[Jensigner]

Some RMAA audio performance measurements comparing line-in, optical S/PDIF and USB modes using headphone loading of 33 ohm:
  http://www.jensign.com/FiiOE17/RMAA/ComparisonFiiOInputs.htm
Results are discussed at bottom of this page:
  http://www.jensign.com/FiiOE17

 

[stv014]

Note that in the USB case you likely created a ground loop between the X-Fi sound card and the USB port on the same PC. That is why it looks much noisier than the S/PDIF input, and also has (on the graphs) different crosstalk, even though a different digital input should not affect crosstalk.
 
By the way, Tyll Hertsens at InnerFidelity has recently measured the amplifier of the E17 with an Audio Precision analyzer. The results can be seen here and discussed here. The results are similar to yours for frequency response and crosstalk (if you ignore the ground loop affected USB mode, and take into account that RMAA lies about crosstalk and shows it to be 6 dB better than reality, as can be seen here as well), but show some odd distortion issues.

 

[Jensigner]

Thanks for the link to the InnerFidelity measurements and comment on RMAA xtalk lie. Yes I thought that the USB might cause some extra noise. BTW there is no difference in the noise/distortion for the USB case with/without USB charging enabled on the E17.
Also, with optical SPDIF input, there is no apparent advantage according to my RMAA measurements with using the line-out of the L7 doc accessory vs. the headphone as a line out (when loaded with 1 kohm).
My audio measurements for the E17 line-in RMAA measurements are at the limit of my measurement capability http://www.jensign.com/XFiRMAA  (an X-Fi Elite Pro sound card) and obviously not as good as the Audio Precision results

 

[ClieOS]

I do some measurement via RMAA form time to time as well. It won't be a bad idea to invest into a small USB isolator to help with the ground loop issue. They are around $50 or less on Amazon or eBay these days.

 

[Jensigner]

Quote:

I do some measurement via RMAA form time to time as well. It won't be a bad idea to invest into a small USB isolator to help with the ground loop issue. They are around $50 or less on Amazon or eBay these days.

Are the InnerFidelity measurements for the line-in connection?  Are there audio measurements with USB digital input demonstrating difference between no-isolation and isolator for say a "typical" laptop computer? I noticed that FiiO in their product specs (they use Audio Precision also) don't include a distortion vs freq. plot for SPDIF or USB playback.  I think most users that get the E17 for computer USB playback will expect good performance and probably won't like having to spend an extra $50. for USB ground isolation to get hopefully similar great performance to the line-in connection.
 
I have a question into FiiO about the SPDIF and USB performance specs.

 

[stv014]

The isolator should normally not be needed if the E17 is not connected to anything else than the USB port and the headphones, which is the typical usage. A ground loop is created when you also connect it to the line input of a sound card in the same PC, because the E17 is then grounded to the PC through both the audio cable and the USB cable. PCs tend to have a noisy ground, so it is likely that there is some noise voltage between the ground at the USB port and at the input of the sound card.

 

[ClieOS]

An USB isolator will be helpful when you are doing measurement on USB stuff to your soundcard.It really won't affect the sound of E17 since you don't typically feed it back to sound card, but to headphone - so there is no loop. I only mentioned it since I saw you have done quite a few RMAA measurements in your site.

 

[Jensigner]

I understand about the ground loop. However I'd expect ground loops to only affect the noise floor. However both the SPDIF and USB inputs show much higher distortion levels so that sounds like the DAC part to me. Just a thought.
 
I should ad that I am not an audiophile by any stretch (as you can tell my by equipment). But I like a good challenge and the recent problems with many Samsung Galaxy S4 http://www.jensign.com/S4Distortion related to low-impedance headphones got me interested in this stuff again.

 

[stv014]

Quote:

I understand about the ground loop. However I'd expect ground loops to only affect the noise floor.

 
That is not necessarily true, for example, the ground loop affected crosstalk here (with a low impedance load connected to a TRS output, most of the crosstalk is from the impedance of the common ground). It can also affect the distortion in some cases.
Quote:

However both the SPDIF and USB inputs show much higher distortion levels so that sounds like the DAC part to me. Just a thought.

 
In the S/PDIF case, there is no ground loop, so that probably does correctly show higher distortion. However, it is theoretically also possible for the line input test to show better distortion than reality because of grounding issues. That could be the case if the E17 has a poorly designed virtual ground (like the AMB Mini3), which is basically bypassed in the sound card loopback measurement, although the crosstalk results (consistent between line in and S/PDIF) do not suggest that is the case. According to the InnerFidelity PDF file, the amplifier in the E17 does have relatively high distortion, but maybe that is just a random bad sample.

 

[Jensigner]

Quote:

 
..... According to the InnerFidelity PDF file, the amplifier in the E17 does have relatively high distortion, but maybe that is just a random bad sample.

Here are some more tests of my E17 for the line-in connection comparing 16bit/44.1kHz with 24bit/96kHz:
   http://www.jensign.com/FiiOE17/RMAA/LineIn/ComparisonE17_LineIn.htm
Loaded into 33 ohm and at a headphone output level of 1.8Vrms. Unless I misunderstand the InnerFidelity measurements, my results look quite a bit better so yes maybe a bad sample.

 

[Jensigner]

another tidbit ... tested the coax SPDIF input and it is (probably not surprisingly) identical to the optical SPDIF input. This just verifies that there is nothing odd occuring between optical/electrical digital conversion. The RCA/phone-plug adapter provided with the E17 sits with about 1mm+  of the gnd part of the plug showing ... not a very satisfying feeling ... hopefully the hot part of the plug isn't bottoming out and touching the optical lens for optical SPDIF!
 
I also tried another optical spdif output from another configuration and it is also identical. Therefore the IMD distortion for the E17 with SPDIF input would appear to be coming from the DAC part of the chain.