[amb]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif Yes, I read that yesterday. But I don't see that it has any particular benefits if proper grounding practices are exercised.

A topology that keeps rail current constant isn't a difficult concept to grasp and I presented plenty of graphs to illustrate it. It has nothing to do with grounding practices.

Quote:

Though I didn't find the measurements terribly informative. Primarily because they were made using two physically different amplifiers. There's no way of knowing exactly what the ground scheme was in each amp (beyond just the issue of using an active ground or not) and grounding isn't the only thing which can contribute to crosstalk.

The 2-channel β18 is a single-PCB amp where the power supply, bulk caps and ground are all on a single board, connected to a large plane. Yes, it is physically different than the M³, but I've observed similar differences while experimenting with other circuits, and I do know a thing or two about proper grounding practices.

I didn't invent the 3-channel active ground topology. I have to give Morsel and Phil Larocco credit for that, who championed it in the PPA amplifier, as well as Larocco's commercial designs. Tangent also uses it in the Pimeta. I was convinced of its benefits after all the testing I've done.

If you're not convinced, that's fine by me. No one is forcing you to build a 3-channel amp, and β22 could be built in many configurations including the conventional 2-channel with passive ground.

 

[n_maher]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif By the way, I just noticed the Resistorous Conflagorous in your particulars. What's the story behind that? What, did you re-enact the firebombing of Dresden in one of your amps or something? k

Not to completely derail this thread but the resistorous comes from this thread.

 

[Steve Eddy]

Quote:

Originally Posted by amb /img/forum/go_quote.gif A topology that keeps rail current constant isn't a difficult concept to grasp and I presented plenty of graphs to illustrate it. It has nothing to do with grounding practices.

I grasp the concept just fine.

And I don't understand why you say it has nothing to do with grounding practices.

Virtually everything I see written about the "active ground" relates it to ground issues.

From the PIMETA page over at TangentSoft:

3-channel setup so return currents from the headphones don’t disturb the signal ground.

From the Technical Highlights section of the B22 on the AMB website:

The ground channel amplifier sources or sinks the return current from the transducers, which would otherwise have been dumped into signal ground or power supply ground. This shifts responsibility for the high current reactive load of the headphones from signal ground to the tightly regulated power supply rails, thus removing the primary source of signal ground contamination.

From the post that digger referenced previously:

This makes sense because the headphone load's return current doesn't contaminate signal ground. If the signal ground, which is the zero-volt reference for both channels, is allowed to "wiggle" as a result of return current from one channel, that would result in an effective leakage of signal into the opposite channel.

and

Real headphone loads are reactive rather than purely resistive, which means that the return "ground" current can become phase shifted compared to the swinging voltage. If that current is allowed to pollute signal ground (which might be the case in a passive-ground 2-channel setup), it would induce a voltage wiggle that is also phase shifted from the main signal voltage, when the two are summed together you get a form of dynamic distortion.

The "signal ground" being disturbed, contaminated, allowed to wiggle, etc. due to load currents depends on grounding practices.

And if the raison d'etre of the active ground is to keep power supply current constant, then how could anyone recommend, as you did, using two power supplies in a five channel situation with one power supply for the four balanced boards and the other power supply for the ground channel, or three power supplies where one supplies the left hot/cold channels, the other the right hot/cold channels, and the third supplies the ground channel?

In both of these examples, the ground channel would do absolutely nothing with regard to keeping power supply current constant. In order for a ground channel to do that, it would have to be powered by the same supply that powers the channel or channels that it's "grounding."

So what advantage would the ground channel serve here?

k

 

[FallenAngel]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif And if the raison d'etre of the active ground is to keep power supply current constant, then how could anyone recommend, as you did, using two power supplies in a five channel situation with one power supply for the four balanced boards and the other power supply for the ground channel, or three power supplies where one supplies the left hot/cold channels, the other the right hot/cold channels, and the third supplies the ground channel? In both of these examples, the ground channel would do absolutely nothing with regard to keeping power supply current constant. In order for a ground channel to do that, it would have to be powered by the same supply that powers the channel or channels that it's "grounding." So what advantage would the ground channel serve here? k

So you're basically suggesting a completely separate power supply for ground instead of a separate amplifier for ground. I guess that can work and will lower stereo crosstalk and if it's a good power supply, it'll be able to sink current quickly and easily, but I'd say that an amplifier does a better job at driving (or in this case sinking) a headphone load over a power supply, especially one in Class-A and you still haven't addressed the impedance point. I'm pretty sure the ground channel of the amp helps with stereo crosstalk more than the PSU because it isolates headphone driving ground from input signal ground. Not to mention it's cheaper to build a Beta22 board than another Sigma22.

