Neve 1290 gnd wiring, Whatever i do i'm getting hum!!!!

[Spencerleehorton]

Hopefully this modded ssl 9k design will surfice for the supply of the 4 channels, I’m building an external Psu but was going to try it out inside the case just to see if I can get away with it!!!!
Pushing my luck I reckon!
I’ve got at 24v 0.83A toroidal now, before I was powering it with an 18v!!? Need to go to specsavers!!!!
From what I’ve read the LM317 can handle 1A and can go up to 1.5A if you change the zeners?

 

[Spencerleehorton]

Just put together Ssl 9k Psu, could only find a used lm317 and I seem to be getting 75v on the 48v rail!
Regulator must be damaged , will wait until I get some more and replace.

 

[peterc]

Is there a protection diode on the 317? Anode to output, cathode to input?

Check the polarity of this and make sure it is not short circuit.

Peter

 

[Spencerleehorton]

dodgy 317 as i suspected, fine now with new 317, 47.7V and 24.63v no ripple, so i'm going to go ahead and finish up this other 2 channel version, will start on the other 4 channel after tested this one.

 

[Spencerleehorton]

installed the newly improved SSL 9k modded psu which was giving 47.7v and 24.65v and when testing the ch1 furtherest away from the toroidal it was super clean, testing with U87 48v phantom.
Then plug in ch2 and hey presto hum is now back!!! and then swapped over the channel 1 again and hum is now on both channels??
whats going on?
when i turn off 48v the hum goes, so is it my 48v wiring?
I wired the +48v to the led correctly as it wouldn't light up obviously, i've put the - to the screen of the input XLR lead which is connected to pin1 which is connected to the tab on the socket.
I have chassis connection to 12ohm resistor which goes to 0v on the pcb psu.
there is something im doing wrong, also i get a fair thump when putting on the 48v on the other channel that i'm using through the channel that im using, if that makes sense?

 

[Spencerleehorton]

right ive worked out what is happening, the LM317 has gone to 75v again!!

there must be something wrong with the design of this PSU, could someone check this design please?

 

[Spencerleehorton]

ive just put a new 1n4007 diode in along with another diode from adj to output on the lm317 as per the JLM schematic, also changed the LM317 as it had blown again?
only thing i can think of is i had put in 1n4001 diode? or the 48v supply when drawing current is blowing it?

 

[Spencerleehorton]

tested again now and both channels are good! 47.1v and no blown LM317 yet!!!

 

[Spencerleehorton]

i've gone through 10 x  LM317T on this JLM psu version!!! aaahhhhg!

going to make another SSL 9k version as i've had success with this version in the other units.

Still no luck with these LM317T getting 75v at diode bridge and 75v after LM317? must have a dodgy batch?

 

[ruffrecords]

Unlikely to be a bad batch. More likely to be a wiring error.

Cheers

Ian

 

[Pusch3l]

Spence,
You should try Volkers POS48PSU.
He posted it in the White Market.
And maybe you could upload some images as well from your preamp.

Greets

 

[Bo Deadly]

i've gone through 10 x  LM317T on this JLM psu version!!! aaahhhhg!

You have to be careful not to exceed the Vin/out max. Usually TL783 is used for high voltage. It's Vin/out max is 125V.

 

[Spencerleehorton]

As always Ian was correct, bad wiring, 8.2k was put down to the neg rail of the 48v diode bridge rather than the 0v!!
All fine now, haven’t tested fully but 4 channel unit voltages all check out, 48v works on all channels and no smoke!
Just hope there is no hum!!!!

 

[Spencerleehorton]

Crap, failed again!!!

48v rail at 62v now?

That’s where the hum is coming from, looks like LM317 has failed again!

Must’ve dodgy batch!

 

[Spencerleehorton]

The design isn’t right apparently, I’m going to redesign so it is correctly done.
Seems to be some confusion about double or tripling, I’m using a 24vac txf so doubling is fine for me.
24vac x 1.4v = 33.6v x 2 = 67.2v
Then I’ll calculate the resistors needed for the drop of 19.2v.
Which  by my calculations is 240R
So do I get rid of the 8.2k and the 6.8k?

 

[Harpo]

The design isn’t right apparently, I’m going to redesign so it is correctly done.
Seems to be some confusion about double or tripling, I’m using a 24vac txf so doubling is fine for me.
24vac x 1.4v = 33.6v x 2 = 67.2v

+/- 10% for line voltage variation, see your own above measurements of 62VDC ... 75VDC...

Then I’ll calculate the resistors needed for the drop of 19.2v.

No voltage drop calculation. Better read the vregs datasheet.
Roughly (parts come with tolerances) Vout=1.25V x (1+ R1/R2), giving 47.84VDC on paper with R1=8K2 and R2=220R. (R2=240R would regulate down to 43.96VDC on paper). Current setting resistor R2 is set for 1.25V/220R=5.7mA. Reading the datasheet closely, you'd notice this min.current draw might be enough for the LM117 type but probably not enough for the LM317 type. The additional min.current load is drawn by the 47.84VDC/6800R=7mA, giving a total min.current of 12.7mA for a predictable result without any phantom switch per channel activated. As the current in this series string of R1 and R2 is constant, the 8K2 R1 woud require a rating of at least (47.84V-1.25V)x0.0057A=0.265W, so use a maybe 1/2W rated part for R1 and the 6K8 here. Reading the datasheet, you'd as well notice the max.i/o voltage differential of 40V for the LM317 part, where the LM317HV type would allow for a 65V max.i/o voltage differential in order to give this part a fighting chance to survive (you could design in some clamping zener across the vreg i/o to prevent a part exceeding i/o differential, or as squarewave already pointed out, even better use a TL783).
Good luck

 

[Spencerleehorton]

ok, thanks, thats some good info there, so i'm better going for a TL783, will try these.

Ive also found the problem in the design, i will fix the schematic and post it up,  i've got it working now but a little low on the 48v supply, 46v.

seems a lot better now with this correction but i do keep blowing the first in line led for the on switch for the 48v?

 

[Spencerleehorton]

I get it, brackets first!!!!
My 48v design is wrong, I’m going to rebuild it as per access 312 48v design.

 

[Spencerleehorton]

I’m still having a nightmare with this, hopefully will suss out why the original triple design works but it’s wrong?
In the other mod design of this ssl 9k Psu the C5 and C11 are not down to 0v they are only down to the neg of the diode bridge which is connected to the positive of the 24v diode bridge?
I have been informed that this is wrong?
Any comments?

 

[Ilya]

Check JLM PowerStation schematic. There's voltage multiplier for phantom and standard vreg sections. That schematic is known to be correct and working. Check it against the one you're referring to and see what's different.