Placid-BP Bipolar Shunt Regulated Power Supply

[Russ White]

Hey Folks,

Here is the Placid-BP under test.

Just wanted you to see that the new boards are here and they are tested. It is working perfectly. I have it set up to drive both sides of a BUF32S at 15V rails.

It is very easy to calculate fixed Rs to set the standing current and output voltage, but you can also use pots and make it adjustable. Fixed Rs are just a bit less noisy if you know what you are aiming for already.

My calculations proved very close, I ended up with +14.94 and -14.75V. Not too bad.

I sounds great, and is nice and stable.

There are two green LEDs per rail that set the reference voltage. Those can be replaced with other references if you need, but the LEDs are a very low noise reference. It would be hard to do any better.

There is also an RC filter prior to the opamp which drives the CFP shunt.

The shunt employs a shaped compensation scheme that keeps the regulator stable even into large capacitances (I have tested 1000uf low ESR Pannasonc FMs). You typiclaly do not want to load a shunt reg with a very high capacitance. It defeats the purpose.

The CCS and Shunt transistors are situated so that they can be mounted directly to a suitable chassis where vertical space is at a premium.

More details to come shortly. I have two more new circuits to test.

Cheers!
Russ

 

[TheShaman]

Good job, Russ.
An objective and subjective comparison to LCBPS would be appreciated!

 

[Beefy]

They look like pretty tall heatsinks! Here's hoping I can get away with the regular 1 inch ones to allow for easy stacking......

 

[BrianDonegan]

1-inch heatsinks will work in many situations. The kits will ship with 1.5" heatsinks to allow more flexibility. There will be times when you may need something larger (2" or chassis). Shunts generate heat. Stacking may or may not be a good idea, depending on what you want to do, and the case you are using.

 

[Beefy]

I was thinking along the lines of dual mono supplies for Buffalo. I'll just wait for more info to come out before I make any decisions......

 

[tailspn]

Hi Russ,

I have need for a bipolar +/- 24V supply at 500MA. Can a Placid with adequate heat sinking accommodate that need? The current is fairly static. If required, I would parallel the pass transistors, say with hard wires and low value sharing resistors. I anticipate using a forced air cooled tunnel heatsink.

Thanks,
Tom

 

[Beefy]

Obviously I'm not sure whether Placid could handle it...... but if it can't then the Sigma22 is a perfect option; assuming it fits in your case, no modifications would be necessary.

http://www.amb.org/audio/sigma22/

 

[tailspn]

Thanks Beefy. Unfortunately the Sigma 22 is not a shunt regulated power supply. There's no problem finding series regulated supplies to fit my power needs, and I've built a bunch of them, including the Sigma 22. There's even a great thread here on DIY about optimizing the 317/337 that looks very interesting:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=143539&perpage=25&highlight=&pagenumber=6

Thanks again,
Tom

 

[Beefy]

And the practical difference is.........?

I've read a little bit on shunt versus series, and from what I can see they are basically just two different sides of the same coin.

 

[tailspn]

Actually not. It's all about supply impedance symmetry. A series regulator turns on more to supply current to a dropping load voltage caused by the load drawing more current. However, when the load voltage is rising, caused by its drawing less current, the series regulator delivers less current, but it's the load that has to absorb the excess energy. That's OK as long as the load can absorb the excess faster than it can be reduced. Otherwise, the voltage at the load rises. In effect, the output voltage of a series regulator only regulates in one direction, ie, when the load is asking for more current. In the other direction, the regulator is coasting.

A shunt regulator regulates in both directions, both by supplying more current, and absorbing excess current from the load. the effect is a much more constant source impedance as seen by the load.

 

[BrianDonegan]

The Placid (kit) uses FJP5200/FJP1943 devices (80W, 230V, 15A). The only other limiting part are the caps, but they are >=50V. I think you should be fine.

 

[Beefy]

Actually not. It's all about......

As I said...... and the practical difference is?

Let me explain. I am equal parts technophile, physiologist, and practical to a fault. Part of me wants the absolute best technical solution possible. Part of me wants to know that it actually makes a scientifically testable audible difference. Part of me wants to know whether I can de-OCD to the point where any of it will affect my musical enjoyment at all.

So I guess I'm wondering where we sit in this particular situation. Pure technical pursuit? Extremely minor audible differences? Critical loss of musical enjoyment?

 

[tailspn]

Thanks Brian!
This should work very well for my applications. Looking forward to the kits.

 

[iko]

As I said...... and the practical difference is?
...
So I guess I'm wondering where we sit in this particular situation. Pure technical pursuit? Extremely minor audible differences? Critical loss of musical enjoyment?

On high transient current draws the shunt can supply it on demand, with minimal voltage drop. This, of course, depends how good the regulator is. If a voltage drop is audible or not depends on how large the sag is. Most people tend to hear a very non-subtle difference with a shunt, usually as improved dynamics.

 

[Spirit02]

Hi Brian/Russ,

Any idea when this will be available (next 4-6 weeks?)
I was just about to grab a 32S and the 'standard' ps option, but if this is just around the corner, then might hold out for it.

Would love to hear some comparisons between the two

Cheers,
Phil

 

[Russ White]

It should be able to handle even 1A load as long as the drop is not crazy into into the reg.

The reg only needs about 3V drop for optimal regulation.

500ma should be no trouble at all.

 

[tailspn]

Thanks Russ and Brian. This kits should work very well in the applications I have planed for them. Could you please advise what percentage of the load current would you recommend be sunk by the shunt transistor? Aside from the loads inrush current at start-up (charging all the capacitors), the + and - 400MA appears relatively constant under signal conditions. All of the opamps run Class A, and the load on them is above 2K Ohm. I'll measure the delta current before I build the kits and know exactly.

Thanks again,
Tom

 

[Russ White]

I don't have any solid data yet to back this up, but I would recommend using your maximum AC load(delta current) + 20-50ma + static load.

The shunt part of the circuit utilizes a CFP. The idea is to keep somewhere between 2-10ma going through the CFP driver. This helps it run in its ideal region.

I currently have mine running at 300ma with a static load of about 200ma and a dynamic load of 10-50ma depending on if I am driving headphones. This seems to give me plenty of reserve current.

You can experiment and find the sweet spot for your application.

The supply is pretty forgiving.

 

[tailspn]

Please excuse my being dense. I interpret from your explanation that the entire supply (one polarity) would draw the maximum AC load(delta current) + 20-50ma + static load. The shunt element of the regulator would be adjusted to draw the max AC Load (delta current) plus 20 to 50ma. I'm just trying to get a rough idea of the power dissipation of each of the power transistors on each polarity of the supply. With 4V or so drop across the series pass transistor, the dissipation for .5A is only 2-3W. With 100ma through the shunt transistor, at 24V, that's less than 3W. This is of course only one polarity, but much less heat than I originally thought. Am I missing something?

Also, what does CFP stand for? Constant something I'm sure.

Thanks again,
Tom

 

[eclectic2k]

...
I currently have mine running at 300ma with a static load of about 200ma and a dynamic load of 10-50ma depending on if I am driving headphones. This seems to give me plenty of reserve current.

You can experiment and find the sweet spot for your application.

The supply is pretty forgiving.

How do you think this would work with IVY or Ballsie? If I understand correctly how you've configured it, Placid should be a good upgrade to your LCBPS for most applications...?

I'm planning to order an Opus ASAP, along with an IVY as a balanced buffer and SE converter. A Placid would be a great touch. (even if I need to wait a bit for it)

Thanks!