Bottlehead Hybrid Shunt Regulator....can it be 'scaled up'?

[Mikey]

Hi All,

As some of you might be aware, I've been playing around with a balanced differential preamp this winter.
Recently, I upgraded the constant current sinks under each differential pair to a 'hybrid' approach with
Paul Joppa's coaching, with a noticeable sonic improvement!

Now I'd like to turn my attention to the power supply:

(https://forum.bottlehead.com/proxy.php?request=http%3A%2F%2Fi3.photobucket.com%2Falbums%2Fy80%2Fmpaschetto%2FAtma-Sphere%2520MP-1%2FDSC07354.jpg&hash=fe10ae0c68988431676ab1aa1fd528473538e1e9)

(https://forum.bottlehead.com/proxy.php?request=http%3A%2F%2Fi3.photobucket.com%2Falbums%2Fy80%2Fmpaschetto%2FAtma-Sphere%2520MP-1%2FDSC07357.jpg&hash=948acc7fb1793d71653940e4fd11d38b9e77899c)

The supply has both a 250VDC B+, and a 250VDC B-.  The block diagram for each is as follows:
230VAC into bridge rectifier (325VDC output), 100uF filter capacitor, VB408 series regulator (set up for 250V), 100uF filter capacitor > 60mA out to circuit

I can't help but think there would be some improvement to be made by replacing the regulators with
Bottlehead style Hybrid Shunt Regs....

- Can the schematic shown in the Extended Foreplay Manual be scaled up to handle (shunt) more current?
- Is the C4S upstream of the shunt regulator absolutely necessary?

Thanks,

Mike

[Paul Joppa]

You can scale it up, subject to the limitations of the components (voltage, current, dissipation) with suitable safety margins. Be wary of start-up transients! That is for instance why we have a Zener diode array in the Paramount driver circuit.

There is no high-power replacement for the 431 or other shunt reg chips, so you may have to resort to a discrete version - see John Camille's article in Sound Practices, a long time ago.

Be sure to look at behavior with varying line voltages; I use +/-10%. This analysis is why the stock Foreplay has twice the power supply voltage as the Extended Foreplay, and why the big resistors dissipate so much heat.

The C4S at the top supplies a bit more than half of the dB power supply noise rejection, so I would say it's important but not necessary. It's in the Extended but not the stock Foreplay.

[Mikey]

There is no high-power replacement for the 431 or other shunt reg chips, so you may have to resort to a discrete version - see John Camille's article in Sound Practices, a long time ago.

Paul, I've got the Sound Practices archive on CD, I think I've found the relevant articles:

- Development Notes on a Single Ended 211 Amplifier, by Dennis Boyd and John Camille
- Development of a 211 Amplifier, part 2; Shunt Regulator Concepts, by Johm Camille

Do those sound like the right articles?  I will give them a look tonight, I'm sure they will
raise more questions than they answer.... ;-)

Mike

[Paul Joppa]

Yes, those are the ones I was thinking of. Incidentally, those were designs he was still developing; I believe John was never entirely happy with them and they had stability and reliability problems even in later iterations. I said it could be done, I didn't say it was easy!

We've built a SR45 which provided 36mA at 320 volts. I am not happy enough with that design to recommend it for future uses. I have developed an improved design based on that experience, which I will eventually prototype - we'll see.

We did build a SR2A3 with 380v at 55mA regulated; it took some fixing to get it to work and I would not do it again. I have not yet come up with an improved design that fixes the problems satisfactorily.

Those are the outer limits of what I have actually done; anything more is theoretical at best.

[Mikey]

Hi Paul,

I just finished re-reading the two articles during my lunch break, two things popped out at me:

- They claim the MOSFET shunt regulator sounds as good as the tubed version, with slightly lower dynamic impedance
- Constant current feed provides greater power supply isolation

I believe Allen Wright's 'SuperRegs' are a CCS-fed, MOSFET shunt regulator...in kit form!
Maybe this is the route I'll persue instead of reinventing the wheel, gotta mull it over a bit.

Mike

[Mikey]

One other quick thing:

The Fairchild TL431 datasheet (attached) shows a high current shung regulator circuit on page 6.
Could that transistor be replaced with a vacuum tube?
Would something similar to this circuit work satifactorily?

Mike

[Mikey]

Hi PJ,

I took the easy way out, and ordered a pair of Allen Wrights 'SuperReg' kits for this application.
Basically, the kit is very similar to the approach used by Bottlehead, with a constant current
source preceeding a shunt regulator....all packaged on a 3" x 4" circuit board.

In various Bottlehead designs, I don't see any capacitance added after the shunt regulator.
Also, how much power supply filtering do you suppose is need ahead of the regulator?  Will a
single capacitor before the regulator be adequate, or will I need a 'pi' filter at the least?  I see
that the C-R-C-R-C filter is retained in the Extended Foreplay 3 before the raw B+ is regulated.

I'm asking because I'd like to do away with those (4) big black filter capacitors if possible, to
open up some room to install these circuit boards!

Mike

[coca]

Mikey, can the SOUND PRACTICE archive mags, be purchased on cd, and if so, where?

Bernie.

[Paul Joppa]

...In various Bottlehead designs, I don't see any capacitance added after the shunt regulator.
Also, how much power supply filtering do you suppose is need ahead of the regulator?...

The shunt regulator includes a Zobel-like RC damper across the output which adds a pole to the transfer function to provide feedback stability. If you want passive filtering after, I'd advise to use a series element first, i.e. an RC or LC filter.

Upstream, the shunt reg provides about 90dB reduction of power supply ripple and noise. You can do the math; a single cap is probably adequate from a calculated signal to noise perspective. However, I have not yet found any power supply improvement that was not also an audible sonic improvement, so I'm reasonably confident that the CRCRC plus shunt reg will sound better than a CRC plus shunt reg, or a C plus shunt reg. Only one way to find out for sure!

[VoltSecond]

http://www.siteswithstyle.com/voltsecond/shunt_regs/shunt_regs.html

I started a shunt regulator page along time ago and never finished it.  The link is above. Shunt regulators work good if you don't get greedy with the performance. 

( I did a multiple kV shunt regulator about 20 years ago. That was fun.)

If you really want low noise, you cascade two shunt regulators.  It is easier than making one super high performance one.

[Mikey]

Mikey, can the SOUND PRACTICE archive mags, be purchased on cd, and if so, where?

Bernie.

Bernie,

I got my CD from a seller on eBay:

http://cgi.ebay.com/SOUND-PRACTICES-MAGAZINE-ARCHIVE-CD-/330542286548?pt=Vintage_Electronics_R2&hash=item4cf5db06d4

There is a wealth of information in those articles!

Mike

[Mikey]

I'm reasonably confident that the CRCRC plus shunt reg will sound better than a CRC plus shunt reg, or a C plus shunt reg. Only one way to find out for sure!

Thanks PJ, I'll see what I can cram into the power supply chassis!

VS, thanks for the link...I will need to re-read it a couple of times before it sinks in.

Mike