Line mixer (active)

[NewYorkDave]

"It is rejection of signals from external sources that are results of poor construction."

This statement doesn't seem to make any sense. Care to elaborate? How can the builder of a console have any control over "external sources?"

"However, it increases internal noises about 6 dB up and doubles distortions... "

No, the real-world increase of noise and distortion is more on the order of 3dB. The amplitude of the desired signal is, however, increased by 6dB. The noise and distortion voltage would only double if these signals were correlated, which is very unlikely in real life.

"makes distortions symmetrical that may be good for measurements and spec reports, but bad for sonic qualities."

Any truly symmetrical distortion would be cancelled out in the differential circuit.

 

[Guest]

NewYorkDave, can you reveal the secret of generating anti-distortions and anti-noises?

 

[NewYorkDave]

That's cute, but it does nothing to clarify apparently inaccurate statements made in your last post--nor does it disprove anything I said in mine.

 

[Svart]

It's not rejection of external sources necessarily, it's more like rejection of sources within the console itself. Various clocks and currents radiating inside of this theoretically enclosed cabinet, which for this purpose is rejecting all outside sources but keeping others within!

cooling fans, plasma meters, lcds, LED switching, digital ground bouncing on rails right next to the audio path etc.

Careful design and layout are only part of the solution to the obvious problems. We will then have harmonics and harmonics of distortions that can be additive and subtractive and all of this can lead to a noise floor many dbs higher, especially when picked up by cables running back and forth inside of the console.

So we balance our signals for those problem areas, especially long traces and long cables that would normally act as antennas.

So to answer your question, there is NO sonic advantage aside from keeping the s/n ratio higher.

 

[Guest]

So. remembering that all wires have resistances and inductivities, and about capacitances between them, is the way to go.

 

[Svart]

well yes. I for one consider that part of a proper design but sometimes getting all that right doesn't completely "fix" an infiltration problem.

 

[Guest]

[quote author="NewYorkDave"]That's cute, but it does nothing to clarify apparently inaccurate statements made in your last post--nor does it disprove anything I said in mine.[/quote]

Do you enjoy listening the music you play?

 

[mikep]

[quote author="Wavebourn"]... Also, it makes distortions symmetrical that may be good for measurements and spec reports, but bad for sonic qualities.[/quote]

If this is an issue you could use a pseudo-balanced buss, where the cold side is not driven, it simply has termination resistors physically right next to the feed-on resistors. noise will cancel, f2 will not.

 

[Guest]

[quote author="mikep"][quote author="Wavebourn"]... Also, it makes distortions symmetrical that may be good for measurements and spec reports, but bad for sonic qualities.[/quote]

If this is an issue you could use a pseudo-balanced buss, where the cold side is not driven, it simply has termination resistors physically right next to the feed-on resistors. noise will cancel, f2 will not.[/quote]

If you balance impedances of wires and feed them to balanced input it is actually balanced design, not a pseudo-balanced. You loose power on additional resistors, you need more amplification to replenish it. Power loss is the price you pay in order to cancel external to the network noises. But internal noises and distortions raise up.

If you shield properly the single wire and avoid odd currents through the shield results will be better because you don't need to amplify lost on balancing power, and don't need symmetrical inputs that add distortions.

 

[sodderboy]

[quote author="Svart"]
cooling fans, plasma meters, lcds, LED switching, digital ground bouncing on rails right next to the audio path etc.

[/quote]
Whoa. I didn't see anything like that on Dave's block diagram. How can I fit all that ***** in here?!?

This is an IAR class built TLO74 based mikser (short for passive summing buss with active electronics n' stuff) circa 1987 built from Canal St. surplus. KISS- switchable line and instrument inputs, pan, fader, master fader. Construction wise, I have improved my wiring a bit. . .

Mike

 

[Guest]

[quote author="sodderboy"][quote author="Svart"]
cooling fans, plasma meters, lcds, LED switching, digital ground bouncing on rails right next to the audio path etc.

[/quote]
Whoa. I didn't see anything like that on Dave's block diagram. How can I fit all that ***** in here?!?

[/quote]

Mike, you may take a switched power supply, and connect it's ground wire to the ground wire right on the input socket. After that you will defenitely discover advantages of symmetrical design that excuses such stupid mistakes.

