How "super" is a Jung Super Regulator (or similar)?

[rock soderstrom]

The super regulator has excellent PSSR properties and a very low noise if you choose the right opamp but you could also use LT3045 which has similar properties although a bit slower and sadly a very un-DIY friendly physical size.

Here you have a LT3045 based power supply.
https://www.mpaudio.net/product-page/sd-hpuln-ps

Thanks for posting this!

 

[peranders]

Is that important? As I said, I like sloppy rails..

Personally I had those ideas before but now when you have extremely advanced (and small) voltage regulators I think they can actually be better. I believe in a rock stable supply voltage,

These "super-duper" regs, they just propagate disturbances back to the PSU, instead of keeping them confined to a controllable area.

Can you elaborate this?

 

[JohnRoberts]

Personally I had those ideas before but now when you have extremely advanced (and small) voltage regulators I think they can actually be better. I believe in a rock stable supply voltage,
Can you elaborate this?

Try designing consoles with several feet (meters) long power supply rails that will never be very low impedance (Ignore the pictures of heavy copper bars).

Local decoupling can keep the supplies stiff locally... but I have seen strange interactions with tens of regulated strips inside a single system.

JR 

 

[abbey road d enfer]

Personally I had those ideas before but now when you have extremely advanced (and small) voltage regulators I think they can actually be better. I believe in a rock stable supply voltage,

We try as much as we can here, to deal with science, not beliefs.

These "super-duper" regs, they just propagate disturbances back to the PSU, instead of keeping them confined to a controllable area.
Can you elaborate this?

A sudden current demand at the output of the reg reflects directly on the input, which in turn propagates to the PSU, and to all teh nodes that are connected to the rails.
Decoupling capacitors, by containing the disturbance very close to the stage that produces it reduces its propagation. If the feed is taken via a resistor, it is even more contained.

 

[peranders]

If you want to take advantage of the super regulators then you must have very short wires. We talk about centimeters.

 

[abbey road d enfer]

If you want to take advantage of the super regulators then you must have very short wires. We talk about centimeters.

Which wires?
Is it between reg and load? I agree. That's part of "keeping them (disturbances) confined to a controllable area".
Is it between PSU and reg? It's not always feasible. And should not matter IMO.
But I don't see how a super reg improves performance significantly over well-designed decoupling.
And at system level, all the ground currents that are returned to the PSU create longitudinal noise.
IMO super regs must be associated with isolated supplies, either with isolated DC/DC converters or with resistors, that need to be complemented with strong decoupling caps.
All the claims made to super regs superiority are made in the context of a stand-alone stage.
Remember this thread started with someone who wants to improve a mixer, where a number of channels draw their power from a single distant PSU.

 

[rock soderstrom]

So my question is, why are these types of regulators so rarely used in ProAudio circuits?

Wouldn't it be an advantage to use such super regulators in a mixer, for example, as a bipolar voltage source for the operational amplifiers?

Couldn't a 70dB microphone amplifier noticeably benefit from the super low noise values ​​or the very good PSRR values of these regulators?

Well, the mixer was just an example, maybe not the best. I was and I am still interested in whether the use of these super regulators makes sense or not. What about DA converters, mic preamps or mastering compressors or EQs?

I think the thread has already shown a lot of interesting aspects of this topic, thanks for that. I look forward to whatever comes next, Groupdiy.com is one of the best places for such discussion because of the knowledge available and the (mostly 8)) unagitated way it is discussed.

 

[JohnRoberts]

Well, the mixer was just an example, maybe not the best. I was and I am still interested in whether the use of these super regulators makes sense or not. What about DA converters, mic preamps or mastering compressors or EQs?

I think the thread has already shown a lot of interesting aspects of this topic, thanks for that. I look forward to whatever comes next, Groupdiy.com is one of the best places for such discussion because of the knowledge available and the (mostly 8)) unagitated way it is discussed.

I am probably repeating what I or others have already said, not that there are many experienced console designers posting here.

DC precision is not important for console regulators.

The last "big" console (more than 50 strips) I did used 3-terminal regulators on every strip... that came back to bite me when a batch of unusually noisy (shot noise) regulators caused audible noise in the master bus noise floor. When Motorola declined to guarantee a max noise spec for their regulators I black listed them (at the time Motorola was very proud of their six sigma quality performance, but they apparently attained that by not rejecting noisy regulators.)

Sorry about the veer... Good practice for console design is to use modest build out resistors at each strip with electrolytic reservoir capacitors. Also popular to use flame proof resistors for these build outs that will serve as cheap fuses preventing a PS short on a strip from taking down the entire console. 

Modern electronics has such good power supply rejection that you could probably get away with unregulated rails for most line level circuitry, but no need to with cheap regulation and decent sized rail capacitors.

JR

 

[abbey road d enfer]

Well, the mixer was just an example, maybe not the best. I was and I am still interested in whether the use of these super regulators makes sense or not.

