Phantom power horrific-ness. (Is that word? It is now!)

[CurtZHP]

Any chance you got the wrong pairs of secondary leads?

Never mind measuring volts on open windings, capacitive coupling can be deceiving, ohm out the windings for continuity.

Gene

PS: What is the html tag to get a fixed-width font? I was going to do an old school ascii drawing. Anyone remember usenet?

Can't help you with html.  I can't even get Ohm's law right these days.

The wiring is correct.  The other voltage rails on this supply are working (B+ and 12VDC), and I'm measuring the 15VAC with the transformer in the circuit.  But, somewhere in the multiplier stage, it gets lost.

Stand by for actual measurements....

 

[buildafriend]

If you can afford or have access to a 36 to 40 vac secondary transformer you can skip the extra parts and build a normal bridge rectifier that is RC smoothed in and out of the regulator. Voltage doublers and triplers work but I prefer to stay away from them. Use the protection diodes like gyraf said.

Another thing to know is that phantom power isn't always 48 volts..

 

[CurtZHP]

FOUND IT!!

And of course it was the dumbest of things.

There was a break in ground (0V) after the multiplier stage and the first filter cap before the regulator (C111).  So basically, the rest of the circuit had no ground reference, so the poor little electrons were like, "Dude, we don't know where you want us to go!  Swearing at us and calling us names doesn't help!"

Or rather, there was no ground connection in the first place because some dumbass forgot to put it in!  :-[

The reason it worked before was because it was likely seeing ground from the filament supply rectifier.  Now that they've been separated, no ground.  Gotta add one.  Oh, look!  48VDC!

But the electrons and I are still not quite on speaking terms just yet.  Not being that easily fooled this time around, I connected the 48V supply to the sillyscope.  Still seeing some nasty square-wave looking grunge.  (FWIW this is with the time base set to 10 ms/div.)

Perhaps adding some additional filtering caps south of the regulator would help?  Just having one there seems kind of stingy.

Bottom line:  Got it working again, but no better than before.  Back to Square One.

On a side note:  Jakob, if you're still following this, here's a tiny piece of good news.  I found a set of ZD39 diodes in the drawer, so I can add those around the B+ regulator.

 

[CurtZHP]

Looks like the issue with the grunge on the rail goes back to something PRR said....

Are you brave enough to look at the input of the regulator?

I have doubts you are getting >51V at the *dips* of the raw DC. That throws down-spikes into the "48V".

Sure enough, I'm only seeing 50VDC on the input to the regulator, which means this whole supply is just underpowered.  Like Jakob's design, I need to get it to where I'm seeing about 60VDC before the regulator to give it something to work with.

 

[ruffrecords]

Looks like the issue with the grunge on the rail goes back to something PRR said....



Sure enough, I'm only seeing 50VDC on the input to the regulator, which means this whole supply is just underpowered.  Like Jakob's design, I need to get it to where I'm seeing about 60VDC before the regulator to give it something to work with.

A TL783 really needs about 20V across it to be effective at reducing ripple. I use a 50VAC winding which when rectified gives about 70V dc into the regulator.

Cheers

ian

 

[CurtZHP]

I'm even starting to wonder if the little bit of ripple I see is what I actually heard.

I think the noise I heard had more to do with the other problem that PRR spotted, namely that the multiplier got crossways of the heater rectifier and was creating a short somewhere.

That makes sense in light of what I heard when I switched the phantom on for one channel.  I wasn't hearing just a little hum.  It sounded like a buzz saw!  And there was NO sound from the mic.  (Hope I didn't kill it!)  It sounded like what you'd expect to hear if something was either shorted out, or you'd stuck the connector in your mouth.  Loud and nasty.  I wouldn't be surprised if the changes so far have tamed things considerably.

Still, I want to see what I can do clean it up a little more.

 

[ruffrecords]

Can't help you with html.  I can't even get Ohm's law right these days.

It's either:

Teletype

or its:

code

more likely the latter.

Cheers

Ian

 

[PRR]

> What is the html tag to get a fixed-width font? I was going to do an old school ascii drawing

Try (as Ian said)  "tt" (teletype!) tag.

