Damping Factor

[castleofargh]

  I just got my iBasso DX90 (0.3 ohm) and will order a resistor so I can rule out different levels of DAP performance when comparing damping factors.
 
I think it`s clear that a high output impedance (=> low damping factor) degrades sound quality with headphones that have large variances in impedance.
Here are measurements from Benchmark Media.
It would be great if someone could make similar measurements with headphones that have a flat impedance response.

 they indeed show that higher source impedance means higher distortion, but part of that is simply because with the volume matched at 1khz, something like 50khz will be louder on the higher impedance source.
and the headphone simply has more distortion when going louder. so I think a good part of what they show is explained by the frequency response variation.
would the 50hz at matched level have that much distortion? I don't think so.
 
still the result is the same and I'm all for impedance bridging on systems that were intended for it.

 

[Kaffeemann]

We are talking about a 0.7 dB increase in bass response. That's never enough to explain the huge rise in distortion.

 

[ab initio]

  We are talking about a 0.7 dB increase in bass response. That's never enough to explain the huge rise in distortion.

Frequency balance/equalization is a linear process: namely, frequency components present in the signal are scaled by a (complex) factor. It is a process where the relationship between the system's input and output is independent of the amplitude.
 
Distortion is a nonlinear process: namely, frequency components are created that weren't in the original input signal. It is a process that does depend on the input's amplitude.
 
Based upon the article, it's not clear to me where the distortion is actually coming from. My understanding from skimming through the article is that the all measurements were purely electrical (they didn't mike the headphones at all). Basically, one measurement was measuring the output of the amp (while connnected to a headphone) while the other measurement was conducted between a resistor and the headphones. 
 
My mental picture of the configuration is as follows (where 'x' denotes location of the measurement probes):
[AMP (-)]---x---[(-) HEADPHONE (+)]---x---[(+) AMP]
vs
[AMP (-)]---[RESISTOR]---x---[(-) HEADPHONE (+)]---x---[(+) AMP]
 
It wasn't clear to me whether the observed distortion was due to nonlinearities in the headphones or if it was due to distortion in the amplifier. It didn't seem like they were using unreasonably high output levels, so the amp *shouldn't* have been stressed too hard. Nevertheless, we don't really know the complete story because we don't know what the changes in the sound output were.
 
 
Cheers

 

[stv014]

It wasn't clear to me whether the observed distortion was due to nonlinearities in the headphones or if it was due to distortion in the amplifier. It didn't seem like they were using unreasonably high output levels, so the amp *shouldn't* have been stressed too hard. Nevertheless, we don't really know the complete story because we don't know what the changes in the sound output were.

 
For a test with both acoustic and electrical measurements, check my link from a few posts ago. While these were not made with as expensive equipment as the Benchmark one, they do show that the acoustic distortion increases approximately by the same amount that is measured on the voltage input to the drivers. Of course, it is not "100 times worse distortion" then, like in the Benchmark article, but more like ~2 dB at the resonance frequency, at least with the headphone and output impedance used in the test.
 
The increased distortion with a high source impedance is most likely simply the headphone distortion fed back via back EMF (the similarity of the distortion vs. frequency graphs in both types of measurement agrees with this theory as well), and it is
reduced by electrical damping when the source impedance is low.

 

[SilverEars]

A lightbulb hit when I was discussing about amp output impedance and Ethan's video helped.  I have a better understanding of what the damping factor really is.  It's basically damping caused from the field built up from the motion of the cone.  The damping is the back EMF or the oppposite force that is resisting the movement caused by the audio signal from the built up field in the coil.  What really interesting is when you look at the graphs of German headphone, they have a large resonace peak where half of it is inductive since it's positive phase.  Those are the frequencies where the FR, distortion, and decay will be affacted.  
 
What I'm curious about is what is happening when the amp's output impedance is large and for this reason the cones have less damping which will cause it to move to more extremes and the EMF that should have build up is sent to the amp.  What is this EMF doing to the signal to the headphones?

 

[SilverEars]

Top is result of adding resistance of different values to the amp's output to the HD800 which has that huge resonance peak. You see the phase is 0 at the resonance peak where it's resistive, and the phase is positive where the impedance is rising where it's inductive.
 

 

[castleofargh]

I also noticed that some time back, more by luck than real understanding. but my practical "no fun" thinking decided that the impedance response was telling as much as the phase response, and some more. so I just dismissed the phase altogether as "something useless" ^_^.

 

[stv014]

As it is likely close to a minimum phase response, the phase can be approximated reasonably well from the magnitude response. Also, on the graphs above, the frequency response is simply modified according to how the electrical output of the (high impedance) amplifier is affected by the load reactance.

 

[castleofargh]

   

 

just realized, but maybe you should have said what values where used here. because from my simplified calculations I ended up, just like someone mentioned on the now derailed hard drive topic, with about 1DB for a 120ohm amp. so What are these graphs supposed to be? 200ohm and 1000ohm amps ^_^?
 
as is, it gives the idea that the HD800 will actually change signature from higher impedance amps, when even something as ludicrous as the crack amp would make an almost unnoticeable bass boost(1DB in the bass is actually not that easy to identify compared to 1db in the mids at least for me).

 

[SilverEars]

  just realized, but maybe you should have said what values where used here. because from my simplified calculations I ended up, just like someone mentioned on the now derailed hard drive topic, with about 1DB for a 120ohm amp. so What are these graphs supposed to be? 200ohm and 1000ohm amps ^_^?
 
as is, it gives the idea that the HD800 will actually change signature from higher impedance amps, when even something as ludicrous as the crack amp would make an almost unnoticeable bass boost(1DB in the bass is actually not that easy to identify compared to 1db in the mids at least for me).

It's measured through a CMOY, and the impedances are 0, 170, and 680ohms.  FR is not the whole story as it only tells us the difference in loudness, there are other aspects that could be affected.

 

[Kaffeemann]

Can you measure distortion or time domain characteristics?

 

[SilverEars]

  Can you measure distortion or time domain characteristics?

I found the graph online, and I don't have measurement tools.

 

[Kaffeemann]

My 30 Ohm resistor finally arrived and Iistened to my iBasso DX90 (0.3 Ohm) and this IEM with and without the resistor.
The bass does sound different with the resistor. Softer, more blurred, it's harder to hear single bass notes.

 

[SilverEars]

30ohms will mostly likey only reduce loudness based on what I see of the impedance response.  When you swap out the adapter, your volume well not be the same, you are raising the volume.  Try to see if that is the case.  Given the flat impedance graph, I don't know what the adpater would do besides attenuate the power to the iem. 

 

[Kaffeemann]

double post