[digger945]

Quote:

edit: looks like the output resistance of the amp is related to the gain. hmm.
 

...and gain is controlled by? (in this case using an IC opamp)
 

 

[Shike]

Quote:

the output impedance isn't shared...

I'm just going to focus on this right now.  The active grounds output impedance can't be fixed using feedback, so it gets shared between the L/R channels forming the voltage divider.
 
NwAvGuy goes into a lot more depth about this on ABI, and will probably explain it infinitely better than I could hope.
 

 

[Costia]

I thought you were talking about the output impedance of the op amps that actually do the amplifying not the one used as a buffer for the active ground - which is shared
I did read nwavguy's post on active grounds. not really convinced about the degradation. with a gain of x1 most op-amps should cope fairly well. (i.e. measurable but not audible difference)
But it does seems like a waste of parts\space when it is not necessary (such as a single battery)
I guess when you are talking about a required impedance of a few mOhm to get the 90db separation, the ground channel buffer's resistance will be the dominant one.
 
I find it quite strange that TI don't have output impedance in the sheets of opa132 and opa552/1 although their op-amp guide states that it could be significant
And i actually didn't know that an op amp's output impedance depends on the closed loop gain. learned something new  

 

[fubar3]

Just looked at the schematic. Interesting design but I am wondering if it is good to have a diode between the regulator and the load. I don't know.. just asking.  It is possible to improve the power supply using adjustable regulators, instead of fixed which tend to be imbalanced. Adjustable regulators are flexible since they allow slow turn-on and other tweaks.  A resistor between the large caps on the input side might be helpful since it becomes an R/C filter.  Of course, I have forgotten the design details.. just relying on intuition.

 

[Willakan]

NwAvGuy said that the power supply regulators are not really a "bottleneck" in the design, so better ones wouldn't offer any real measurable improvements. It also already incorporates input filtering, although the AC filtering cap was removed.
 

 

[anetode]

After reading up a bit more on the design and the background debates, I've noticed an interesting parallel. The power output criticism of nwavguy's O2 amp kind of mirrors his own concerns over the Mini3's power output discrepancies among different test setups. Whereas amb's specs were the culmination of a long design cycle and the optimum setup for the greatest measured performance, nwavguy's methodology aims at the average user and uses average componentry. By amb's exacting standards, the O2 has a very low usable gain range that the predetermined (though since revised) figures would have pushed to clipping. Contrarily, by nwavguy's standards we have an amp that would work across the greatest practical range of headphone impedances and sensitivities to very low distortion, but sacrificing the usable output to meet a very exacting crosstalk spec. A kind of reverse situation from when nwavguy challenged the mini3's crosstalk due to the OPA690 3-channel output decision. So, from an outside perspective it seems like a fanatical resolve was placed to designing to a specific ideal. Personally, I don't quite see how a possible difference from inaudible but otherwise unexemplary crosstalk to an equally inaudible but altogether impressive crosstalk justifies the inherent limitation in design choices. Sure, O2 is a concept amp built on achieving great measurements, but also a budget and wide usability. Both amplifiers appear to be the product of carefully weighed design consideration.
 
Unfortunately, what with the unfortunate subjective/objective rift in the head-fi camp and nwavguy's holy crusade and the designer egos, it almost seems like these approaches are somehow adversarial, rather than complementary.

 

[SpaceTimeMorph]

 
Quote:

After reading up a bit more on the design and the background debates, I've noticed an interesting parallel. The power output criticism of nwavguy's O2 amp kind of mirrors his own concerns over the Mini3's power output discrepancies among different test setups. Whereas amb's specs were the culmination of a long design cycle and the optimum setup for the greatest measured performance, nwavguy's methodology aims at the average user and uses average componentry. By amb's exacting standards, the O2 has a very low usable gain range that the predetermined (though since revised) figures would have pushed to clipping. Contrarily, by nwavguy's standards we have an amp that would work across the greatest practical range of headphone impedances and sensitivities to very low distortion, but sacrificing the usable output to meet a very exacting crosstalk spec. A kind of reverse situation from when nwavguy challenged the mini3's crosstalk due to the OPA690 3-channel output decision. So, from an outside perspective it seems like a fanatical resolve was placed to designing to a specific ideal. Personally, I don't quite see how a possible difference from inaudible but otherwise unexemplary crosstalk to an equally inaudible but altogether impressive crosstalk justifies the inherent limitation in design choices. Sure, O2 is a concept amp built on achieving great measurements, but also a budget and wide usability. Both amplifiers appear to be the product of carefully weighed design consideration.
 
Unfortunately, what with the unfortunate subjective/objective rift in the head-fi camp and nwavguy's holy crusade and the designer egos, it almost seems like these approaches are somehow adversarial, rather than complementary.

Do we really have to go here again?  A lot has been written on the input overload "issue" already.  I recommend you go over this thread again, some pertinent information is pretty recent with regards to what you are talking about.  To summarize, however: the only way in which you will push this to clipping is with a "hot" source driving too high of a gain or if you are driving a rather extreme set of headphones you may have to lower the gain so that you will have lower volume on certain recordings.  Both of these are rather extreme examples and hardly limitations of the design that you will be encountering every time you power the amp up.  There's a difference between this amp's power output capabilities on AC vs. on battery... make sure you understand that as well as it bears importantly to this topic.  Any other questions, feel free to ask them.  Also, NwAvGuy has a good summary article that talks about this issue as well as some other things, I recommend reading that.
 
As for the comparisons between NwAvGuy's and AMB's "philosophy" on design...  Frankly, that isn't the point of this discussion; this is a thread about the O2 amp.  However, if you want to know more, check this thread out before you criticize NwAvGuy too much:
http://www.diyaudio.com/forums/headphone-systems/188983-3-channel-headphone-amps-virtual-grounds-5.html#post2572899
 
The fact is, there has been some debate over the validity of AMB's crosstalk figure, and there have been independent measurements taken that show it is within an audible range(~-33dB to ~-46dB), so make of that what you will.

 

[anetode]

Quote:

 

Do we really have to go here again?  A lot has been written on the input overload "issue" already.  I recommend you go over this thread again, some pertinent information is pretty recent with regards to what you are talking about.  To summarize, however: the only way in which you will push this to clipping is with a "hot" source driving too high of a gain or if you are driving a rather extreme set of headphones you may have to lower the gain so that you will have lower volume on certain recordings.  Both of these are rather extreme examples and hardly limitations of the design that you will be encountering every time you power the amp up.  There's a difference between this amp's power output capabilities on AC vs. on battery... make sure you understand that as well as it bears importantly to this topic.  Any other questions, feel free to ask them.  Also, NwAvGuy has a good summary article that talks about this issue as well as some other things, I recommend reading that.
 
As for the comparisons between NwAvGuy's and AMB's "philosophy" on design...  Frankly, that isn't the point of this discussion; this is a thread about the O2 amp.  However, if you want to know more, check this thread out before you criticize NwAvGuy too much:
http://www.diyaudio.com/forums/headphone-systems/188983-3-channel-headphone-amps-virtual-grounds-5.html#post2572899
 
The fact is, there has been some debate over the validity of AMB's crosstalk figure, and there have been independent measurements taken that show it is within an audible range(~-33dB to ~-46dB), so make of that what you will.

Quite. As I said, I just spent a couple of hours reading up on it. I meant to offer a reflection on some of the controversy behind the amp's development which was relevant to the prior contents of this thread, regardless of the multitude of other takes on what this thread should really be about. I don't mean to seriously criticize NwAvGuy's work, or AMB's for that matter, as I simply don't have the requisite engineering knowledge to do so. The previous comment was just a helpful generalization I came by while tying up a number of disparate threads in my mind.

 

[SpaceTimeMorph]

^ Fair enough.