 

[Steve Eddy]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif Not to completely derail this thread but the resistorous comes from this thread.

Thanks, Nate.

Wow, what a rollercoaster of a ride!

k

 

[Beefy]

Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif So you're basically suggesting......

I think the suggestion is that a 'proper' ground scheme automagically cures all ills, and that such a scheme could never be polluted by return current from the headphones.

I really don't see that this could be true. The current has to go somewhere, is transient in nature, and I think that AMB has provided more than substantial arguments and evidence to support his position that buffering it through an extra amplifier channel has benefits.

 

[Steve Eddy]

Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif So you're basically suggesting a completely separate power supply for ground instead of a separate amplifier for ground.

If the "you're" here is referring to me, I wasn't suggesting anything. I was questioning how, if the purpose of the active ground is to maintain constant power supply current, anyone could suggest using an active ground where the active ground channel is powered by a separate power supply from the channel it's supposed to be "grounding."

This would leave you with the variable supply current that you would have without the active ground, i.e. just using a passive ground.

k

 

[amb]

There are two different issues I'm addressing here that you are mixing together.

1. Signal ground pollution removal
2. Rail current cancellation

#1 is the primary benefit of an active output ground. An actively driven high current amplifier with wide bandwidth and low output impedance is going to be better at sourceing/sinking current than a passive PSU ground. The latter's impedance is dominated by the bulk capacitor's ESR and wires. As long as there is ground impedance (which I hope you won't dispute, even the best star ground scheme has some) then there will be ground potential wiggle when a varying current flows through it. Basic Ohm's Law. Divert that current away from ground and you've eliminated that wiggle.

#2 is a synergistic benefit of active ground when all channels are operating in class A. I've posted the link to the thread with my findings with plenty of graphs to show the rail current cancellation (so the result is a constant zero summed current), and I won't rehash it. This has nothing to do with ground, per se. But you don't get this rail cancellation if there is no ground channel amplifier providing the opposing current to cancel with. Obviously, to get this benefit you need to use the same power supply for all three channels. Nowhere did I state otherwise.

 

[Steve Eddy]

Quote:

Originally Posted by Beefy /img/forum/go_quote.gif I think the suggestion is that a 'proper' ground scheme automagically cures all ills, and that such a scheme could never be polluted by return current from the headphones. I really don't see that this could be true. The current has to go somewhere...

Sure, it has to go somewhere.

And even with an active ground, it ultimately goes to the same place it does with a passive ground. And that is, to ground.

What the active ground does, providing it's being powered by the same power supply that's powering the channel it's grounding, is make the power supply current constant.

If the power supply current is constant, then even if your grounding scheme isn't quite up to snuff, you're not going to pick up anything from the other channel via IR drops across any of the resistances in the ground scheme.

Quote:

...is transient in nature, and I think that AMB has provided more than substantial arguments and evidence to support his position that buffering it through an extra amplifier channel has benefits.

It ain't exactly being buffered. Again, what it's doing is keeping supply current constant. It's essentially a push-pull compliment to the push-pull output of the amplifier.

k

 

[amb]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif Sure, it has to go somewhere. And even with an active ground, it ultimately goes to the same place it does with a passive ground. And that is, to ground.

In an active ground amp, the load current doesn't go to ground. It's sourced from one amplifier's positive rail and sinked to another amplifier's negative rail (and vice versa).

