High Voltage Regulator
Hi. I have the Pete Millett HV regulator board and something is wrong. It looks like current flow isn't correct in LM317.
With R4=10k and R5=120R (RV1=0), voltage should be 105V @ 34mA but this is what I get: input=140V, Vout=130 with a resistor load of 3,12k.
Specific parts:
Q1: IRF840
U1: LM317 (not HV version)
C1: 33µF
R6: 3.6R
Readings:
Vgs=3.7V
V(R7)=5.4V (meaning 54mA)
LM317 (in-out)=1.6V (should be around 15V)
Suggestions? Thanks
Note: I used PM reg in another application at higher voltage and current and had no issues. LM317 was HV version. Just had to reduce R7 value a bit.
Hi. I have the Pete Millett HV regulator board and something is wrong. It looks like current flow isn't correct in LM317.
With R4=10k and R5=120R (RV1=0), voltage should be 105V @ 34mA but this is what I get: input=140V, Vout=130 with a resistor load of 3,12k.
Specific parts:
Q1: IRF840
U1: LM317 (not HV version)
C1: 33µF
R6: 3.6R
Readings:
Vgs=3.7V
V(R7)=5.4V (meaning 54mA)
LM317 (in-out)=1.6V (should be around 15V)
Suggestions? Thanks
Note: I used PM reg in another application at higher voltage and current and had no issues. LM317 was HV version. Just had to reduce R7 value a bit.
I assume that you've checked for all the obvious things (open solder joints, backwards diodes...)
Normally I would expect that this means that the LM317 is in dropout. That would typically be because the path from ADJ to ground is open. Double check with an ohmmeter that there is ~10k from ADJ to ground (after the caps are all totally discharged).
Pete
Normally I would expect that this means that the LM317 is in dropout. That would typically be because the path from ADJ to ground is open. Double check with an ohmmeter that there is ~10k from ADJ to ground (after the caps are all totally discharged).
Pete
It would be nice to know the voltage at the Vadj pin. Actually the difference Out to Vadj. Is it 1.25V, more, less?
Agree that a "10K" not really connected would do this. Or a 100K where should be 10K (how would I know that??). But various shorts (solder bridges) might do it also.
BTW, I *think* the Gate leg "should" come on the Source side. That's how I usually see MOSFETs, and it makes sense this is a Source follower. And the wrong voltages do not look like a flopped MOSFET (I could be wrong).
Data presented so far, on diagram:
EDIT : illustration removed at request of PRR, please see post #10
Agree that a "10K" not really connected would do this. Or a 100K where should be 10K (how would I know that??). But various shorts (solder bridges) might do it also.
BTW, I *think* the Gate leg "should" come on the Source side. That's how I usually see MOSFETs, and it makes sense this is a Source follower. And the wrong voltages do not look like a flopped MOSFET (I could be wrong).
Data presented so far, on diagram:
EDIT : illustration removed at request of PRR, please see post #10
Last edited by a moderator:
I hope you guys have a lot of fun on New Years Eve! Thanks for your help. Just took a reading and found 10.6V on D1 (1N5245)...
Thanks PRR! We were writing at the same time I guess. I think your diagram will be very helpful.
Actually the input goes to drain on the board.
The voltages you indicated with a ? are correct (just have 131.4V instead of 131.7V)
Thanks PRR! We were writing at the same time I guess. I think your diagram will be very helpful.
The in-adj = 2.85V unless this is not you want to know.
Right on 1.25V
Actually the input goes to drain on the board.
The voltages you indicated with a ? are correct (just have 131.4V instead of 131.7V)
I'm wondering about dropout as well. But remember he wants 105V output.
Crazyfrog - solder a 7-8k resistor (or so) across R5 and then see if you can get any adjustment in output voltage with the pot.
The other trick to try is to make R5 100ohm or so, which will pull more current through the 317
Crazyfrog - solder a 7-8k resistor (or so) across R5 and then see if you can get any adjustment in output voltage with the pot.
The other trick to try is to make R5 100ohm or so, which will pull more current through the 317
Found it!
Re-checked all ( possible) parts with DMM: GOOD
Re-checked for discontinuity from part to part: GOOD
However, since 10.6V on D1 was "weak" (see post 9), I took readings of voltage backwards from there until the adjust pin: BAD from D1 to R3. The bobo was 1N4148. And it was testing good... Never occured to me to check there. Anyway I just put a new one and it's working fine. Thanks to all for your help and interest (especially to PRR for the "educational" diagram
).
Pierre
Re-checked all ( possible) parts with DMM: GOOD
Re-checked for discontinuity from part to part: GOOD
However, since 10.6V on D1 was "weak" (see post 9), I took readings of voltage backwards from there until the adjust pin: BAD from D1 to R3. The bobo was 1N4148. And it was testing good... Never occured to me to check there. Anyway I just put a new one and it's working fine. Thanks to all for your help and interest (especially to PRR for the "educational" diagram
).Pierre
Output at zero. D1 holds Q1 Gate at +15V. Q1 Source rises near +18V. The LM317 is entirely safe for voltage; Q1 must stand about 120V but is a 700V part (tho shunted with a 530V TVS). LM317 can still shut-down on over-current, any large output cap will charge at ~~1A. Less if it goes on long enough to make the '317 hot.
The whole D1-Q1-LM317 rig rises as output voltage rises. '317 should never have over about 18V across In/Out. If we screw-up, D5 should limit to 39V across the '317.
There are more stumbles possible when power turns off. That is why you need D5 D3- so any output cap is discharged around the '317 (which dislikes reverse current). Also D2 takes-down bypass C2.
If D2 is shorted, output setpoint is 210V. With only 140V coming in, it may go to 130V. If D2 is installed (or marked!) backward, output setpoint is 155V-160V, so again we may get 130V.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.