H  H EEEEE L    L      OOO      W  W  OOO  RRRR  L    DDDD  !!
H  H E    L    L    O  O      W W W O  O R  R L    D  D !!
HHHHH EEEEE L    L    O  O      W W W O  O RRRR  L    D  D !!
H  H E    L    L    O  O  ,,  W W  O  O R  R L    D  D   
H  H EEEEE LLLLL LLLLL  OOO  ,,    W W  OOO  R  R LLLLL DDDD  !!

H   H EEEEE L     L      OOO       W   W  OOO  RRRR  L     DDDD  !!
H   H E     L     L     O   O      W W W O   O R   R L     D   D !! 
HHHHH EEEEE L     L     O   O      W W W O   O RRRR  L     D   D !! 
H   H E     L     L     O   O ,,    W W  O   O R   R L     D   D --  
H   H EEEEE LLLLL LLLLL  OOO  ,,    W W   OOO  R   R LLLLL DDDD  !!

 

[Gene Pink]

Try (as Ian said)  "tt" (teletype!) tag.

Actually, that's the first thing I tried. Didn't work in the preview.

I'll try again, a large X, and transformer windings in TT

 \    /
  \  /
   \/
   /\
  /  \
 /    \

   ________
  |        1
  |
  |___  ___2
      \/
   ___/\____
  |        3
  |
  |________4

OK. figured it out, my fault.

I am used to the post window being a fixed width font from the old days. And this one ain't.

Pardon the topic swerve. To get back on point:

1) 47uF for the three trippler caps? Wimpy. This isn't a television CRT trippler that runs at the horizontal output freq of 15Khz, I would go with 470 or 1000uF caps, to really stiffen it up. This may fix your problem.

2) Why are you loading the 48V regulator with 16mA just for the reference input? Those reference inputs are high Z and just need a voltage, perhaps you can increase the values in that 1.24 volt divider by a factor of 10.

Gene

 

[ruffrecords]

2) Why are you loading the 48V regulator with 16mA just for the reference input? Those reference inputs are high Z and just need a voltage, perhaps you can increase the values in that 1.24 volt divider by a factor of 10.

Gene

The TL783 and friends do need at least 10mA or so drawn from the output in order to regulate. - the actual value depends on a variety of factors.  Check out the data sheet. I see lots of designs that do not adhere to this and I wonder just how well they work.

Cheers

Ian

 

[CurtZHP]

1) 47uF for the three trippler caps? Wimpy. This isn't a television CRT trippler that runs at the horizontal output freq of 15Khz, I would go with 470 or 1000uF caps, to really stiffen it up. This may fix your problem.

Last night I bumped them all up to 100uF/100V caps (what I could get my hands on...)  That took the ripple right out!  Looked at that rail on the scope, and it was as flat as the workbench.

I also added another "stage" to the multiplier to bump its output to a little over 63VDC to give the regulator more to work with.
Now the final output is more like 44-45VDC.  I'd like to get it a shade higher, so I'll need to swap out a resistor or two, but I'm almost there.

 

[mjrippe]

Now the final output is more like 44-45VDC.  I'd like to get it a shade higher, so I'll need to swap out a resistor or two, but I'm almost there.

The spec for phantom power is 48 +/-4 volts, so you are fine.  In my experience, standard value resistors won't get you any closer.

 

[CurtZHP]

The spec for phantom power is 48 +/-4 volts, so you are fine.  In my experience, standard value resistors won't get you any closer.

But this was measured on the raw supply output.  I haven't yet accounted for the 6.81K resistors to the input jacks, or the mic load.

Testing all that tonight.

 

[buildafriend]

A TL783 really needs about 20V across it to be effective at reducing ripple. I use a 50VAC winding which when rectified gives about 70V dc into the regulator.

Cheers

ian

thanks for sharing

50x1.41=70.5

then the 20v drop assists in smoothing.. nice.

 

[CurtZHP]

Good news!

Changed a couple resistors and then tested the phantom supply under load.  50VDC!  Smooth as glass!

This weekend, I'll hook the supply back up to the audio side of things and see how that goes.