5 : Excellent 4 : Good 3 : Average 2 : Bad 1 : Terrible

Superregs for your line-level projects

Posted 9th December 2013 at 08:52 AM by jan.didden
Updated 10th December 2013 at 07:02 PM by jan.didden

---

It’s a recurrent issue: you want to build a preamp, a DAC, a phono stage, anything that needs a nominal supply voltage between 3.3 and 15VDC, positive and/or negative polarity. Sometimes you want several supplies to isolate stages from mutual interference via the power supply. So you want a power supply regulator that approaches an ideal DC voltage source as best as possible within reasonable cost. In your search, you inevitably run into the term ‘superreg’ – so where does the name come from and what is it?

The history of very high performance low-voltage regulators is well documented on Walt Jung’s website (www.waltjung.org – look under Library|Regulators & References). An early design that attracted attention was Mike Sulzer’s, published in 1980 and 1981 in Audio Amateur. I added something to that in 1987, and then I was invited by Walt Jung to work on a further improved version. This was published in a series of four articles in Audio Amateur in 1995 by Walt (part 1, 2) myself (part 3) and Gary Galo (part 4). To the best of my knowledge, it is to this design that the term ‘superreg’ refers to. Later on, Walt published some additional refinements in 2000 in AudioXpress, the successor to Audio Amateur.

There’s a lot to say about these designs and although they do look relatively simple, almost all components and details are optimized in long hours of testing. The articles on Walt’s’ website give all the details. If you are interested in detailed measurements, see again the original articles or the recent (2012) measurements comparing many regulator designs with listening tests by Jack Walton in Linear Audio Vol 4. The most important of the comparisons are on www.linearaudio.net under Articles (look for ‘Color graphs for Jack Walton's regulator article in Vol 4’).

One of my contributions to the 1995 series was a compact PCB design for the positive and negative regulator that was sold by Old Colony Sound Lab for more than 10 years. When OCSL folded I received many requests for the PCB and finally I decided to redo it and update it to Walt’s’ 2000 article, and this article is about this PCB.
This article reviews the superreg design and gives an updated PCB design for the positive and the negative regulator. The PCBs are available from the diyaudio store. Each PCB contains one pos and one neg reg, but they are completely separate so you can cut them apart if needed.

Read the full article with schematics, board layout, stuffing guides etcetera here.

(c) 2013 Jan Didden

Posted in Uncategorized

Views 13172 Comments 18

Comments

1. 	Thanks Jan! A good, inexpensive regulator board is sorely needed! Thank you! Now, if we can just get someone to make a good discrete opamp board in single and dual configurations, we'd not need much else! All the best and happy holidays Vince
   	Posted 20th December 2013 at 02:51 PM by vdi_nenna

2. 	Yes, that's a good idea! Let me think about that.... Jan
   	Posted 21st December 2013 at 08:51 AM by jan.didden

3. 	Hi, Jan. I don't have any of the Audio Amateur articles you reference handy for reference, but I have always assumed, perhaps incorrectly, that the term "super-reg" referred to the work of Alan Wright. As I'm sure you know, his regulator consists of a current source driving a shunt regulator, similar conceptually to the Salas HV regs.
   	Posted 22nd December 2013 at 05:52 PM by jlevro

4. 	HI Jan , this baby is good for mA max ? edit I just see the D44H11 are rated 10A !
   	Posted 22nd December 2013 at 10:00 PM by nicoch58 Updated 22nd December 2013 at 10:07 PM by nicoch58

5. 	Max current spec The 10A is a max spec for the pass device but not for the regulator! Check the thread about this, about 1A is possible but for higher currents you need to make some changes.
   	Posted 23rd December 2013 at 08:03 AM by jan.didden Updated 23rd December 2013 at 08:07 AM by jan.didden

6. Super reg term The first use of the term 'super reg' I found was for the Sulzer regulators in TAA, probably early 80-ies. But at the time the term didn't really catch on yet. After the Jung/Didden series in 1995 it DID catch on, so that's why I believe that most people, when they refer to the term 'super reg' refer to those regulators. Alan may have used the term for his design as well; what time would that have been?

   Posted 23rd December 2013 at 08:06 AM by jan.didden

7. 	My guess is that would have been mid-90's as well. So perhaps solid-state guys would think of Jung/Didden, whereas tube guys would think of Wright!
   	Posted 23rd December 2013 at 02:16 PM by jlevro

8. 	Hello! Ive just read your build-guide. I have a question please, before i order this board. how do we calculate the voltage divider values for R6 and R7, R13 and R14 to achieve a given output voltage.. for example, i will require +/-12.0v for one project. Can you please explain and provide examples? I dont want to make any false or erroneous assumptions. thanks Andrew
   	Posted 19th April 2014 at 08:12 PM by FormosaWest Updated 19th April 2014 at 08:16 PM by FormosaWest