 

[mikeaj]

Quote:

After reading up a bit more on the design and the background debates, I've noticed an interesting parallel. The power output criticism of nwavguy's O2 amp kind of mirrors his own concerns over the Mini3's power output discrepancies among different test setups. Whereas amb's specs were the culmination of a long design cycle and the optimum setup for the greatest measured performance, nwavguy's methodology aims at the average user and uses average componentry [emphasis added]. By amb's exacting standards, the O2 has a very low usable gain range that the predetermined (though since revised) figures would have pushed to clipping. Contrarily, by nwavguy's standards we have an amp that would work across the greatest practical range of headphone impedances and sensitivities to very low distortion, but sacrificing the usable output to meet a very exacting crosstalk spec. A kind of reverse situation from when nwavguy challenged the mini3's crosstalk due to the OPA690 3-channel output decision. So, from an outside perspective it seems like a fanatical resolve was placed to designing to a specific ideal. Personally, I don't quite see how a possible difference from inaudible but otherwise unexemplary crosstalk to an equally inaudible but altogether impressive crosstalk justifies the inherent limitation in design choices. Sure, O2 is a concept amp built on achieving great measurements, but also a budget and wide usability. Both amplifiers appear to be the product of carefully weighed design consideration.
 
Unfortunately, what with the unfortunate subjective/objective rift in the head-fi camp and nwavguy's holy crusade and the designer egos, it almost seems like these approaches are somehow adversarial, rather than complementary.

In my opinion one of the most compelling points here is that you can get very accurate performance driving headphones-level impedances with relatively low-cost components.  If you check the op amp article, you'll see that many of the more expensive alternatives seem to behave worse.  This is especially true if you use expensive parts with great specifications for some applications but not so much for audio, and if you aren't diligent in designing the circuits around them properly.
 
Maybe I misinterpreted you, but the crosstalk has little to do with the design decisions regarding input clipping, so it's not like there's a design tradeoff there.
 
I do agree there's too much ego involved and things getting personal.  But unfortunately, good intents and good efforts don't always result in the best designs.  It's best if you call it like you see it, but some of the reactions to that were not ideal.

 

[Olli1324]

A little note about the group buy...
 
Tomorrow is the last full day before the GB closes and the order is placed.

Please make sure that your order is placed in time so that you don't miss out. Also, if you haven't paid yet then please make sure that you check the online spreadsheet and send meyour payment asap.

Many thanks!

 

[Reticuli2]

[size=10pt][size=10pt]AMB made large mistakes in its crosstalk measurements. I don't think it's appropriate to be drawing some equivalency, attempting to defend that mistake, and thereby bring further criticism when AMB itself is no longer attempting to defend it, either. Furthermore, AMB itself says they do not believe moving the pot back has that much of an effect on the specs, and therefore your parallel is even weaker.  It was probably a bad choice on their part to make that argument using the Mini3 as an example, but that’s water under the bridge.  The gain changes in the stated design of the O2 were always capable of being done by anyone building one and the limitations and possible remedies were disclosed pretty early on. That's very different than making measurement mistakes or using a virtual ground. Even if you just made an honest bad analysis with this parallel, if anyone else tries to bring this up again, we should just assume it's some provocateur trying to push the conversation to bashing AMB. It's not necessary. Give it a rest. [/size][/size]
Quote:

After reading up a bit more on the design and the background debates, I've noticed an interesting parallel. The power output criticism of nwavguy's O2 amp kind of mirrors his own concerns over the Mini3's power output discrepancies among different test setups. Whereas amb's specs were the culmination of a long design cycle and the optimum setup for the greatest measured performance, nwavguy's methodology aims at the average user and uses average componentry. By amb's exacting standards, the O2 has a very low usable gain range that the predetermined (though since revised) figures would have pushed to clipping. Contrarily, by nwavguy's standards we have an amp that would work across the greatest practical range of headphone impedances and sensitivities to very low distortion, but sacrificing the usable output to meet a very exacting crosstalk spec. A kind of reverse situation from when nwavguy challenged the mini3's crosstalk due to the OPA690 3-channel output decision. So, from an outside perspective it seems like a fanatical resolve was placed to designing to a specific ideal. Personally, I don't quite see how a possible difference from inaudible but otherwise unexemplary crosstalk to an equally inaudible but altogether impressive crosstalk justifies the inherent limitation in design choices. Sure, O2 is a concept amp built on achieving great measurements, but also a budget and wide usability. Both amplifiers appear to be the product of carefully weighed design consideration.
 
