Opamp power supply, super regulator alternatives?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Ok let me start of by saying hat im pretty much a newbie to diy audio, ive built a bunch of kits that consisted of stuffing a pcb and connecting wires...etc. But I haven't really ever designed my own circuits. I recently heard a pre-amp a diy audio friend of mine built it was a basic opa627+buf634 circuit but he used a very high end power supply using a few chokes(CLCLC) and Jung super regs...etc. I was amazed at how good this thing sounded I never imagined such a simple circuit could sound so good, im thinking it has alot to do with the high end overbuilt power supply? So anyways, I just built my own version of the preamp using Opa132+buf634 and an alps "blue" pot. Ive been powering it from a pair of 12v 8ah SLA batteries I scavenged and wired up in series to give a basic +/_ 12v supply. I guess I could stick with batteries and come up with a better way for splitting the rails, but id really like to built a regular power supply for it. Id like to do something like my friend did but a bit less complex and less expensive. So any ideas on what the "best" power supply design for a simple pre-amp like this? DO I want to go with a CLC or CRC before the regulators? I rarely see chokes in SS designs except maybe Musical Fidelity and Ayre. Whats the advantage of a choke(clc) over a resistor(crc)?? WHat kind/size choke or resistor would I wanna use, is there a simple formula to figure this out? What kind of regulators do I wanna use? Id like better performance then a basic 78xx or LM3xx but the Jung regulator is to complex for me. I found these two ideas for regulators which look pretty simple and are supposed to perform better then standard 78xx types.

Neil Mcbride Naim mods

Super "Raygulator"

Wow i have so many questions! I almost don't really know where to start. I haven't really built a "high end" power supply before . Ive just built very simple ones that consisted of a transformer, rectifier, large cap, 78xx/79xx, small cap and thats it. The thing that really got me interested in higher quality power supplies was hearing my friend's pre-amp with its super overbuilt preamp. Most of my questions probably seem dumb to most of the people around here, but im a newbie ad trying to learn so go easy on me:) If you guys could maybe draw up some simple schematics of your ideas for a high performance opamp power supply that would be really cool, but id just settle for some info and maybe some ideas on what books or websites I can check out to learn more about this stuff. Thanks alot
frank
 
Hi Frank

You may call yourself a newbie but your questions don't really confirm your newbie status :)

Batteries do some things incredibly well and no matter how much you overdesign a regulator it won't match your batteries in some respects.

First of all download PSU Designer II. It will really help you understand what caps, chokes and transformers do in a PS.

I have found Jung type regulators to sound better than any other type. And that goes against my prejudices: opamps + complexity.

There is an incredible number of circuit topologies you can choose: series or shunt; feedback or open loop. IC or discrete.

These will all sound different and unless you really want to build them all, the easiest solution is to clone whatever your friend got. Down to component types.

There have been several group buys for Jung type regulator pcbs here. Stuffing them with parts is easy and leads to almost guaranteed results. Why don't you do it?
 
thanks guys, as for the reason I dont want to do the Jung regulator is because its way beyond my understanding of circuits and im in the learning stage right now and would really like to understand the circuit im building. So id like to stick with something a bit simpler for now, I am going to build the pre-amp with sort of a modular design so i can swap in different power supplies. I can always try a Jung regulator later on when i have a better understanding of some of this stuff. Im building this pre-amp more for the learning experience then anything else really, but i would like to get good sound too:) As for being a newbie, I have some basic understanding of circuits and what different components do, but no where near the understanding of some of the experts on this forum:) I will download that program and give it a try thanks for the tip!

Frank
 
1) a caution about substituting LM317 (in http://www.neilmcbride.co.uk/preamp.html ) for a 78xx ... they have different pin connections, but you already knew this.

:hot:

2) Concerning differences between op-amps ... basically there is very little as long as the basic specs match up = slew rates, unity gain response, etc. etc. Examples:

http://focus.ti.com/lit/ds/symlink/opa627.pdf ...
http://www.analog.com/UploadedFiles/Data_Sheets/AD825.pdf ...
http://focus.ti.com/lit/ds/symlink/opa132.pdf ...

