[psgarcha92]

Guys,
i read all this stuff about the Mini^3 amp.
now, whats the problem with the results here? what i think is that the author used a Mini^3 with a gain of 5, the default gain, which ofcourse wont be good at all for IEMs and low impedance phones.can he really disregard the Mini^3 like this?
And seriously, how is the Mini^3 an improved CMoy? whats this thing about the Crosstalk? after reading his reviews about the Nuforce amps, i really started disliking them, i dont want this to happen to the Mini^3 too.
Thanks

 

[qusp]

a gain of 5 is perfectly fine with iems and has no bearing on how well it will drive the low impedance and he had to test with a range of loads, so was the best choice. i'm not going to go into the rest of it, as i would rather not get involved, read up on crosstalk if you want to know what it is in detail, basically its one channel bleeding into the other. in this case its proposed this is through the ground channel.
 
its harsh what he wrote, but lets just say i'm not always a big fan of ground channel amps either, the ground has to be good and stiff, at least twice as stiff as the other channels and seemingly stiffer than here. basically hes saying that the ground channel is underpowered and since both channels are linked to and supported by it, they bleed into each other through it when driving loads that exceed the current delivery of the ground and need high current like low impedance iems. there are many other problems he reports too, but i do think it could have been handled better on both sides.
 
again i'm not taking sides here, the mini^3 is well respected by many and this is no reason to stop

 

[psgarcha92]

OK, the ground thing and leaking sond from one channel into the another, taken.
but, if he chose a gain of two maybe, atleast the Mini^3 would have been more compatible with the low impedance phones?
i think the noise floor then would have been lower and thus better. what say?

 

[limpidglitch]

I remember Steve Eddy once questioned the assumed fact that a 3ch B22 would be technically superior to a 2ch one, and caused some significant indignation. Doesn't look so stupid now.
 
About driving low impedance 'phones. I seem to remember AMB advicing against it, regardless of gain. MisterX at least certainly does, so it's obvious that at least someone realizes the limitations of the mini3.
 
According to the article author it's the opamp that is the culprit with reference to THD (not noise, the noisefloor seems perfectly fine), and chances are that it would only operate less linearly at lower gain.

 

[psgarcha92]

okk, so, is the impedance a factor? or are sensitivity and impedance going hand in hand to decide how well the Mini^3 would perform?
i mean comparing between the two:

1. Impedance: 64 Ohm @ 1kHz
2. Sensitivity: 100dB @ 1mW

 
and
Impedance: 16 Ohm @ 1kHz
Sensitivity: 100dB/1mW
 
which none would let the Mini^3 shine better?

 

[goodvibes]

Well done article. Noise floor is fine with correct gain but if you use both hi and low impedance phones there could be an issue. Distortion is generally fine driving something over 20ohms to a volt or so. Output resistor coupling resistor is the wrong way to go for short protection. Use a coupling cap. I know tweeks hate coupling caps but sometimes the alternative is far worse. Distortion and driving low impedance isn't just the Op AMP used though it's in the equation. Feedback and how it interacts in circuit has a lot to do wit it. That resistor is limiting as is the way they've done the ground channel. Stiff ground is always a good thing. The problem with 2 batteries is that they can voltage vary from each other without regulation and cause other instabilities but done right or with a non voltage critical circuit should be better. I'm sure that there are virtual ground amps that work extremely well but it's also a solution that introduces new challenges to address.

 

[psgarcha92]

Quote:

okk, so, is the impedance a factor? or are sensitivity and impedance going hand in hand to decide how well the Mini^3 would perform?
i mean comparing between the two:

1. Impedance: 64 Ohm @ 1kHz
2. Sensitivity: 100dB @ 1mW

 
and
Impedance: 16 Ohm @ 1kHz
Sensitivity: 100dB/1mW
 
which none would let the Mini^3 shine better?

 

 

[qusp]

Quote:

I remember Steve Eddy once questioned the assumed fact that a 3ch B22 would be technically superior to a 2ch one, and caused some significant indignation. Doesn't look so stupid now.
 