 

[digger945]

FWIW, in my balanced amp(not a Beta 22) with Ti's Balanced/Unbalanced Source Switch used to ground the inputs of two amp boards(all boards have a gain of 11) for balanced headphone listening with an unbalanced source(in this case the Gamma 2), the audible difference to my ears is not subtle. No need for measurements or graphs to hear the improvement of active grounding.
I don't think it has been noted that the active ground amp for the Beta 22 has a different gain setting than the "hot" amps do. Recommended configuration is a gain of 1.

 

[Steve Eddy]

Quote:

Originally Posted by amb /img/forum/go_quote.gif #1 is the primary benefit of an active output ground. An actively driven high current amplifier with wide bandwidth and low output impedance is going to be better at sourceing/sinking current than a passive PSU ground. The latter's impedance is dominated by the bulk capacitor's ESR and wires. As long as there is ground impedance (which I hope you won't dispute, even the best star ground scheme has some) then there will be ground potential wiggle when a varying current flows through it. Basic Ohm's Law. Divert that current away from ground and you've eliminated that wiggle.

But you're not diverting current away from ground.

All of the current flows through the ground node whether you're using an active ground or a passive ground. An active ground makes absolutely no difference whatsoever in this regard.

The only difference is that instead of the "return" current going directly to the ground node via a conductor from the opposite side of the driver, it instead goes to the ground node via the output stage of the ground amp, the supply rail, any voltage regulators that may be used, and finally the power supply caps.

If you think current is somehow diverted away from ground, then where exactly is it going? Is it dribbling out onto the floor? Is it evaporating into the air?

Where is it going if it's not going to ground?

Quote:

#2 is a synergistic benefit of active ground when all channels are operating in class A. I've posted the link to the thread with my findings with plenty of graphs to show the rail current cancellation (so the result is a constant zero summed current), and I won't rehash it.

Yes, as I said, I already read that.

Quote:

This has nothing to do with ground, per se.

Well, actually it does in a sense. With a constant supply current, you can't couple any signal from one channel to another due to IR drops and poor grounding.

Quote:

But you don't get this rail cancellation if there is no ground channel amplifier providing the opposing current to cancel with.

Well, you wouldn't get that constant supply current with something like the B22. But you can get it by other means without a separate ground channel amplifier.

Quote:

Obviously, to get this benefit you need to use the same power supply for all three channels. Nowhere did I state otherwise.

Then why did you suggest a five board solution using two power supplies, one for the four balanced boards and one for the one ground channel? Or three power supplies, two for the left right balanced boards, and one for the ground channel?

Since the ground channel doesn't divert any current away from ground, and since it can't provide a constant supply current if it's being powered from a separate supply, then what function does the ground channel serve?

k

 

[Steve Eddy]

Quote:

Originally Posted by amb /img/forum/go_quote.gif In an active ground amp, the load current doesn't go to ground. It's sourced from one amplifier's positive rail and sinked to another amplifier's negative rail (and vice versa).

And from the amplifier's negative rail to where exactly?

k

 

[cfcubed]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif Hmmm, I'd swear that either amb or someone had done comparisons of a 2ch vs. 3ch beta22. But I can't find them now either, crud. And if they weren't for the beta22 maybe they were in the M³ section?

Assuming you mean measurements comparisons, perhaps this thread?
Knew it wasn't just déjà vu happening for me (again)

.

 

[amb]

Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif ... If you think current is somehow diverted away from ground, then where exactly is it going? Is it dribbling out onto the floor? Is it evaporating into the air? Where is it going if it's not going to ground?

I'll just cite the M³ (and PPA, Pimeta, etc) as an example of why that current does NOT go to ground because it's easier to illustrate. In these amps the power supply is "single rail" and split into two to get the positive and negative rails. The ground is virtual, created by the TLE2426 rail splitter chip. That chip could only source/sink 20mA. Yet the M³ could deliver amperes of current through the load. Don't tell me all that current is "going" back through the poor little TLE2426!

Quote:

Then why did you suggest a five board solution using two power supplies, one for the four balanced boards and one for the one ground channel? Or three power supplies, two for the left right balanced boards, and one for the ground channel?

If you're referring to a private message I had with someone, it's only because two doesn't divide into an odd number of amplifier boards very well, and it's too much to ask one σ22 PSU board to power five β22 channels.