Super Regulator, collecting the facts

[peranders]

The thing with LM329 is the very low noise.

 

[johnferrier]

The thing with LM329 is the very low noise.

The LT1634 is also low noise (10uVpp). The highest voltage is 5V.

Okay, just noticed another consideration. The LT1634 (which I use in a precision current source) is a 9$US part. The LM329 is a $1 part.

If it needs to be 6.9V, then the LM329 is the part. It's possible to solder such a part to SMD pads (if you want to go no through-hole).


JF

 

[jackinnj]

The super regulator in the 4/00 issue of Audio Electronics used Panasonic HFQ or FC capacitors -- 120Uf, 25V

Panasonic has discontinued these:

http://www.maco.panasonic.co.jp/www-data/pdf/ABA0000/ABA0000CE74.pdf

 

[jackinnj]

the HFQ's are discontinued, not the FC's -- this was the point i was trying to make

here's a link through DK to the Panasonic datasheet
http://rocky.digikey.com/WebLib/Panasonic/Web data/EEA-FC, EEU-FC (Series%20FC,%20Type%20A).pdf

 

[jwb]

The LM329DZ is a TO-92, if i recall correctly. Most SMD packages are not smaller than that. I would say use the DZ package.

Am I the first to have implemented a variant of the SOAPSUDS? I got four working right here on my headphone amp. They sound dandy, and no startup or oscillation problems.

BTW, I used Panasonic FC caps for everything. Good, cheap cap in a compact packages. And LM329AH, the ridiculous mil spec reference.

 

[jwb]

Good to hear. Any photo?

Any comments from Fred?

I put a very bad photo up here. That's a negative reg in the nearest part of the board. You can see the tin can on the LM329AH, and the AD825 behind it.

Sorry for the bad photography. My mobile phone is the only digital camera I have.

 

[peranders]

Looks very very bad because you have som many parts. I have heard that many parts is bad for the sound.... no just kidding. Looks good. It's fun to do what ever you like.

 

[jackinnj]

RFI protection --- thinking of decoupling the AD825AR with 499R resistors aw WJ did in the tail end of his article -- before I burn some boards here I was thinking to decouble the inputs to the error amp with resistors and 100 pF caps from the inputs to the output -- since I am working in SMT it would be virtually impossible to do it later.

 

[peranders]

It's no harm to add "just in case" parts.

 

[dimitri]

about groundplanes

... large groundplane ...
not everything that simple

http://www.elecdesign.com/Articles/Index.cfm?ArticleID=5944

 

[jackinnj]

meant to part you from your money

now I recognize that AD825AR's cost about $3, and AD797's about $9 when you buy a tube of them -- this is taking profit margin to an extreme:

 

[jackinnj]

instrumentation or differential amp for error amp?

the regulators we discuss in these posts use an error amp open loop. why would one not choose to use more modest gain in a differential configuration?

 

[Fred Dieckmann]

open loop?

The circuit is an emitter follower inside the closed loop of an op amp configured for unity AC gain. The open loop impedance of the follower is divided by the open loop gain of the op amp. The AD 825 is nice because it has a constant open loop gain over most of the audio frequency range instead rolling off at 6dB per octave in this range. A constant output impedance instead of an impedance that rises with frequency is consirdered desirable by many designer as is the transient response in the range where the impedance of a reglator goes from inductive to capacitive ( where the impedance of the output is dominant.The circuit is an emitter follower inside the closed loop of an op amp configured for unity AC gain. DC gain is the output voltage divided by the reference voltage. The open loop impedance of the follower is divided by the open loop gain of the op amp. The AD 825 is nice because it has a constant open loop gain over most of the audio frequency range instead rolling off at 6 dB per octave in this range. A constant output impedance instead of an impedance that rises with frequency is considered desirable by many designer as is the transient response in the range where the impedance of a regulator goes from inductive to capacitive, where the impedance of the output is dominant. You could use cascaded close loop circuits (op amp or discrete or a combination of both) as the error amp as long as stability requirements are met with the loop closed.

 

[fdegrove]

HORSES FOR COURSES...AGAIN.

Hi,

A constant output impedance instead of an impedance that rises with frequency is consirdered desirable by many designer as is the transient response in the range where the impedance of a reglator goes from inductive to capacitive

Yes, yes and yes...

However, there are situations where this can be relaxed...
An RIAA stage is one example...
In short, it all depends where the reg starts to change behaviour and what you want it to do.

Having one that's wide band and has low Zout across the target frequency band hasn't hurt anything yet as far as I know...

Certainly, you can purposely design for less broad passband to keep the nasties out...

Fred may also agree with me here that the nasties should be designed out at the source and not as an afterthought downstream...

Unfortunately, we're all too often at the receiving end.

Cheers,

 

[transducer]

PS On the " Jung regulator - Didden PCB layout" note that the gnd plane was purposefully omitted over some of the more (capacitively) sensitive circuitry. At the time, it seemed to me the best way to avoid making possible oscillatory tendencies worse. Would I make a different layout now, with the hindsight of 8 years of experience and other opamps availble? Yes, in some details. Not in the general grounding and remote sense arrangement. And I would probably leave out the gnd plane, since almost all circuit nodes are low impedance enough to be insensitive to capacitive and RFI interference.

Jan Didden

Jan, are you saying that the layout as you designed it would work just as well without the ground plane, and therefore the board could be single sided? I'm in need of several of these regulators, it would be easy enough for me to copy the layout of the original and burn my own boards from the drawing in the original article. Eliminating the ground plane would make it much easier to do, especially since the ground plane was never published on its own, but with component placement on top of the drawing.

Are there any connections to be made through the groundplane? I can check to see, but maybe you have the info handy.

Thanks,
RonS

 

[scott wurcer]

A quick note on the Jung Super Regulator. The distortion neutralizing capacitor on the AD797 has the feature of reducing the output impedance over frequency. 0.000001 Ohm out to 100kHz was achieved, but this was left out of the articles as too esoteric. Moderators are welcome to spank me if they think I get out of line promoting my own products.

 

[jcarr]

As far as I know, Scott is the designer of the AD797, among other Analog Devices devices.

jonathan carr

 

[dimitri]

yes, jonathan, you are right - USP 5,166,637 "Distortion cancellation amplifier system"

 

[jackinnj]

United States Patent 5,166,637
Wurcer November 24, 1992

--------------------------------------------------------------------------------
Distortion cancellation amplifier system


Abstract
A distortion cancellation amplifier system operational amplifier system includes a current mirror circuit having an input, an output and a common terminal; a device for providing a pair of differential current signals to the input and output terminals of the current mirror circuit; a control device, responsive to said output terminal of the current mirror circuit, for controlling the voltage at the common terminal to drive the voltage at the input terminal of said current mirror circuit to track the voltage on the output terminal of the current mirror circuit; an output amplifier stage having a predetermined gain and having an input and an output terminal with its input terminal connected to the output terminal of the current mirror circuit; a gain control device having a predetermined impedance connected with the input terminal of the output amplifier; and a distortion suppression device connected between the output terminal of the output amplifier and the input terminal of the current mirror circuit and having an impedance equal to the predetermined impedance of the gain control device divided by the predetermined gain of the output amplifier for cancelling signal distortion introduced by the output amplifier stage.

 

[millwood]

scott seems to be quite recognized among those who know what they are talking about.

http://www.analog.com/library/analogDialogue/archives/30-2/editnotes.html

so his major work includes ad524, 712, and 797 which he eluded to in his last post. sounds like AD owes a lot to Scott.

I doubt any of us, even in our own field, has as much success as scott does.