 

[NewYorkDave]

Whoa. I didn't see anything like that on Dave's block diagram.

Nor are there any balanced busses on my diagram, which makes it seem especially odd that Wavebourn has chosen to pick a fight with me over the merits of balanced transmission lines.

Do you enjoy listening the music you play?

I do not respond to audiophile mysticism. Talk about technical facts and then I'll respond.

If you balance impedances of wires and feed them to balanced input it is actually balanced design, not a pseudo-balanced.

No, it is not truly balanced. The source impedances may be balanced, but the signal voltages in the two lines are not symmetrical.

You loose power on additional resistors, you need more amplification to replenish it.

That is true when someone tries to create an impedance-balanced line after the fact by sticking an extra resistor in the low side return. It doubles the source impedance and also reduces the output level by some degree. But this is not the case in a properly-designed balanced interface, whether it's truly balanced or only impedance-balanced.

symmetrical inputs that add distortions.

You still haven't given any technical facts to support that assertion.

It seems that every year or so, some provocateur with a little bit of technical knowledge and a lot of ego shows up and starts picking fights with everybody. If that's the reason you're here--if you're this year's agitator--then please start a new thread for that. I'm sure most of us would prefer that this thread remain focused on the particular project at hand.

 

[Guest]

[quote author="NewYorkDave"]

symmetrical inputs that add distortions.

You still haven't given any technical facts to support that assertion.

[/quote]

The technical fact is obviuos: any active element adds audible distortions. Even some passive elements such as transformers add them. Period.

I can't put on stage SPICE simulators. I can't show to artists specs and opinions of people who consider that they know well about technical details. I have to play on people's perceptions using deep knowledge of electronics, psychology, and my own experience.

And it is not an audiofile ********, it is the opinion of Engineer Designed and Technologist of Radio and Electronics Equipment, something like American Master's degree in Electrical Engineering, also I have a certificate from Richard Bandler and John LaValle, "Design Human Engineer", it is about practical psychology. And the whole my life I experiment with different kind of audio distortions in order to get needed imaginations.

 

[NewYorkDave]

Did it not occur to you that you don't necessarily have to add an amplifier, a transformer or any other component to implement a balanced input? If you're starting out with a differential amplifier, for instance an op-amp, the choice is yours whether you're to sense the difference between a single, unbalanced signal line and ground or between the two sides of a balanced line.

I can't put on stage SPICE simulators.

I don't design with SPICE, either. I use paper, pencil, calculator, breadboards, test equipment and ears. So if you're looking to spar with a SPICE jockey, I'm not your man.

And it is not an audiofile ********, it is the opinion of Engineer Designed and Technologist of Radio and Electronics Equipment

And I'm the King of Prussia. What's your point? That your opinions must be correct because of a title or degree? Someone with a "deep understanding of electronics" should be able to explain the physical facts behind his assertions without having to resort to "it is, because I said so."

 

[Guest]

My opinion is, OpAmp is not just a perfect black box that allows to build what do we want assuming only feedback and input resistors. It is the complex beast that has own character. The more of such beasts you use, the more is cumulative impact on audio quality, such bad as increased level of distortions on higher frequencies and on low powers that is against our experience of perceiving natural sounds. Also, we always amplify power, even when we want "purely voltage amplification". If we loose a power, we need more power gain. It always mean less deep feedback, or more stages. Both add noises and/or distortions. Feedback is such a weapon so if you use it you have to use as more as possible for less distortions, otherwise don't use it.

Also, by my experience usage of non-inverting input of opamps must be limited by cases when it is absolutely necessary. I mean audio amplifying, not just college books and datasheets with examples. Distortions may be not audible when you use one non-inverting amplifier, but when you use a lot in the console the result is accumulated.

 

[NewYorkDave]

OK, now we're getting somewhere.

OpAmp is not just a perfect black box that allows to build what do we want assuming only feedback and input resistors.

True, at least until they create an op-amp with inifinite open-lop gain. And that's not going to happen.

The more of such beasts you use, the more is cumulative impact on audio quality

Right, which is why it made me cringe when posters in this thread started asking for a bunch of additional features. That defeats the whole purpose of the project, which was to accomplish the task of mixing some line-level signals cleanly with the fewest amplifiers needed to do the job.