About driving low impedance 'phones. I seem to remember AMB advicing against it, regardless of gain. MisterX at least certainly does, so it's obvious that at least someone realizes the limitations of the mini3.
 
According to the article author it's the opamp that is the culprit with reference to THD (not noise, the noisefloor seems perfectly fine), and chances are that it would only operate less linearly at lower gain.

according to the article a large part of the problem is the current limit of the opa690, which is operating at unity gain, less than unity gain you say?
 

 

[limpidglitch]

Ha! I'm confusing opamps
Sorry about that.

 

[psgarcha92]

What if the Mini^3 someone uses is the mixture of the High performance version (AD8397 and OPA690)
and the High Runtime version (LMH6643 and LMH6642) as in, using the AD8397 and the LMH6642 (inplace of the OPA690)? would that rectify any of the problems here?
and does only impedance of the load affect the Amplifier's performance? or does the sensitivity too?

 

[mikeaj]

Based on the author's speculation and reading of the alternative op amp specs, the high runtime version would probably be worse, so don't bother.  The alternative would have even more severe limits with lower impedance loads.
 
The amplifier's performance changes with different impedance and different output (voltage, which tells you the current at a given impedance) levels.  The amp falls flat on its face at low impedance loads at levels like 1V (or a little higher).  Then again, most low-impedance 16 ohms IEMs/headphones tend to be pretty sensitive too, so you don't need to get anywhere near 1V to be plenty loud.  The author is a bit harsh in judgment, but he uses a consistent and thorough methodology to test gear in each review.  It's fair to do apples-to-apples comparisons, and it's fair to compare published specs to real world specs.  Just note that relatively poor performance at low impedances at high levels doesn't mean the mini^3 is suddenly no good.  There just are better and cheaper options for low impedance headphones.
 
The mini^3 does much better at 150 ohms and would probably be much better at 64 ohms than 15-33 ohms.  As noted above, the problems mostly stem from the ground channel op amp not being able to supply enough current (for low impedances) and the output resistor (which matters more at lower impedances).

 

[Armaegis]

There's a bit of hooplah over at AMB's forum if you take a look there. TL;DR, the testing methodologies were different. Each person's numbers support their own claims. Meh.

 

[CEE TEE]

Hi Guys,
 
I have a recently-built high-performance version at gain of 2.
 
I am using UE Reference Monitors out of the mini^3 and they have the following specifications:
 
Efficiency:  112 db SPL @ 1 Khz, I mw
Impedance:  35 ohms @ 1 Khz
 
According to the full review and your quote, I might hear distortion in the high frequencies at high levels of volume?
 
So far, I haven't noticed any distortion unless the battery runs low.
 
I might check mine later for this possible distortion issue but so far I just wish I had more/wider headstage/imaging.  
 
The mini^3 is a bit "narrow" in presentation.  I like the sound of it, though.  Brightens my Sony X and my iPhone.
 
Quote:

Based on the author's speculation and reading of the alternative op amp specs, the high runtime version would probably be worse, so don't bother.  The alternative would have even more severe limits with lower impedance loads.
 
The amplifier's performance changes with different impedance and different output (voltage, which tells you the current at a given impedance) levels.  The amp falls flat on its face at low impedance loads at levels like 1V (or a little higher).  Then again, most low-impedance 16 ohms IEMs/headphones tend to be pretty sensitive too, so you don't need to get anywhere near 1V to be plenty loud.  The author is a bit harsh in judgment, but he uses a consistent and thorough methodology to test gear in each review.  It's fair to do apples-to-apples comparisons, and it's fair to compare published specs to real world specs.  Just note that relatively poor performance at low impedances at high levels doesn't mean the mini^3 is suddenly no good.  There just are better and cheaper options for low impedance headphones.
 
The mini^3 does much better at 150 ohms and would probably be much better at 64 ohms than 15-33 ohms.  As noted above, the problems mostly stem from the ground channel op amp not being able to supply enough current (for low impedances) and the output resistor (which matters more at lower impedances).

