B22/Active Ground Query
Sep 22, 2009 at 9:35 PM Post #106 of 204
Quote:

Originally Posted by n_maher /img/forum/go_quote.gif
I think it's because some of your posts read as if you are convinced that you are 100% correct and have no interest in seeing another side of the debate but rather are simply here to prove that you know more/better and are correcting other's false assumptions?


All I'm doing is simply stating things as I see them. Ultimately that's what we all do. And we state things as we see them with the assumption that they are correct until such time as evidence is presented to show otherwise.

I see no reason whatsoever why that should cause any sort of consternation on anyone's part.

Quote:

But what you're attacking (and it does come across that way, Steve)...


All I've done is disagree with certain claims made. That's not any sort of attack nor should it be taken as one.

Quote:

...is not something that was happened upon or decided without a great deal of thought or by people without a well formed understanding of the principles in play (amb, tangent, morsel, and a few others).


That's certainly possible, though up to this point, I haven't seen any convincing evidence of that.

There's no shortage of apocryphal information in audio (or in other pursuits for that matter), and often times claims are made that are accepted as fact without ever bothering to question or analyze them to see if they're in fact true and they just get perpetuated.

Perhaps that's the case here. I don't know.

Quote:

But it would seem to me that it should be a simple experiment to show the benefits or lack thereof. Would it not stand to reason that someone could build a 3ch version, test it as such and then simply disconnect the active ground channel from both the power and output and properly connect it to be a 2ch setup and retest?


With regard to the claim that with the active ground the load current doesn't go to ground, there's no need to build anything. It would be trivially easy to show it in a sim. Or even a sketch on the back of a napkin for that matter.

I've presented my argument that it does (indeed that it must according to Kirchhoff) go through the ground node. I've given my reasons why and even provided illustration of the current path.

If I'm incorrect, it should be a simple matter to show where I am in error and show where the load current does go if not through the ground node. I even provided a blank slate for someone to use for that purpose.

k
 
Sep 22, 2009 at 9:37 PM Post #107 of 204
Quote:

Originally Posted by nikongod /img/forum/go_quote.gif
I have nothing to add to this thread, because it is a dumb thread


frown.gif


Quote:

Originally Posted by n_maher /img/forum/go_quote.gif
But it would seem to me that it should be a simple experiment to show the benefits or lack thereof. Would it not stand to reason that someone could build a 3ch version, test it as such and then simply disconnect the active ground channel from both the power and output and properly connect it to be a 2ch setup and retest?


Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif
*something* is going on, since there are measurable results with and without 3rd channels.

are they audible? I'm not sure. but they do seem to be measurable.



It seems to me that Steve's initial point was that this hadn't really been done, that the comparative tests were not quite comparative. To wit,

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.




Quote:

Originally Posted by Pars /img/forum/go_quote.gif
It's an interesting discussion. Active ground is only bandied about here (AFAIK). Mention it somewhere like diyaudio for example and see what kind of response you get (waste of a good channel seems to be the consensus) . At least from what I have seen. None of KG's amps have active grounds as well, so there seem to be some differing lines of thought related to the subject.


Indeed, put me in the skeptical camp, if for no other reason that it seems like an awful lot of feedback which never seems like a good thing to me. But then, I don't believe in using "complementary pairs", either
wink.gif
 
Sep 22, 2009 at 9:52 PM Post #108 of 204
Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif
*something* is going on, since there are measurable results with and without 3rd channels.

are they audible? I'm not sure. but they do seem to be measurable.



I've not seen any measurements using the same B22 amplifier and where nothing has changed except whether or not the headphone ground went directly to ground, or to ground through the active ground.

Are you aware of any? If so, where can I find them?

Quote:

what is the cause? I actually don't know enough about 3ch designs; but I do believe that *something* useful is going on since it does get better specs.


Just for the record, I did say previously that it could have some benefit in an unbalanced amplifier that's using a single power supply.

But that benefit would only come from the ground channel's keeping the power supply current draw constant, which would prevent interchannel crosstalk via supply rail coupling.

And that wouldn't be an issue with dual supplies and as I've said before, the active ground won't work as it's supposed to in a dual supply situation.

What the present discussion is about is the claim that with the active ground, load currents don't go to ground.

I'm sorry, but until someone shows some pretty damn convincing evidence to the contrary, I'm considering that claim to be patently false.

Quote:

the only arguing I do about it is if I want to justify the space/cost/powering of the 3rd channel. but I do admit that things measure better with a 3rd channel.


Again, where can I find some apples/apples measurements of a B22 in this regard?

k
 
Sep 22, 2009 at 10:00 PM Post #109 of 204
Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif
It seems to me that Steve's initial point was that this hadn't really been done, that the comparative tests were not quite comparative. To wit,


Yes. And those were the only measurements I've seen.

Though I do concede that it can help reduce crosstalk in an unbalanced, single supply amp. But it wasn't an unbalanced, single supply amp that was under consideration when I started this thread.

k
 
Sep 22, 2009 at 10:25 PM Post #110 of 204
Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif
I've not seen any measurements using the same B22 amplifier and where nothing has changed except whether or not the headphone ground went directly to ground, or to ground through the active ground.

