High output impedance from HP amp and HP frequency response.

2016-02-21 22:48:02

hi there,
i recently traded my sennheiser hd650 (300 ohm) for hifiman he-400i (35 ohm) and now i'm wondering what's the best soundcard for them.  i used the popular asus xonar stx with its 10 ohm output impedance for hd650, but from what i read it wouldn't be that great with the 35 ohm hifiman and might mess up its frequency response pretty bad. my onboard soundcard is alc892 and from what i managed to google its output impedance is 2 ohms and it measures pretty good in other parameters. unfortunately it doesn't quite reach the volume i need on some tracks even when it's maxed out, obviously xonar does it effortlessly. 

so the question is, is there a way to predict how badly/whether at all xonar messes up the FR when paired with 35 ohm hifiman? and given the fact that hifiman are planar magnetic - is there a reasons to expect that this impedance mismatch would act differently?

(it would be interesting to abx the two soundcards with hifiman, but i just don't have the right tools...)

thanks.

Re: High output impedance from HP amp and HP frequency response.

Reply #1 – 2016-02-22 20:51:44

1) The STX is fine.
Here's why: frequency response will only change due to high output impedance if the load (the headphones) has varying impedance vs. frequency. Planar magnetic headphones (like your he-400i) however are pretty much resistive. In other words, the impedance is pretty much constant vs. frequency.
If you had a dynamic headphone with, for example, 30 ohm at 1 kHz and 60 ohm at 100 Hz then you'd get a little "bass boost" with 10 ohm output impedance.

2) Onboard usually has high output impedance.
Expect >100 ohm from the line-outs in the back, and >75 ohm at the front panel outputs.


Output impedance does have an effect on output volume btw. That's where the "bass boost" in the example above comes from. Part of the voltage will drop across the output impedance, and the rest will drop across the load.
For example
10:30 => 30/(30+10) = 0.75 = -2.5 dB
10:60 => 60/(60+10) = 0.86 = -1.3 dB
You see, if the impedance is 60 ohm at 100 Hz then less voltage drops across the output impedance, so you effectively get a little bass boost.

With a 100 ohm output impedance, even if the onboard could output 2V unloaded, the max voltage across a 30 ohm load will therefore be only 0.23V which is almost -13 dB.

Re: High output impedance from HP amp and HP frequency response.

Reply #2 – 2016-02-22 21:31:41

thanks for the explanation, i think many folks are unaware that planar magnetics are unaffected by this.

regarding the alc892 output impedance, it is actually 2 ohm when its HP amp is engaged (via driver), i just didn't have the right source to back my claim in the first post. this is legit though: http://www.hardwaresecrets.com/datasheets/ALC892-CG_DataSheet_1.3.pdf  - look at page 82.

Re: High output impedance from HP amp and HP frequency response.

Reply #3 – 2016-02-22 23:05:08

You're welcome. I think that a few years ago most headphone audiophiles were pretty unaware of both frequency response and impedance measurements. This somewhat changed when NwAvGuy and other guys made posts about it. But still, most of them have only heard of a "1/8th rule of thumb" without knowing what or why. And there is also still a lot of confusion about amplifiers and impedance, sensitivity etc.

If the output impedance is <1/8th, then for a headphone with an impedance peak that's 2x (e.g. 30 ohm at 1 kHz, 60 ohm at 100 Hz, which is not uncommon for a dynamic headphone) then the max difference in frequency response will be +1 dB.

--

I am aware of the datasheets. The problem is that the front panel stuff typically contains >75 ohm resistors, and that raises the output impedance.
Your motherboard or notebook may be a special case, but I have measured different Realtek audio codecs (both higher and lower end than yours) and they all had >100 ohm at the rear connectors and >75 at the front panel.

Maybe building a custom front panel header adapter would help... never tried.

Re: High output impedance from HP amp and HP frequency response.

Reply #4 – 2016-02-22 23:23:15

yeah, i mean there are basically 3 rough types of audiophiles: the whathifi/Steve Guttenberg/most of headfi types; the people who are able to see that the former are simply talking baseless garbage and recognize actual scientific argument though without fully understanding it; and the people with with actual knowledge like here on hydrogen, NwAvGuy and maybe audioholics. i fall into the second category but without the explanation i would also automatically apply the 1/8th rule, i just had some hunch that it might be different with planars.

regarding realtek, my guess is that you measured it when it was in line out (speaker out) mode, you can "retask" its outputs to HP amp mode in the control panel. it really seems unlikely that they would mislead by such large margin really.

Re: High output impedance from HP amp and HP frequency response.

Reply #5 – 2016-02-24 17:35:27

Planars tend to have pretty much ruler flat impedance responses, so they are not too critical WRT output impedance per se. They tend to require a good amount of output stage "oomph" though, due to typically lowish sensitivity at low-mid impedances (shifting requirements closer to speaker amp territory than what you typically need for headphones). With ~1 Vrms and change on a typical onboard audio output plus the common 75 ohm series resistor, a 35 ohm load won't see an awful lot of output level, like 300-ish mVrms. Hence why levels on the onboard output are uninspiring. Even more so on the rear outputs, obviously, which tend to be at ~200 ohms from the chip itself when in line-out mode.

The STX with its lower output impedance and higher intrinsic output level capability will go a fair bit louder, but it's not exactly too fond of driving low-impedance loads at mid-high levels. I remember something like -70 dB THD (mostly 3rd) when driving 1 Vrms at 1 kHz into 32 ohms, which is not especially exciting. I do think a TPA6120A2 should fare better than that, maybe it's a layout or power supply issue. An amp that'll drive such a load effortlessly is generally going to cost more than the entire card though, mostly due to additional components (case, power supply, heatsinks) and unfavorable economics of scale.

Re: High output impedance from HP amp and HP frequency response.

Reply #6 – 2016-02-24 20:51:55

93 db/mW for hifiman isn't that low i think, even my phone (note 4) can get them pretty loud at max volume. 

the modern realtek outputs are dual purpose though, i keep saying it but you guys seem to be ignoring that it can act as either line out (i guess in this mode it sports a high output impedance) and HP amp mode. the latter according to the specs is 2 ohms. it does get loud, but i just miss a few more notches when i really wanna blast it.

the highish thd on STX is interesting though, is -70 db close to being audible though?