, specified at 125V, has not yet failed for me in a 250V-setup - in/out differential protected by zeners, reverse-voltage protected by a 1N1007. Look at G9 psu.

Jakob E.

 

[Brian Roth]

Using a LM-317 for Phantom regulation is a "fool's quest".  WAAY too many things can go wrong when using a part spec'ed for 30-whatever VDC max from pin-in--to-out.  Pay the extra quarter/dollar and get the correct part, as Ian suggested.

If it has worked for YOU, then ...cool.  But, I would NEVER allow that design into my own systems.

My 2 cents. I want my designs to last more than a few weeks and NOT depend upon which mood  the Electrtical Gods had in mind that particular week after adding a bunch of external parts in order to keep that poor 'lil 317 happy..

Bri

 

[[silent:arts]]

It depends Brian.

While I never heard of a blown LM317 in my small PSUs I do recognize many failing TL783s in the GDIY51 or Jeffs Floor Box PSU.
Mainly because DIYers don't care to spot the TL783 next to the 4 x LM350 and misplace them 

(We should do a custom run of TL783 in a red TO220 case)  8)

 

[777funk]

It's working now, I used the LM317HV this time (had some on hand). I'm not sure what I did wrong. Maybe when things were still loose, I shorted something (very possible).

 

[777funk]

Looks like there's no max input voltage; only differential on the LM317. That's what I thought I saw on the data sheet but it didn't say there 'wasn't' a max input voltage just didn't list one (only max differential). I did a little searching and these guys say that you could run 10kV into a 317 as long as you don't drop more than the max differential...

http://www.edaboard.com/thread261784.html

and this seems to also confirm:
http://electronics.stackexchange.com/questions/94554/what-are-the-maximum-input-current-and-voltage-for-an-lm317

I'll have to try one of these on a tube circuit on the preamp tubes supply and see what happens. I never would have considered this before.

 

[Harpo]

Looks like there's no max input voltage; only differential on the LM317. That's what I thought I saw on the data sheet but it didn't say there 'wasn't' a max input voltage just didn't list one (only max differential). I did a little searching and these guys say that you could run 10kV into a 317 as long as you don't drop more than the max differential...

Where is your 67V number (probably DC) from "with 67V down to 48V?" coming from ?
Most likely from a rectified transformer secondary voltage, and as AC mains isn't a constant that might vary by +/-10% depending on location or time of day, the transformers secondary voltage will reflect this same variation%. If this 67V was from a nominal AC mains voltage, this could rise to above 74V at a different time or location. Your voltage regulator will have at least one reservoir capacitor connected at the regulators 48V output side that takes some time to charge from zero (exceeding the 317 max. i/o differential) to 48V. A maybe fitted capacitor at the regulators adjust pin for improved ripple rejection will prolong this charging time. To build this safe with a 317 or a 317HV, you will need some safety belts, such as clamping zeners, that will cost you more than using a TL783 in the first place. A blown short voltage regulator, giving out at least your 67V, might/will cost you a lot more from blown up microphones. YMMV.

 

[ruffrecords]

Looks like there's no max input voltage; only differential on the LM317. That's what I thought I saw on the data sheet but it didn't say there 'wasn't' a max input voltage just didn't list one (only max differential). I did a little searching and these guys say that you could run 10kV into a 317 as long as you don't drop more than the max differential...

What you say is correct, BUT you need to think about what happens when you power up. The output of the LM317 has some capacitance hanging on it, even if it is only the recommended 10uF or so. At power up, the voltage across this capacitor will be zero and until this capacitor charges up the LM317 has the full 67 volts across it. It may only happen for a few microseconds but that can be enough to kill it.

Cheers

Ian

 

[777funk]

Ahhh charging the output cap from 0 Volts... makes sense! I didn't think about that case.

It's interesting how many 317's I see in 48V Phantom schematics out there. I wonder what the big guys do in say interfaces etc.

 

[ruffrecords]

Ahhh charging the output cap from 0 Volts... makes sense! I didn't think about that case.

It's interesting how many 317's I see in 48V Phantom schematics out there. I wonder what the big guys do in say interfaces etc.

There is a National Semiconductor application note on using the LM317 at higher than rated voltages and if memory serves me they always had a zener diode across the part to protect it.

Cheers

Ian

 

[777funk]

There's a note in one data sheet I've seen that shows 2 diodes in different places, and a cap on the adjust pin. It mentions using this config with the higher voltages.

 

[Matador]

I suppose the G9 "5W Zener from in to out" trick can be used for the LM317 as well, just lowering it down to the 40V maximum differential level.

I had a older design that uses a naked LM317 for phantom, regulating down from about 60V.  I had one bare wire on the phantom output which slipped against the case, which caused a little spark.  The regulator promptly died, even though the short lasted only a few hundred milliseconds.  I think the Zener would have saved it.

So I guess the idea is that at startup, the input rail rises quickly, and one the differential exceeds 40V the Zener turns on and quickly drags up the output node until the until node stops rising. 

It's like a "fast bootstrap" feature.