The devil is in the details. In the above comparing the AD825 (page 7 & 8) to your OPA132 (page 6 & 7), one real difference (the AD825 appears to be better) is caused by the manufacturers' Device Under Test circuit used. The '825 gets a better looking 'scope picture (uSeconds settling time of sq. wave) than the '132. This is basically 'cause Analog Devices uses a DIY technique of putting plastic (polystyrene) snubbing caps close coupled to the op-amp supply pins. Carefully examine the circuit as tested on AD825 page 8 and note the double caps on each V+, V- pin. :bigeyes:

3) to be continued when the guffaws die down ... :D
 
" a caution about substituting LM317 (in http://www.neilmcbride.co.uk/preamp.html ) for a 78xx ... they have different pin connections, but you already knew this"

Thanks, your right I knew about the different pin outs of different regulators, maybe im not as much as a newbie as i thought:) I think even the 78xx and 79xx are different. And as for the different opamps, I don't spend to much time worrying about the specs Ive found that the specs don't really tell the whole story of how good an opamp will sound. I tried many different ones in my cd player and I like the opa132 for its price/performance ratio.


"Hey analog, maybe complex psu (which includes, as in my case, power conditioning) + simple signal circuitry is best? That's my experience."

Im starting to think this is right, especially after hearing my buddies preamp which was just a simple opa627+buf634, gain is set at 3. I probably wont need that much gain since my source is an Emu1212m and it has quite a bit of output voltage. Im wondering now if I even need a gain stage and a buffer stage, a simple buffer and pot maybe all I need and would keep the signal path even simpler. Also is there any sort of formula for figuring out the size of the choke and cap to use in a CLC type of supply? Ive downloaded power supply designer 2 but haven't had much time to to mess around with it and it looks to be made more for high voltage tube power supplies rather then low voltage SS?? .

Thanks
Frank
 
There is nothing simple about a 627/634 pre. Have you considered the number of active devices along the signal path?

By "simple signal circuitry" serengetiplains means something like a single active device, not two dozens :)

PSUD does not have any particular prefence for tubes. As far as formulas go, you'll find them in any electronics text. What simple formulas won't show is the behaviour of the ps with pulse signals, similar to music. PSUD is quite good at that.
 
Yes, my current "simple" signal setup is a single-stage parallel tube amp: 2 parallel tubes, 2 parallel output transformers, 4 resistors and an input transformer, a handful of bolts to hold the tubes and transformers in place, a few dabs of solder. That's it. Cut out the parallel part and it's 1 tube, 2 resistors, 2 transformers.
 
" ... I think even the 78xx and 79xx are different. ..."

Yes, the '78xx is a positive regulator, the 79xx is a negative regulator = with same pin out configuration. These were communly used in "mirror image" pairs, decades ago ... Ex: 7815 / 7915 pair for +/- 15 VDC, etc. from a full wave bridge of diodes and electrolytic caps of up to +/- 30 VDC (or a split ground reference to establish a a "false" ground from up to a + 60 VDC source) = made a nice, compact, easy to design PS board, still very useful ... :cool: ... just be sure to add plastic snubbing caps, close coupled to your op-amps' power pins.

analog_sa: " .... By "simple signal circuitry" serengetiplains means something like a single active device, not two dozens ..."

What he said ... :cool: ... right on ... and daisy chains of op-amps just don't make it either as errors and noise get amplified at each stage in a chain ... :bawling: ... My own "rule of thumb" is no more than three op-amps between signal in and signal out and when coupled between a pre-amp and then out to a power amp, no more than a single op-amp in the signal path, otherwise the dumb thingies just inject too much noise for "golden ear" work ... :whazzat:
 
" Last time i used the little horrors pinouts used to be different in both TO-3 and TO-220. ..."

Yes you are correct = different pinouts ... I boo booed there. :bigeyes:

Actually, the LM7815 and 7915 "complimentary pair" are quite good linear regulators in the TO-3 configuration ... just gotta use dem polystyrene snubbing caps on the outputs, in parallel with the usual electrolytics, usually on a 20 to 1 to a 100 to 1 ratio. ( 10uF electrio to ~100nF plastic works and is noticably better on the 'scope.)
 
What is «super» anyway?

Hi «imperfectcircle»,
I'm not sure I understand your self-proclaimed «illiteracy» in electronic design/circuits. You have been stuffing some (a lot?) DIY kits, for some time; sounds a little like my own background. I guess you do rather easily read schematics, not necessarily in terms of the exact operation, but maybe in terms of the general flow of things? But still you are not confident to stuff a proto-board to test the Jung «super regulator»? You should really find yourself a good, intuitive SPICE simulator to play with. Just don’t get to entangled in «theory», like me.

I would not proclaim my self illiterate to electronics, but in terms of design and general understanding, I'm definitely a novice. Though that have never stopped me...

