PEterC/Green PSU phantom trouble

HI, so I built this thing to do +24V, and +48V. I left the negative rail stuff off with no problem (see big green X in pic). My +24V is working OK (23.9) but the phantom is not and I'm not sure if I've messed up the doubler. (using 220R and 8K25 resistors with the LM317, by the way)

I'm using a 2x20 VAC trafo, which gives me 28VDC at the junction of the first two diodes in the multiplier. At the cathode of the second diode I was getting 59V yesterday (and at the regulator input) but my output was 30VDC (triple-checked the resistors).  I then put in the 100uF cap marked in the attached picture, but now am getting 48VDC at that 2nd diode cathode (so this keeps the regulator from working atm, what happened to my 59V?).  The green line is a jumper in place of the third diode and the small X is it's respective cap which I left out (for an unneeded tripler, I thought, tried it anyway but gave me 89VDC or something).

I had to scrounge to find caps to finish this board, and ended up using 220uF in place of the 2x 47uF 63V caps in the multiplier. I didn't think that would make a difference ???   I think the multiplier was working fine yesterday (is the addition of the 100uF the reason I am now getting 48VDC after the multiplier?) but at no time has the regulation worked. I swapped the regulators (with each other) and got the same results (+24V worked fine, but still got 30VDC on the 48V output). Anything obvious stick out?

thanks for any help!

Hi mitsos,
I built the original green psu years ago and had to improvise on the caps as well from what I had at hand. I ended up using smaller caps 22u at 63V rating for the later one only (where voltage is already doubled!) and 100u caps at lower rating on the pre-doubler stage without any problem at all. Also used 2x24V zeners instead of 12V zeners. The circuit seemed pretty fool proof to me (i'm a certified fool btw...)....
Exact capacity is (or should be) not relevant for reaching double/triple voltage, but voltage rating is critical.
How did you replace the 47u caps exactly? Please link your pic, this is not clear to me.

So, if you got 60V yesterday, your 100u might be damaged by voltage if not rated 80V (rated 63V with +-10% ?? standard?).
This would explain voltage loss, cap partly shorted....

Just a thought.

Kind regards,
Martin

OK, I went back and put the pic at imageshack instead of attaching here. Seem to have worked, so now my initial post probably makes more sense.

I went ahead and removed the cap I marked as "left out initially" (was 100uF/100V) and now am back to 59VDC after the voltage divider (and at the regulator input). The two multiplier caps are 220uF/63V. hmmmmmm..... so why would a regulator in a circuit that should give 48V put out 30V? (as I said, I have swapped the two regulators into both 24V and 48V spots and both display the same behavior, that is, they work at he + rail, but give 30V instead of 48V phantom).

(And now there's a damned mosquito under my desk and it's bitten my right foot at least 3 times! when I find him...he's a goner!)

What is the connected load to the supposed 48V out?
Most LM317 regulators need at least 10mA load current (Fairchild need 12mA) and the typ. 1.25Vref./220R is only drawing 5.5mA. A 6K8 between +48 and gnd will draw the maybe needed difference.
With the 10uF across the 8K2 for improved ripple rejection you also need a protection diode across the 220R and until charged, this cap nearly shorts out the 8K2, resulting in a voltage difference between reg.input/output exceeding chips limits. Your regulator might be blown at least from your tripler check.
Just some lesser important ideas than catching this damned mosquito ...

the mosquito is still here! I know cuz he just bit my left foot! and with the laptop fan going, I can't hear him to try to squash him!

The reg should be OK. I tried it after my 80-something V experience in the 24V spot and it worked just fine.  But.... I think I know the problem now... There are 2x 10uF caps (on in the adjustment  and one on the output). I have 63V caps there. combining your and smallbutfine's comments, maybe the adjustment cap has partially crapped out since it saw 80V (and is now maybe partially shorted?) and is now causing the reg to see a 7K5 load (when in parallel with the 8K25), giving me 30V. Or the output cap is partially shorted and is causing half my voltage to go down the drain?  Both of these caps were in there during my 80+V experience (they didn't blow up! go Nichicon! so they both saw over 80V).... Later tonight I'm gonna swap them and see if that fixes things...

thanks for the tips

hehe, I like those Nichicons -good caps, maybe they saved you from a mess all over the pcb....time to retire them now, they already saw the light on the end of the tunnel. (Overvoltage)

Hi Mitsos!
How did you solve your problem, exactly the same happens to me. I use the transformer 2x20V a phantom power output gets 30 V
Thank You
bpucekov

I'm pretty sure this was for my 1290, and coincidentally, I have that open next to me and it has a Keith PSU. So Im not sure but I'm guessing I gave up on it cuz I needed the preamp working fast. I'll look around for the PSU board and see where I ended up using it, and if I remember what I did. Sorry not much help.

Thank You
solved
bpucekov

How did you solve it ?
I tried to use keith's for phantom only
but gave up and built a different circuit
[ i was getting too much voltage out ,
somehow couldn't regulate it down ]

What is the voltage at the secondary winding of Your transformer, my problem was that I used a 2x19 V transformer, the voltage doubler was 60 V, which is too small so I just added D4 and C13 and everything works normally. If that is the voltage output is too high, try for R8 (8k2) to put the trimmer 10 k and set output voltage using it.
bpucekov