LM317 based PSU for 48v troubleshoot

[ohm]

thanks for the answers,

regarding current draw both supplies are for 8 1272 type mic/line amps


well I wired the supplies in and connected it all up how I thought it would work and smoked the large 2200uf cap on the 24V supply

I've discovered this was from trying to create a common ground wiring both 0v from each circuit together

Is there a way I can do this?

The power inputs on the preamps have only 3 points 0v 24v and 48v

It doesn't make sense to need two separate 0v points

Checked both circuits for shorts and they both work fine independently and put out the correct voltages

If I leave the 0v on the 48v supply 'hanging' and try to use the 0v from the 24v supply to get 48v this way then I end up with around 26v  :-\

Perhaps I've missed something?

 

[abbey road d enfer]

thanks for the answers,

regarding current draw both supplies are for 8 1272 type mic/line amps

Then it means you should design for 120mA; the regulator would dissipate about 2-3 W. I guess your heatsink is sufficient. Anyway in practice, the current draw would not exceed 30 mA.

well I wired the supplies in and connected it all up how I thought it would work and smoked the large 2200uf cap on the 24V supply

I've discovered this was from trying to create a common ground wiring both 0v from each circuit together

That means something is stupidly wrong, because both 0V normally can be connected without problems. The fact that it's the cap that blew seems to indicate that there's AC or largely exceeding voltage there, which should not be the case.

Checked both circuits for shorts and they both work fine independently and put out the correct voltages

Can you check the voltage between both 0V?

If I leave the 0v on the 48v supply 'hanging' and try to use the 0v from the 24v supply to get 48v this way then I end up with around 26v  :-\

I don't see what you mean. An updated schemo and explanation of your measurement would help.

 

[ohm]

between both 0v is 15.7v

for the record

The 24v circuit outputs exactly 24v

The 48v circuit outputs 48.7v

using the 24v circuit's 0v as the 0v reference for the 48v circuit there is 26.2v

the voltage between 24v and 48v is 6v

does any of this help?

 

[ohm]

here is an updated diagram

Voltages between

A + B = 48.7

C + D = 24

A + D = 26

A + C = 6

C + B = 36

B + D = 16

 

[abbey road d enfer]

You can't do that! C2 shorts out one leg of the secondary.

 

[ohm]

ahhh ok, was just following the voltage doubler from this... unless I read it wrong?  :-[

 

[abbey road d enfer]

ahhh ok, was just following the voltage doubler from this... unless I read it wrong?  :-[

It is correct, but not compatible with the configuration used for the 24V rail.

 

[ohm]

ok thanks! will rewire...

are both the caps on the input on your diagram 2200uf?

Think I've only got 1000uf 63v at hand, will they be too low?

 

[abbey road d enfer]

ok thanks! will rewire...

are both the caps on the input on your diagram 2200uf?

Think I've only got 1000uf 63v at hand, will they be too low?

They should be ok.

 

[PRR]

> I never had any problems with that.

I have.

I threw-out three perfectly good HV regulators before I noticed the minimum current spec. Since the volt-setting resistors were already there, I just added a dummy-load, around 10mA. ITC, say 4k7. Which would be 0.48W in a probably 1/2W part, HOT, but proved the point (the reg was fine, I was just underloading).

Maybe the actual specs have improved over 30 years (better process). I'd still tend to observe the ratings.

 

[ohm]

All working!

Thanks for all the input

final schematic

What to do about the value of R1 then if 240R is too high?

 

[tubegeek]

Ian your diagram has 100k across the DC inputs. Is this needed?

I'm not Ian, but those resistors are used to bleed the charge off of the filter caps. I recommended you keep them. Otherwise, when you work on the power supply you will have to manually short out the caps to remove the charge or risk being shocked by stored charge, which can be quite a vicious bite!

There should ideally be bleeders on each end of the umbilical connector if your power supply is not in the same chassis as the audio circuit.

 

[abbey road d enfer]

> I never had any problems with that.

I have.

I threw-out three perfectly good HV regulators before I noticed the minimum current spec. Since the volt-setting resistors were already there, I just added a dummy-load, around 10mA. ITC, say 4k7. Which would be 0.48W in a probably 1/2W part, HOT, but proved the point (the reg was fine, I was just underloading).

Maybe the actual specs have improved over 30 years (better process). I'd still tend to observe the ratings.

I should have mentioned that I always had some kind of minimum load - typically an LED running at about 10mA. For the typical application we see here (voltage differential <50V), the minimum total current draw is 6mA, so adding a 22k bleeder resistor would be safe, or lowering the 240r to 180, and the foot resistor to 6700 ohms.

 

[moamps]

All working!

Hi,
please confirm that voltage on the C8 is lower than his voltage rating.
Regards,
Milan

 

[ohm]

hmm voltage on + of C8 is around 71

Looks like I need higher rated capacitor than current 63v?

Hi,
please confirm that voltage on the C8 is lower than his voltage rating.
Regards,
Milan

 

[ohm]

Shall I put 100k in parallel with C8 then?

I'm not Ian, but those resistors are used to bleed the charge off of the filter caps. I recommended you keep them. Otherwise, when you work on the power supply you will have to manually short out the caps to remove the charge or risk being shocked by stored charge, which can be quite a vicious bite!

There should ideally be bleeders on each end of the umbilical connector if your power supply is not in the same chassis as the audio circuit.

 

[ohm]

If I'm to put an LED + resistor across each supply to give an indication of power and supply a load how do I determine the value?

I'm getting 4800R (4k7 I guess) from 48v / 10ma

(guess 2k7 for 24v?)

Is this correct?

Actually I don't want to guess.. how do I find out how to work this stuff out?

Where do we arrive 10mA is this the typical draw for an LED?

I should have mentioned that I always had some kind of minimum load - typically an LED running at about 10mA. For the typical application we see here (voltage differential <50V), the minimum total current draw is 6mA, so adding a 22k bleeder resistor would be safe, or lowering the 240r to 180, and the foot resistor to 6700 ohms.

 

[moamps]

hmm voltage on + of C8 is around 71

Connect - pin from the rectifier bridge DB2 to the ground and recheck voltage.

 

[moamps]

how do I determine the value?

Use your law 

Voltage across LED is 2V typ.
So, R=(48-2)/0,010=4600ohm, any resistor from 4k7 to 10k will be fine.
Regards,
Milan 

 

[ohm]

Are you saying that the - of DB2 should not connect the + DB1 and instead go to ground?

Connect - pin from the rectifier bridge DB2 to the ground and recheck voltage.