Also, we always amplify power, even when we want "purely voltage amplification".

Swinging a voltage across anything less than an infinite impedance always requires some amount of power. The relative importance of voltage or power transfer depends on the type of interface that has been chosen.

It always mean less deep feedback, or more stages. Both add noises and/or distortions. Feedback is such a weapon so if you use it you have to use as more as possible for less distortions, otherwise don't use it.

I agree. Many have fallen into the trap of "minimal feedback = better", while ignoring the fact that minimal feedback will not be effective in cancelling the distortions generated by its use (which tend to be of higher order and therefore are more audibly objectionable than the distortion of the "naked" amplifier itself, if that amplifier was reasonably linear to begin with).

So two stages, using more feedback, may result in fewer objectionable distortion products than a single stage using less feedback. Or, one might opt for a single stage that is sufficiently linear with no feedback applied. That's why the 6SN7 has always been such a popular tube; in a well-designed circuit, it can be astoundingly linear even run open-loop.

 

[Guest]

So, KISS principle in any art is vital... Audio art is not the exception.

 

[sodderboy]

[quote author="Wavebourn"]
Mike, you may take a switched power supply, and connect it's ground wire to the ground wire right on the input socket. After that you will defenitely discover advantages of symmetrical design that excuses such stupid mistakes.[/quote]

I have been wrangling with switching supplies for almost 20 years. My answer to their cost effective addition to the studio has been. . .

Iron.

I can show you a MIDI room with 64 synth outputs, most with dubious power supplies and matter-of-factual racking, and da' jernt is QUIET. I would show the users how much distortion is inherent in the Rapco DI transformers used to normall them to the SSL, but they are too busy making, uh, making. . . music. That's it! They are happily making music every day.
Sometimes it is necessary to put the nit-pick comb away: "sodderboy, why is the sky blue?" "Because it makes your eyes even more beautiful!"
you all know what I mean about how a tech can kill the moment
Mike

 

[Guest]

Symmetrical ins/outs of the gear is the different story, however. I think, Dave may use them to communicate to the outside world with his mixer, getting real advantages.

 

[PRR]

> what is advantage of balanced amps in terms of the target, I mean sonic qualities?

When the customer wants 96 channels with pre, post, and extra inserts, plasma meters, and all different signals going all different ways.... "Ground" is full of junk.

Even if 50 channels have input but all the sliders are slammed to zero: that is 5K/50= about 100 ohms from a lot of signal into your "ground".

Of course ground is never ground, and never silent. But in large consoles you have "garbage" which exceeds the distortion of any chip, and is not related to the signal in that channel, but is the sum of many signals.

That is before you get to class AB stages dumping garbage into their power rails and thus through rail caps into "ground".

Steve Dove's essay explains how bad it can get.

Above some size or complexity, console designers usually have to avoid ground referencing across the console (they may get away with ground reference inside the channel strip). Even though it means at least doubling complexity and increased/changed distortion.

Some projects "need" a massive mixer. And many musicians (and most listeners) do not have Golden Ears, can't hear the number of chips between original microphone and final car speaker.

And some folks can hear a difference. Which is why there are a few very good Large Consoles, and why there is interest in small "K.I.S.S." tools like NYDave's plan.

And sometimes you can use a rat-nest plastic-box mixer, like sodderboy, and make nice music. I've done projects like that, and even used some commercial stuff that was not even that neat.

You have a point about using inverting stages. They do not dump feedback current into "ground". And of course they do not excite the common-mode distortions in many input stages. They do tend to have higher-value resistors and thus higher voltage noise. Everything is a compromise.

It's like building a house. If you build a 2-story house, you just run fresh water and waste water pipes up and down. If you build a 100-story building, you can't push fresh water up from the bottom with normal pressure, waste water falls from the top with pipe-breaking force, and you have to use much more complicated pipe systems.

NYDave's goal here is a little house, which is why he isn't using heavy iron pipes or balanced mixing. If Svart wants to add plasma meters, digital ground bouncing, and a coffee roaster, he'll have to work out his own plan. If someone thinks that even a few op-amps is too many, there are passive mixer plans around.