Her is my two cents for a discrete «super regulator».
The attached schematic maybe a little congested, and it may not have a obvious flow to it. Generally speaking it consists of a gyrator/zener-based tracking pre-regulator, giving the regulator proper about 4V input to output, as is it gives a output-voltage of 27V, it should have a input about 6V above output. The design is something I've been playing with to supply a RIAA design have been working on for to long. I needed a regulator that could reflect my (ridiculously high) ambitions for the RIAA it self. Beein a novice in the field of electronic design, it says something ...

The main design objectives have been very low noise, very high ripple rejection (thus the tracking pre-regulator), and other wise good load-regulation. Minimum voltage drop (tracking etc.), absolute voltage precision, temperature stability and general efficiency (tracking etc.) etc. have mostly been considered irrelevant; though I have implemented a rather crude temperature compensation (the tempco should be adjusted to desired nominal output voltage). Maximum output current is limited to just above 1 Amp, with a fold-back circuit. The design as is, it is still nothing more than a SPICE-model I'm playing with in SIMetrix, thus it is not tried and tested.

Simulated performance thus far is like this (AC-analysis in SIMetrix, including simulated ESR for the capasitors in the schematic):

- ripple rejection better than 160 dB loaded, up to 1M Hz (unloaded, better than 180 dB up to1M Hz), this is much due to the tracking pre-regulator.
- output impedance below 20µ Ohm up to 100K Hz (i.e. the AC component of the outputvoltage when loaded with 1Amp AC is somewhere around -95dBV)
- output noise (hold on to your hat) 1.069µ V 1G Hz BW, or around 200p V/sqrt(Hz) above 50 Hz

I guess any PCB-design will ruin the theoretical performance, which may explain why its still only a SPICE model.

For a negative regulator, turn polarities and substitute complementary transistors - the theoretical performance for the negative version is similar.

Enjoy.

Regards,
KJ

PS If any one should be mad enough to make a proto of this design, or should be inspired to explore it in other ways, please let me know.
 

Attachments

  • reg-psu-pos.gif
    reg-psu-pos.gif
    9.3 KB · Views: 1,218
peranders said:
Hallå, norrmannen! :)

Impressive data but unfortunately you have a parameter called "the real life". Does anything in electronics have 180 dB. Can it be measured?

It would be interesting though to know how good the design would be in real life. :nod:
Hei Per Anders, jeg har vært innom WEBsidene dine og hentet insirasjon til mitt eget overambisiøseRIAA-prosjekt, som desverre har lagt på is litt for lenge, jeg har en rar vegring for å sette igang med prototyping og PCB-design ...

Sorry for the short Scandinavian diversion.

«real life» ... huh ... I've heard about that ... was it a series on MTV a few years ago? :cool:

180 dB ... hmmm ... come to think of it, it is pretty close to a shock wave, i.e. when the velocity of air is higher than the speed of sound, which occurs at around 194 dB SPL. I believe a Danish manufacturer of measuring equipment (I don't remember their name) made a THD-metre a few years back, with advertised resolution down to -160 dB.

I am perfectly aware that there «may be» some difference between theory and «real life». As I tried to stress, these are simulations, and simulations only. The real parasitic of a real life implementation will degrade the results. Output impedance for instance, most soldering resistance, trace resistance, connectors and cabling etc. will dwarf the theoretic output resistance.

However, it’s due time to get my hands dirty...


regards
KJ
 
Hey guys I was just doing a little web surfing and I found this little DIY company in Taiwan and they have this dual regulator that you can buy as a kit or just pcb for around $32 or $8USD
regulator schematic

regulator kit

What do you guys think of this regulator circuit?? The price for the kit is almost to good to be true, I mean it would cost me almost $30 to get just the pcbs for the Jung super regulators!
 
I think you can do better on your own. Three things stand out on a cursory glance; the way that reg is laid-out, you need a raw rail at least 5-6v above the desired output (minimum 3v dropout for the LM317/337 pre-reg; opamp might swing within say 2v of its supply rail; a further 0.7v lost at pass transistor base-emitter). Also, the output from the zener reference could use some RC decoupling for lower noise (e.g. add a good-sized resistor between U3 and C7); and R4 could use a capacitor bypassing it (say 10uF)to reduce noise gain in the error amp.

Dig up Walt Jung's articles on improved regulator design and re-read - there's lots of good hints in there. As it is, implementing a basic LM317/337 well will take you much of the way for peanuts, while you decide how you really want ot 'go for it'.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.