Are you aware of any? If so, where can I find them?



not sure where or who's site has this, but its not hard to run a test with a 3ch amp.
Quote:

Again, where can I find some apples/apples measurements of a B22 in this regard?


gimme a month or so (when my backplane arrives) and I'll have a 3ch b22 and I'll be able to disable the middle channel and re-do an rmaa test. that will be apples to apples. well, almost; I'll be using a pc for the tests (lol).

I could run this test today with my m3, though. I might try that, also, since its not too hard to do.

why not also do it yourself? a cheap 3ch amp can be bought for just a little cash and you'd be able to trust your own measurements, right?

even a pimeta would do the test. why not just GET one and try it? its really cheap, as boards go, and it would suffice for at least SOME independant data gathering.
 
Sep 22, 2009 at 10:47 PM Post #111 of 204
great!!!
beerchug.gif
please report your findings.
 
Sep 22, 2009 at 10:56 PM Post #112 of 204
Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif
why not also do it yourself? a cheap 3ch amp can be bought for just a little cash and you'd be able to trust your own measurements, right?


As I said, I can see that there could be some benefit in an unbalanced, single power supply amp. Just that the way you said it, I thought you might have been referring to some measurements I hadn't already seen and was curious to see them.

And as far as the claim about load current not going to ground goes, there's no need to build anything to "test" for that.

k
 
Sep 22, 2009 at 11:17 PM Post #113 of 204
I just read through this thread and there is some good information here. It's good that someone pointed out KVL and KCL.

Why doesn't someone do a Spice/Multisim simulation? That would be a quick and dirty way to do an apples/apples relative measurement.
 
Sep 22, 2009 at 11:36 PM Post #114 of 204
Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif
Though I do concede that it can help reduce crosstalk in an unbalanced, single supply amp.


Is that an apology regarding all the crap you were talking about the MMM? Surely not!
 
Sep 22, 2009 at 11:55 PM Post #115 of 204
Quote:

Originally Posted by Beefy /img/forum/go_quote.gif
Is that an apology regarding all the crap you were talking about the MMM? Surely not!


What I was talking about wasn't crap. And what I was talking about wasn't to do with keeping supply current constant, but rather the erroneous claim that the load current doesn't go to power supply ground with the active ground.

The only "crap" was your attempt to turn it into a personal confrontation rather than a respectful discussion.

k
 
Sep 23, 2009 at 12:01 AM Post #116 of 204
Quote:

Originally Posted by natrix /img/forum/go_quote.gif
Why doesn't someone do a Spice/Multisim simulation? That would be a quick and dirty way to do an apples/apples relative measurement.


Depends on exactly what it is you're wanting to measure.

If you're wanting to measure intrerchannel crosstalk, you'd need to first come up with some reasonable model for rail impedances.

But you wouldn't need to do any sort of simulation to get at the bottom of the load current not going to ground issue any more than you'd need to run a simulation to get at the bottom of how much current is flowing through a resistor of a given resistance with a given voltage across it.

k
 
Sep 23, 2009 at 12:10 AM Post #117 of 204
Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif
All I'm doing is simply stating things as I see them. Ultimately that's what we all do. And we state things as we see them with the assumption that they are correct until such time as evidence is presented to show otherwise.

I see no reason whatsoever why that should cause any sort of consternation on anyone's part.



For the love of Pete, I gave you my impression of how people may be interpreting your posts based on how I read them and you're proving my point. Rather than argue how I'm wrong in interpreting how you type words how about just simply accepting the fact that the way you communicate your point may not be universally neutral?

Quote:

All I've done is disagree with certain claims made. That's not any sort of attack nor should it be taken as one.


Again, you're arguing how I'm interpreting your posts. Let it go, absorb what I said, you don't need to reply or explain how I have it wrong.
 
Sep 23, 2009 at 12:11 AM Post #118 of 204
Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif
If you're wanting to measure intrerchannel crosstalk, you'd need to first come up with some reasonable model for rail impedances.


Is a rail, by itself, effectively a type of transmission line? That's just series inductance+resistance with capacitance+resistance to ground... right?
 
Sep 23, 2009 at 12:30 AM Post #119 of 204
Quote:

Originally Posted by Koyaan I. Sqatsi /img/forum/go_quote.gif
I've not seen any measurements using the same B22 amplifier and where nothing has changed except whether or not the headphone ground went directly to ground, or to ground through the active ground.

Are you aware of any? If so, where can I find them?



You've been commissioned to build a balanced B22 for iPodPJ, so you'll have enough boards to test.
 
Sep 23, 2009 at 12:52 AM Post #120 of 204
Quote:

Originally Posted by natrix /img/forum/go_quote.gif
Is a rail, by itself, effectively a type of transmission line? That's just series inductance+resistance with capacitance+resistance to ground... right?


I wouldn't call it a transmission line. It could be analyzed with lumped parameters.

You'd basically have a series resistance, series inductance, and shunt capacitance.

Though thinking about it, perhaps a simpler approach would be to simply measure the PSRR of the B22.

I suspect it should be pretty good as with regard to power supply issues, the B22's rather like those guys you see in the bomb suits.

The output stage is cascoded, everything upstream of the output stage is fed from a cap multiplier, and everything else is also cascoded.

This is why I said initially that the ground channel would likely have just a "marginal" benefit.

k
 

Users who are viewing this thread

Back
Top