Bipolar supply with phantom rail -- trying to simplify...

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

CurtZHP

Well-known member
Joined
Mar 21, 2005
Messages
634
Location
Allentown, PA
Back in April I had asked about trying to simplify a bipolar supply with a phantom rail.  The big issue was that I had two transformers:  one for the bipolar section, and another for the phantom section.  I'm trying to figure out if it's possible to run both sections from a single transformer without any interaction between the two.  My initial attempts failed miserably because of the interaction between the bridge rectifier of the bipolar section and the voltage multiplier of the phantom section when both are connected to the same secondary winding. 

The phantom section is borrowed from a tube-based build I did a while back, which was using two power transformers anyway; but this is for use with solid state circuits.

If memory serves, the other big problem was isolating the 0V of the phantom section from the 0V of the audio supply section.

Anyway, this time around I'm posting the schematic I have so far.  I had neglected to do that previously.

 

Attachments

  • bipolarsupply.jpg
    bipolarsupply.jpg
    293.4 KB · Views: 52
Thanks, Rob.  Pretty creative way of doing it.  I have to admit, when I first looked it over, I was thinking "He's got that phantom rectifier wired bass-ackwards!  Is that a mistake??"  Nope!  He meant to do that, and it works!
 
I sat down with a piece of scratch paper and kind of walked my way through it.

Let's see if I understand this correctly.....

One "half" of his multiplier is basically a circuit consisting of C3, the two diodes on the left side of BR2, and C5.  A simple voltage doubler.  18VAC goes in, and about 36V comes out.

The other half is C4, the right half of the rectifier and, again, C5.  But, if I'm looking at this right, it's sort of inverted from the first half (for lack of a better term...).  So, 18VAC in gives me -36V out.  (Relative to the same 0V reference...)

Put them in parallel, like he did, and you get a DC voltage that is about 72V.  (The total difference between -36 and +36...)

Run that through the regulator and get about 48VDC.

How'd I do?  Well enough to be dangerous, huh?

 
That is a good design, but a couple things are not quite right.  The obvious one being the 7815/7915 regulators.  These will give you +/-15v not +/-18v.  Additionally, the 317 regulator for phantom is not ideal.  Although the 317 has been used in many phantom circuits, a TL783 regulator would be a better (safer) choice.
 
mjrippe said:
That is a good design, but a couple things are not quite right.  The obvious one being the 7815/7915 regulators.  These will give you +/-15v not +/-18v.  Additionally, the 317 regulator for phantom is not ideal.  Although the 317 has been used in many phantom circuits, a TL783 regulator would be a better (safer) choice.

Yes, I had noticed that as well.  I was planning on using a 783 anyway.
 
mjrippe said:
That is a good design, but a couple things are not quite right.  The obvious one being the 7815/7915 regulators.  These will give you +/-15v not +/-18v.  Additionally, the 317 regulator for phantom is not ideal.  Although the 317 has been used in many phantom circuits, a TL783 regulator would be a better (safer) choice.

The thing is that this version of the 9k board used 18v chips apart from a TL051 for the servo which is a 15v chip so some people ran the whole thing at 15v.  Or if you were like me you had an 18v version of the PSU feeding into another psu board to derive the 15v.

You are absolutely right re the TL783, a much better chip in this position, & that is what I used.

ALSO BR2 isn't orientated correctly in the diagram but is OK on the pcb design.  If you look the diodes between pins 3 & 4  & also 4 & 1 would short the transformer.
 
CurtZHP said:
I sat down with a piece of scratch paper and kind of walked my way through it.

Let's see if I understand this correctly.....

One "half" of his multiplier is basically a circuit consisting of C3, the two diodes on the left side of BR2, and C5.  A simple voltage doubler.  18VAC goes in, and about 36V comes out.

The other half is C4, the right half of the rectifier and, again, C5.  But, if I'm looking at this right, it's sort of inverted from the first half (for lack of a better term...).  So, 18VAC in gives me -36V out.  (Relative to the same 0V reference...)

Put them in parallel, like he did, and you get a DC voltage that is about 72V.  (The total difference between -36 and +36...)

Run that through the regulator and get about 48VDC.

How'd I do?  Well enough to be dangerous, huh?

No.

C3 & C4 just isolate the 48v from the bipolar supply because you will be shorting some stuff out. This is because the bipolar has the ground from a centre tap & the 48v has the ground from one side.

Just ignore the bipolar supply  & imagine C3 & C4 are just wires straight to the transformer.  It's just a normal single ended supply.  The 36v of the transformer becomes 1.414 x 36v =around 51v when rectified.  51v gives you just about 3v for the regulator to work, but if it won't regulate you can back it of to 46 or47v.  The phantom voltage isn't soooo critical & most mics will run off a lot less in their spec sheets.
 
Rob Flinn said:
No.

C3 & C4 just isolate the 48v from the bipolar supply because you will be shorting some stuff out. This is because the bipolar has the ground from a centre tap & the 48v has the ground from one side.

Just ignore the bipolar supply  & imagine C3 & C4 are just wires straight to the transformer.  It's just a normal single ended supply.  The 36v of the transformer becomes 1.414 x 36v =around 51v when rectified.  51v gives you just about 3v for the regulator to work, but if it won't regulate you can back it of to 46 or47v.  The phantom voltage isn't soooo critical & most mics will run off a lot less in their spec sheets.

Ah!  Figured I was off.  But, as you mentioned in an earlier post, BR2 is not correct on the schematic, so that might have been a source of my confusion.  At first it didn't look quite right, but I figured Keith was just doing something creative with what was readily available.  Your explanation makes much more sense!

 
That looks good.

I would add some protection diodes round the 317/337 like on the 783.  Basically the regulators don't like it if the v on the "out" pin is more than the v on the "in" pin & also if the "adj" pin has more volts on it than the "out" pin.  This situation can possibly arise when you switch the unit off.  On the 337 because it's a neg regulator you will have to work out the direction these diodes should face yourself.

Also I normally add a protection diode in the direction "pin adj to pin out" on the 783 & 317 as they don't like the out pin being less v than the adj pin.

On the 337 you will have to work out the direction these diodes should face yourself.
 
In addition to what Rob said, add a 10µ cap from each adjust pin to ground. That reduces the output noise a fair bit, there are extensive tests on the web if you want numbers.
 
volker said:
In addition to what Rob said, add a 10µ cap from each adjust pin to ground. That reduces the output noise a fair bit, there are extensive tests on the web if you want numbers.

Good call Volker.  Should have picked that up myself, I would normally add that cap if I was building this type of supply myself.
 
Back
Top