IMO it doesn't make much sense, just "some".
A full answer would require analyzing many aspects. I'll try to briefly mention a few of them.
First, why do we use regulators for audio circuits? The main reason is because they clean the PSU voltage more efficiently than a huge batch of capacitors, actually replacing the choke that used to be in valve gear.
In logic circuits, a regulated voltage is essential to the operation, but note that microprocessors and DSP do not rely on super regs, they rely on a lot of capacitors strategically positioned in order to minimize track length, return currents to the load, not to the supply, and and avoid cross-contamination of circuits.

Now, as I mentioned earlier, these super regs, if they were operated as to provide the benefits their supporters claim, would displace the problem to the main supply, or require individual floating supplies. Actually, I don't think any of these super regs can be operated without local decoupling.
Some of these super regs use Kelvin sensing to achieve their best performance. How does this fit with the need to interconnect circuits?
I don't doubt that these regs have outstanding specs but in practice, what is the benefit? I would like to see one-to-one comparisons of a same circuit with a standard well-designed supply and one wit these super regs. So far, I've only read about subjective comments.
I may sound very negative, but it's not because I don't advocate technical improvements, it's because I've read and heard too many things that make it smell like snake oil. No wonder the main (only?) market for them is audiophiles.
BTW I have no doubt a badly designed circuit would perform better with super regs than with ordinary regs (particularly the 78/79xx).

 

[dbelousov]

I built these regs with some minimal modifications, adapting them for the LM4040 in the SOT-23 package was the biggest. They power my single channel "reference" mic preamp.

They are really good, indeed, but usage in pro-audio is very limited for them. They make sense only if they can sense. With short wires. Only one board.

And they are absolutely redundant.

My pre is decoupled heavily, 10 Ohm + 470 uF x 2 at the inputs, 0.1 ceramic + 100 uF lytic x 2 for every op-amp + other local caps, so I don't think the regs are even bothered to do something extraordinary.

There are respected men, who are (or were) professional audio gear designers for decades, but for me, it is just a hobby that went too far, my knowledge and experience are close to none-exitance in comparison. Thus, everybody has to take my opinion with a grain of salt.

I think these regs can be beneficial for some discrete designs, but for contemporary IC-s that "cooked right" I doubt it. My ADC has THD 0.00013% and sounds crystal-clear. It's powered by an SMPS. My desk sounds transparent and crystal-clear. It's powered with a 2-meter long cable.

In my (limited) experience with regulators, the 78/79 are inferior, they are noisy and can't even provide stable output, there is always some annoying drifting.  But they are very cheap and "good enough" for many applications, though. 317/337 are fine, and it's possible to achieve a very low noise level and virtually no ripple with them. Jung's regs are very good, but they are just an overkill for most designs.

My (limited) experience screams, that board routing and local decoupling is the most important part. Everything becomes stable, clear, and silent if it is done right.

I have a couple of buffers, powered by 78/79, heavily decoupled. I can switch between them and the direct signal from my DAC with my monitor controller. Nobody can hear any difference with contemporary uber op-amps involved. So what the purpose of better regs in this case? To make an inaudible difference more... inaudible?

In my opinion, 317/337 are the best option for DIY. They are low noise, stable, and, the most important part, adjustable. It's convenient to have only one board that can be used in different projects. I'm gonna build a 2-channel version of my "reference" pre and it will be powered by a 317/337 PSU.

BTW there are fast transient response optimized and low noise regs:
https://www.analog.com/ru/products/lt1963a.html
and its negative counterpart
https://www.analog.com/media/en/technical-documentation/data-sheets/3015fb.pdf

3A versions are available as well, it's enough for a medium-sized console.

 

[abbey road d enfer]

My (limited) experience screams, that board routing and local decoupling is the most important part. Everything becomes stable, clear, and silent if it is done right.

Don't be too modest. This is the opinion of many an expert.

 

[rock soderstrom]

Thank you all for all the very useful info!

 

[dbelousov]

Don't be too modest. This is the opinion of many an expert.

But I'm not an expert in this area. I'm an audio engineer, what I know about audio electronics is a collateral damage.

BTW, I've opened my primary "mixing" ADC, which is made by a local company. It is considered a "mastering grade" device among Russian audio engineers, lots of Russian audiophiles have it as well. They test it against Myteks, Lavrys, etc., constantly with good results.  Its primary designer is known for his "no BS" approach, he will never spend a dime for anything unimportant, no matter what "marketing potential" it has.

So, there are a top of the line DAC IC, pricey ultrafast opamps, digital I/O ICs, some descrete parts, four Dale through-hole resistors. And... LM317, LM337 and LM7805. And huge through-hole ferrite beans that connect the PSU outputs to the "main" part of the board. The circuit itself relies on SMD ceramic and TH lytic caps, placed locally.