Unfortunately, what with the unfortunate subjective/objective rift in the head-fi camp and nwavguy's holy crusade and the designer egos, it almost seems like these approaches are somehow adversarial, rather than complementary.

 

 

[Reticuli2]

[size=10pt]Interestingly, we've known this in the DJ and studio mixer genres for a long time. Some of the best designs in history of "boards" use some of the cheapest, easiest-to-find, and ubiquitous opamps, resistors, and caps. It's what you do with them, and many designers apparently don't know how to "support" the parts with the right stuff immediately adjacent and feeding them the right juice or use good isolated circuit paths. However, you do not want to skimp on the faders, pots, and the power supply on them. And it doesn't hurt if the case weighs almost as much as a tank![/size]
 
Quote:

In my opinion one of the most compelling points here is that you can get very accurate performance driving headphones-level impedances with relatively low-cost components.  If you check the op amp article, you'll see that many of the more expensive alternatives seem to behave worse.  This is especially true if you use expensive parts with great specifications for some applications but not so much for audio, and if you aren't diligent in designing the circuits around them properly.
 
Maybe I misinterpreted you, but the crosstalk has little to do with the design decisions regarding input clipping, so it's not like there's a design tradeoff there.
 
I do agree there's too much ego involved and things getting personal.  But unfortunately, good intents and good efforts don't always result in the best designs.  It's best if you call it like you see it, but some of the reactions to that were not ideal.

 

 

[anetode]

Quote:

[size=10pt][size=10pt]AMB made large mistakes in its crosstalk measurements. I don't think it's appropriate to be drawing some equivalency, attempting to defend that mistake, and thereby bring further criticism when AMB itself is no longer attempting to defend it, either. Furthermore, AMB itself says they do not believe moving the pot back has that much of an effect on the specs, and therefore your parallel is even weaker.  It was probably a bad choice on their part to make that argument using the Mini3 as an example, but that’s water under the bridge.  The gain changes in the stated design of the O2 were always capable of being done by anyone building one and the limitations and possible remedies were disclosed pretty early on. That's very different than making measurement mistakes or using a virtual ground. Even if you just made an honest bad analysis with this parallel, if anyone else tries to bring this up again, we should just assume it's some provocateur trying to push the conversation to bashing AMB. It's not necessary. Give it a rest. [/size][/size]

Yes on the drawing some equivalency, no on attempting to defend anything and no on further criticism. Thank you for acknowledging both the honesty and the badness of my analysis, as well as warning against anyone else offering their input.
 
 
Quote:

Maybe I misinterpreted you, but the crosstalk has little to do with the design decisions regarding input clipping, so it's not like there's a design tradeoff there.

Then the misinterpretation was on my behalf, I recall reading that there pot placement was to optimize crosstalk and that it left less flexibility in configuring the design. I'll go back and take another look. Anyway, I agree that O2 is perfectly suited for portable use given the typical DAP voltages where higher gain would not be a problem.
 
The time to market is also remarkable. Given a couple of months I expect to see a multitude of different builds available from DIYers. And a while after that the new toy syndrome will die down a bit and production might scale to allow for some even nicer pricing.

 

[koolkat]

As I blow at anything DIY, I figured I could buy one of these O2s to test them out but was discouraged by the price..
 
The maker of O2 said that the cost of parts wouldn't exceed 50 dollars. It's selling for about 120.
 
Meh.