[reginalb]

Schiit Fulla 
 
Added a comparison to a loaded measurement with the Merlins, very impressive performance.
 

 
Details for 20Hz to 20kHz, loaded: +0.08, -0.51 dB

 

[reginalb]

HiFiMan HM-602
 
I'd expect it to be not great loaded, but how bad can it be?
 

 
This was once marketed as the pinnacle of audiophile sound. 

 

[HiFiChris]

I'm not much surprised to see that as the output impedance of HiFiMan's DAPS is known to be somewhat too high for driving multi-driver in-ears "correctly".
As it seems too, they have used a NOS DAC, hence the "smoothed-out" treble.

 

[HiFiChris]

Luxury & Precision L3 (single-ended Output):
 
 
No Load:
 

 
 
Triple.Fi 10:
 

 
 
Various Multi-BA IEMs:
 

 

[HiFiChris]

FiiO X3 (first Generation):
 
With the Triple.Fi 10 as my usual critical Multi-BA Load:


 
-> Really low output impedance clearly below 1 Ohm.

 

[HiFiChris]

$1.49 MP3 Player:
 

 
Some cheap no-name stuff I bought on e-bay for $1.49 each some time ago.
 
 
Loaded Measurement with the Triple.Fi 10:
 

 
What the fudge?! This thing that is cheap as dirt has got a low output impedance somewhat below 1 Ohm?! Yes, it really does. Nonetheless, I wouldn't recommend anyone to really use it (noise floor, DAC artifacts, pulsating LED audible in the left channel, ...).

 

[castleofargh]

  HiFiMan HM-602
 
I'd expect it to be not great loaded, but how bad can it be?
 

 
This was once marketed as the pinnacle of audiophile sound. 

very typical of a source without enough impedance difference vs the load(1/10 rule blablablah).  and looking at the impedance curve of the PS1000 http://www.innerfidelity.com/images/GradoPS1000.pdf  your result fits the bump before 100hz and does rise back a little when reaching 20khz, so you have little reason to suspect a mistake on your side. now you can guess most of the general shape of the impedance curve of the merlin(no amplitude scale for ohm relation though). about the same overall shape except that you would need to at least remove the roll off of the hm602 unloaded, to the graph of the merlin. so the merlin's impedance seems to rise from about 9khz to 20khz in changing ways but going up still.

 

[thuNDa]

Hi,
 
-ALC 892 Onboard-Sound
-Soundblaster Z
-SMSL SD-793II
-SMSL Sanskrit 6th + Objective2
 
 
 
Sennheiser HD590 (120 Ohm)
 

 
AKG K712 (64 Ohm)
 

 
(and these guys from tomshardware, still haven't updated their statement: http://www.tomshardware.com/reviews/high-end-pc-audio,3733-19.html , while they actually knew for 2 years: http://www.tomshardware.com/reviews/high-end-pc-audio,3733-17.html)
 
EDIT: crap, pics aren't that sharp...
 
EDIT2: oh, they are if you click on "original"

 

[reginalb]

 
very typical of a source without enough impedance difference vs the load(1/10 rule blablablah).  and looking at the impedance curve of the PS1000 http://www.innerfidelity.com/images/GradoPS1000.pdf  your result fits the bump before 100hz and does rise back a little when reaching 20khz, so you have little reason to suspect a mistake on your side. now you can guess most of the general shape of the impedance curve of the merlin(no amplitude scale for ohm relation though). about the same overall shape except that you would need to at least remove the roll off of the hm602 unloaded, to the graph of the merlin. so the merlin's impedance seems to rise from about 9khz to 20khz in changing ways but going up still.

 
Makes sense! Kind of fun taking these measurements. It's always fun to load them up and see what I got!

 

[reginalb]

Schiit Valhalla 2 via Behringer UCA222
 
Alright, I was a bad objectivist, and pounced on a decent deal for a Valhalla 2 (yeah - it's a tube amp...)
 
I've long wanted to play around with one. I'm too young to have used one in any other context, and I think tubes look cool...so I picked one up Will probably sell again at cost in a while, but thought I'd do some tests with it. Not bad performance. Probably committing a cardinal sin using it through a $30 DAC (that does more than just DAC work, mind you).
 
I did think it sounded like the bass was a little elevated out of My Grados with it, and indeed, that's what I later measured, though it's not much. Interestingly, it behaves a little differently below 50Hz than anything else I have measured with those headphones. Most of them have a bass response just below 0dB, where as this one peaks down there at +.5dB. 
 
Looks a little better at the high end than my AK300. Also of note, I'm now pretty sure that my interface causes the waves that I get north of 10KHz, since everything I measure gets wavy up there. 
 

 

[castleofargh]

did you try setting you ADC part of the loop to more than 44.1khz(or all the chain depending on what you want to test)? I have one input where I also get wiggling stuff in the trebles, but admittedly it doesn't start as soon as shown on your graph.

 

[reginalb]

  did you try setting you ADC part of the loop to more than 44.1khz(or all the chain depending on what you want to test)? I have one input where I also get wiggling stuff in the trebles, but admittedly it doesn't start as soon as shown on your graph.

 
Yeah...that did the trick. Thanks!

 

[RRod]

   
Yeah...that did the trick. Thanks!

 
If you look at your original output on a linear frequency scale, do the amplitude variations look somewhat sinusoidal? I've noticed this in measuring various devices as well.

 

[reginalb]

   
If you look at your original output on a linear frequency scale, do the amplitude variations look somewhat sinusoidal? I've noticed this in measuring various devices as well.

 
I'll have to take a look.
 
I did decide to try it with my Merlins, probably not the right amp for these:
 

 

[HiFiChris]

Chord Electronics Mojo:
 
No Load:
 


Something that could be considered as a slight slow roll-off filter. Nothing wrong with that.
 
 
Triple.Fi 10:


 
Seems like the output impedance is higher in the upper and super treble than it is in the bass and mids.
 
 
Various Multi-BA IEMs:
 
 
 
Seems like the Mojo doesn't fully like low impedance multi-driver loads.
 
 
IE 800:
 

 

Various Full-Sized Headphones:
 

 
 
I unfortunately don't have the link to the test/article in a print anymore, but it had the measurement on hand and stated that while the output impedance was very low (I think measured at) 1 kHz, it was a bit higher at (I think measured at) 20 kHz. 

The question is whether these more or less slight deviations do matter to you or not as most of them are above 10 kHz anyway, however objectively speaking, the Mojo could stay somewhat more linear in the treble with complex and low impedance multi-driver loads.
 
I read a statement from Chord Electronics some time ago that they designed "[...] Mojo['s] analogue section [...] very simple, so giving Mojo's transparency, but the downside is a small variation in frequency response with load impedance" (http://www.head-fi.org/t/784602/chord-mojo-the-official-thread-please-read-the-3rd-post/4320#post_12058437).