 

 

[mikeaj]

I kinda skimmed through the AMB forum's thread.  Was there any explanation in the end as to how somebody got 300mW into 33 ohms off of ~9V battery power?  It seems frankly impossible under any testing circumstance with the current design of the circuit with a ~9V supply voltage, and then further implausible with any supply voltage at usual 1% THD based on the OPA690 specs.
Quote:

There's a bit of hooplah over at AMB's forum if you take a look there. TL;DR, the testing methodologies were different. Each person's numbers support their own claims. Meh.

 
 
 
It seems unlikely you'll have problems with distortion unless you listen really damn loud with the UE.  Since the impedance is higher than 15 ohms, you'll probably get somewhat better performance than the 15 ohms tests quoted in the article, run at the 400 mV reference.  And 400 mV gives you nominally 4.6 mW -> 118.6 dB SPL @ 1 kHz.  So only borderline audible distortion at really high volumes...probably.  The IMD tests and 20kHz / 20 Hz THD tests he runs are used because they're apparently kind of worst-case scenarios for stress testing / benchmarking purposes, so take that as you will.
 
The distortion tests at 1V output level were to highlight the issue with the ground channel and that op amp running out of current under a pretty high level.  (Is the level unrealistic, particularly for IEMs?  I'd say mostly yes in practice, but the AMB spec claimed 300 mW into 33 ohms, which requires a whole lot higher than 1V, so the test is hardly unfair.)  I don't think you'll be using a 1V output into 35 ohms, much less the 1V output into 15 ohms that was tested, so the ground channel won't be causing as much peculiarities.
 
Note that crosstalk was like -40 dB at 15 ohms, -60 dB at 150 ohms.  I'm guessing with the UE, you'll have crosstalk in between those figures, closer to -40 dB?  This might cause some imaging issues, but I think that's more likely due to the sound of a particular IEM than the amplifier, at levels like that.
 
I'm kind of curious about the EMI issue though.  Could you find a large power transformer or other such device, put the mini^3 on or near it, and see if you hear anything? 

 
Quote:

Hi Guys,
 
I have a recently-built high-performance version at gain of 2.
 
I am using UE Reference Monitors out of the mini^3 and they have the following specifications:
 
Efficiency:  112 db SPL @ 1 Khz, I mw
Impedance:  35 ohms @ 1 Khz
 
According to the full review and your quote, I might hear distortion in the high frequencies at high levels of volume?
 
So far, I haven't noticed any distortion unless the battery runs low.
 
I might check mine later for this possible distortion issue but so far I just wish I had more/wider headstage/imaging.  
 
The mini^3 is a bit "narrow" in presentation.  I like the sound of it, though.  Brightens my Sony X and my iPhone.

 

[Yoga Flame]

 
Quote:

I kinda skimmed through the AMB forum's thread.  Was there any explanation in the end as to how somebody got 300mW into 33 ohms off of ~9V battery power?  It seems frankly impossible under any testing circumstance with the current design of the circuit with a ~9V supply voltage, and then further implausible with any supply voltage at usual 1% THD based on the OPA690 specs.
 

I recall seeing a post by AMB in that thread explaining the calculation he used to arrive at that spec. NwAvGuy followed with his own calculation showing different result. I don't have any proper EE knowledge so I can't independently verify for myself either one. (Another "impossible" thing was when NwAvGuy observed a 60Hz hum when running the mini3 off a battery.) Perhaps AMB miscalculated. But if I had to pick a side purely on faith, I'd go with AMB because of his longstanding reputation and contributions to the community, over NwAvGuy who we know next to nothing about.
 
 
Quote:

I'm kind of curious about the EMI issue though.  Could you find a large power transformer or other such device, put the mini^3 on or near it, and see if you hear anything?  

 

 
 
Remember to grind off the Hammond case's interior anodizing so the board's ground strips can make proper contact. NwAvGuy neglected to do that, but said grounding the aluminum case would make the EMI issue worse(?). There wouldn't be as many objections to his measurements if he had only used a Mini3 that was properly built to spec.
 
 
FWIW I'm very happy with my own Mini3 and have not noticed any hum, distortion, or crosstalk on it. Never tried measuring it though, and I mostly use mid-high impedance headphones.