So, the Objective2 headphone amp - designed entirely around the measurements? (PLEASE READ RULES BEFORE POSTING)

[mikeaj]

Quote:

THD and cross-talk increase with gain.

 
Indeed.  However, the differences skeptic quoted are more or less apples-to-apples, at the same output level into the same loads with typical default gains set.
 
Most of the published results are at that 3.1X or maybe 2.5X gain, which is important to note.  The difference in THD+N is given in one graph for 1X gain as opposed to 6X gain at 2V rms, and this shows a pretty small difference in performance (less than double THD+N at any frequency and generally much less than double, so still below accepted levels of audibility) but this is for a relatively easy 600 ohms load.  Then again, you probably won't be wanting to use 6X gain with most 16 ohms headphones.

 

[Shike]

Use a gain of 2 and 6, or 2 and 3 if no low voltage (sub 2V units) are used.  Congrats, you've covered just about every headphone and source.  If a gain of two is too high, then your source is either a very special case (some tube designs) or just plain built way out of standards.
 
It's really a non-issue with even a moment of forethought.

 

[amb]

Quote:

Of the measurements discussed by nwavguy, noise is hardly the chief concern when it comes to comparing the mini3 and O2, amb - as you well know.  How bout the 25db difference in measured crosstalk between the mini3 and o2 at 15 ohms?  How about the THD at 20khz? 

 
The whole premise of placing the volume pot after the gain stage instead of before is about lowering noise, hence the focus on that subject.  The rest is red herring.  The point is not to compare the Mini³ against the O2, only to illustrate that the supposed noise advantage of having the pot after the gain stage is small and negligible.
 
Quote:

Current is less than 0.1A, not 1A (at least according to the designer, when running the clipping tests).  So the voltage drop with 1A is not relevant.  Dare I say that you were mistaken here?

OK, you have a point.  The forward voltage changes with current, and is lower at low currents.  But we're talking about about 0.2V difference each here, which is small compared to the losses in the NJM2068 opamp.
 

 

[Willakan]

Wait a second, I thought we were discussing power output primarily, not pot position. What happened to that?
To cut to the chase, here is graph showing power output. Is it wrong/impossible? Sorry for the Mini3 figures, but I can hardly edit them out of the graph.
The designer has since responded again at ABI - NwAvGuy politely suggests that respond to him there instead if you still think your argument has merit.

 

[Maxvla]

Will a desktop version of this amp put out much heat? My desk is in a corner and tubed amps make it rather toasty over there.

 

[Anaxilus]

Quote:

Will a desktop version of this amp put out much heat? My desk is in a corner and tubed amps make it rather toasty over there.

No worries.  Any heat will be negligible compared to the tubes.  Heat is absolute and not additive so you should notice little to no effect to your ambient temps.  To be more clear, you won't exceed your current tube amps maximum heat output but your total ambient might tick up a degree (F) depending.  More than likely whatever monitor or TV you are using is outputting more heat than the O2 unless you're on your laptop (depending on the laptop). 
 

 

[Maxvla]

Getting more and more interested in the desktop O2. Great for sensitive IEMs, low(er) heat, low price, stackable (no tubes), and more. Just give me a case that doesn't look like shambles (a good portion of the DIY stuff) from a competent maker and I'm sold on trying one.

 

[Reticuli2]

You mean it's absolute when not considering a confined space, right?  So in his little corner heat will not build up, but in the entire room it will.  Unless I misunderstand something about heat so that "heat is absolute and not additive" is completely correct.  I thought the opposite was true.
 
Quote:

No worries.  Any heat will be negligible compared to the tubes.  Heat is absolute and not additive so you should notice little to no effect to your ambient temps.  To be more clear, you won't exceed your current tube amps maximum heat output but your total ambient might tick up a degree (F) depending.  More than likely whatever monitor or TV you are using is outputting more heat than the O2 unless you're on your laptop (depending on the laptop). 
 

 

 

[Steve Eddy]

Quote:

Originally Posted by Willakan /img/forum/go_quote.gif
 
The designer has since responded again at ABI...

ABI? What's ABI?
 
se

 

[chinesekiwi]

Anythingbutipod.com

 

[khaos974]

ABI? What's ABI?
 
se

http://anythingbutipod.com/forum/index.php

 

[Steve Eddy]

 
Ah, thanks! Only other place I'd been following it is over at diyAudio.
 
se
 
 

 

[purrin]

Quote:

Didn't two versions of a defective Cmoy and an E7 also blow your Mini3 away?

 
Quote:

...He also expressed amusement that Ti would rigidly declare, when both the Mini3 and his design use the same battery type, that the batteries output 8.4V, as whilst trying to defend his own designs he insisted they output considerably more.
 
So to paraphrase that, the maths is wrong, the argument is wrong and Ti appears slightly contradictory to the uneducated observer.