9. Andrew, the key thing you need to remember is that in actual operation, the opamp inputs are at the same level.So, if you have a reference of 6.9V at the + input, you want a 12V output, then the resistors on the -input have to be sized to divide down the 12V output to 6.9V. An easy way out: make the lower resistor 6.9V, then the top resistor, which has 12-6.9=5.1V across it should be 5.1k. A more general solution: Vout/Vref=Rtotal/Rbottom. So if you decide to make Rbottom = 1k, then Rtotal = (Vout/Vref) in kohm. Subtract 1k from Rtotal and you have Rtop. Does that help? Jan

   Posted 20th April 2014 at 10:42 AM by jan.didden Updated 20th April 2014 at 10:45 AM by jan.didden

10. 	Thank you Jan! : ) Andrew
    	Posted 20th April 2014 at 06:56 PM by FormosaWest

11. 	How do you modify this to get +/- 24v out? best wishes, david
    	Posted 23rd July 2014 at 04:06 AM by anystereo

12. Hi David, The answer is a few posts above this. Also there is an extensive thread in the forum on customizing the output voltage. In addition to sizing the feedback resistor values you should also adapt the output series zener so that the opamp output is around half the Vout. And of course make sure the electrolytics can withstand the voltages, and that you have enough input voltage for a few volts dropout reserve. Success! Jan

    Posted 23rd July 2014 at 07:19 AM by jan.didden

13. super-reg +/-24v Jan, I got the boards running at +/-24v, with the published values for the schematic, the only modifications was for the resistor pairs that give the 24 volt output, and the more robust 35v caps. I think i understand the suggestion to change the zener from 6.8v to a 12volt zener but i want to make sure, you are talking about D2 & D7 from your schematic? I tried this and the output didn't change, I am assuming it sets up the circuit to operate in its optimum range? Are you also suggesting to change D5 & D10 to a 10volt diode? I tried this and it requires reconfiguring the resistor pairs to get the target output, I am not sure if this results in 'better' performance. I don't have a scope to see the effects of these changes, so any advice you can lend will be helpful. Thanks so much for your assistance etc. David

    Posted 1st September 2014 at 06:16 PM by anystereo

14. Hi David, yes the change in the zeners in series with the output are meant to put the opamp output smack in the middle of its supply for optimum performance. Similarly with the change of the ref to 10 or 12 V which would place the inputs in the mid-supply range (and yes that means you'd need to modify the feedback resistances). But this is more my personal obsession with splitting hairs - as you found out, this circuit is pretty robust and works quite well as-is. Don't bother to make any changes now that you have it up and running! Jan

    Posted 1st September 2014 at 07:47 PM by jan.didden

15. Hi Jan Thank you for sharing this great circuit. In your view is it suitable to provide the 12vdc I need to power my PicoPSU (DC to DC ATX) which is plugged into my audio PC? At the moment I use an old TTI bench supply to provide a low ripple/low noise 12v to the Pico PSU. The bench supply disply shows the computer is drawing about 1.2amps while playing music. Would I need to upgrade the capacitors and diodes to cope with the 18vdc input to give me 12v out? Regards Is there a higher power alternative to the amp AD825 please? can't see one on Mouser or RS

    Posted 18th February 2015 at 11:05 AM by SteveHolt Updated 18th February 2015 at 11:10 AM by SteveHolt

16. 	Hi Jan, The links, both in the web store and the blog, for the BOM and schematic are not working. Regards, ym
    	Posted 26th April 2015 at 04:01 PM by ymwong

17. 	Hi Jan, I'n a newbie so sorry for the dumb questions but I wand to be sure. To get 12VDC our (I just need + and return) and using your example of 1K value to Rbottom I would need 0.74 K for Rtop? Is R7 top and R6 bottom? What should be the VAC output of transformer? Thanks IK
    	Posted 7th August 2015 at 12:09 PM by IZHAKKATZ

18. 	What is the output noise at 15VDC ?
    	Posted 26th June 2016 at 04:37 PM by cobretti

jan.didden

• Join Date May 2002
• Location The great city of Turnhout, BE
• Posts 20,849
• Blog Entries 8

• Blog Categories

  Local Categories

  • Uncategorized
• Recent Comments

  • Can a double blind test really be double blind? by mmerrill99
  • Superregs for your line-level projects by cobretti
  • Don't be such a scientist! by wlowes
  • Tech note: Balanced lines-1 by alfa147x
  • Feedback, or how to be late and be on time the same time, all the time. by jan.didden
• Recent Entries

  • Can a double blind test really be double blind?
  • Superregs for your line-level projects
  • The F-word - or, why there is no such thing as too much feedback
  • Feedback, or how to be late and be on time the same time, all the time.
  • Don't be such a scientist!
• Recent Visitors

  • bighead1707
  • dabenza
  • dognort
  • enser
  • geo2
  • mzperx
  • pronk
  • Sergey Zh.
  • ta12
  • wintermute