 
Quote:

Of the measurements discussed by nwavguy, noise is hardly the chief concern when it comes to comparing the mini3 and O2, amb - as you well know.  How bout the 25db difference in measured crosstalk between the mini3 and o2 at 15 ohms?  How about the THD at 20khz? 

      Quote:

 Sorry for the Mini3 figures, but I can hardly edit them out of the graph.

 
---
 

This is really wierd, how did this thread become O2 > mini3? From my reading, it appears that Ti threw the mini3 into the ring for purposes of illustrating the (small) noise differences related to the placement of the pot.
 
It seems the "friends" of nwavguy or nwavguy himself have some kind of axe to grind with Ti - using pseudo authoritative statements such as "So to paraphrase that, the maths is wrong, the argument is wrong and Ti appears slightly contradictory to the uneducated observer." Hmm, this sounds a little bit like nwavguy himself. This seems like a very aggressive method of discourse, keeping in mind that no one forced the topic of the O2 design in these forums.
 
I thought this thread was meant as sort of a public commentary on the design of the O2. Why can't the "friends" of nwavguy just politely acknowledge Ti's criticisms, i.e. "We've spoken with nwavguy, and he has taken Ti's design concerns into account and lowered the gain. (Thank You!) Nwavguy's real-world measurements of the opamps voltage swing exceed those of the spreadsheets. Therefore Nwavguy feels that this will adequately address the expected use and scenarios of the typical user." 

Case closed.
 
Geez!
 

 

[Willakan]

The graph is NwAvGuy's. It is the only graph showing power output. It compares it to the Mini3 and there is nothing I can do about that. Feel free to make something irrational of it.
As to Ti's criticism, I am simply passing on what NwAvGuy has to say about it. Don't shoot the messenger. The slight gain decrease is to allow for the use of some extremely hot sources (by consumer standards) with the amplifier. The bulk of Ti's criticism about headphone driving capability is strongly contested by NwAvGuy - I am reproducing what he said - otherwise it might appear to people on such a popular website as Head-Fi that the design is flawed - especially considering the prohibition of links to other areas in which it is discussed.
 
As repeating what NwAvGuy says appears to lead to considerably..erm...friction, I would recommend Ti heads over to ABI to read and respond for himself.
 
EDIT: NwAvGuy has added a section on the input stage to the most recent article on his site. 
 

 

[SpaceTimeMorph]

Just to reiterate:
 

 
The above graph indicates fairly symmetrical (i.e. almost identical +/- peaks) clipping performance running from an AC source.  +10Vp -> -10.5Vp for a total of 20.5Vpp.  This is the maximum the O2 can put out before clipping, and it shows great linearity with respect to the input sweep right up until it clips.  Therefore, by this graph you are seeing the voltage on the power supply rails of the O2; 20.5Vpp.  That's precisely 7.24675Vrms.
 
Now calculating the dB equivalent of this into a 1000ohm load (note the 1000ohms and the *1000 to convert to mW cancel out):
 
10*log(7.24675^2) = 17.2 dB.
 
Therefore, with a 1000ohm headphone you would need a sensitivity of <93 dB/mW to find a set of headphones this couldn't drive off of AC power.  Move this value down to 600ohm:
 
10*log[(7.24675^2/600)*1000] = 19.4 dB.
 
This needs a sensitivity of <91 dB/mW before it becomes a problem.
 
300 ohm:
 
10*log[(7.24675^2/300)*1000] = 22.4 dB.
 
Sensitivity < 87 dB/mW. 
 
So, all that being said, what's the problem here?  Even if you want to discount the designers comments about "one size fits all", etc... this is still a <$100 DIY project that can rival amps many times its price.  You are going to struggle to find a gain structure/source voltage/headphone combination that clips.
 
AMB, you claimed you needed a 5.6 Vrms signal to reach 114 dB with your headphones.  The O2 was designed to reach a target spec of 110dB.  That's a little over half the voltage that you are claiming you need.  So even if you want that much headroom, you would only need 3.53 Vrms, which would give you >100% headroom until you hit the power supply rails.  (This still gives you plenty of headroom on battery power: ~5 Vrms supplied).
 
(I will message NwAvGuy to see if he has graphs for the clipping point on battery power, however, he claims 14Vpp here which would still be enough if you actually run a "normal" portable source.)
 
The comments about other amps being able to accommodate any input voltage are fallacious.  Every amp (no matter the topology) is going to clip at a value roughly corresponding to what the power supply will be able to supply to it.  This means that every amp with a specific gain setting will clip with a specific voltage source.  There is no magic that will allow an amp to suddenly not clip if the input exceeds the max allowed.  Even if you are talking about a Class G or H amp, there is still a clip point involved.  The O2 has a relatively high clip point compared to the competition even considering the input overload "issue."  Again, this is an amp that costs <$100 as a DIY project.