1. WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements of RME ADI-2 DAC

Discussion in 'Digital Audio Converters (DACs)' started by amirm, Apr 7, 2018.

  1. Rod

    Rod Active Member

    Joined:
    Mar 24, 2018
    Messages:
    173
    Likes Received:
    22
    Tried to plug the headphones into both the Xduoo and the RME amp at the same time. Got a short warning. That doesn't work.
     
  2. amirm

    amirm Founder/Admin CFO (Chief Fun Officer) Staff Member

    Joined:
    Feb 13, 2016
    Messages:
    11,042
    Likes Received:
    2,918
    Location:
    Seattle Area
    We lost power due to the storm. Running on the generator.

    For now, I remembered vaguely that I had seen a reference to making THD measurements without noise. Dug in and what do you know, it has a sophisticated THD meter in the DSP subsystem. You can instruct it to sum any and all harmonics and derive the distortion total from it!

    Here I have all the harmonics selected but I can click on them individually:

    upload_2018-4-14_16-16-50.png

    Anyway, before the power was lost, I captured one set of measurements. I matched the output level while driving the MDR-V6 to 1 milliwatt per Benchmark test. I then compared both THD and THD+N relative to a 50 ohm dummy load and the 73 ohms of the MDR-V6. Volume on RME ADI-2 DAC was set to 0 dBr and I adjusted the digital samples until I achieved 1 mw (about -32 dBFS):

    RME ADI-2 DAC THD into Sony MDRV6 headphones.png

    We see that THD measurements were pretty necessary as without it, noise dominates and we see the slightest hump with MDR-V6 relative dummy load. We see a 20 dB rise due to Sony load versus dummy resistor.

    Here is the same results from Benchmark:

    upload_2018-4-14_16-41-56.png

    Focusing on the navy line in theirs, the overall distortion is similar to RME ADI-2 at mid to high frequencies. At low frequencies of course, the distortion remains the same and does not rise as it does with RME ADI-2 DAC. What they show happening to the other two amps is what is happening to RME ADI-2 DAC but at about 8 dB lower.
     
    Dro likes this.
  3. amirm

    amirm Founder/Admin CFO (Chief Fun Officer) Staff Member

    Joined:
    Feb 13, 2016
    Messages:
    11,042
    Likes Received:
    2,918
    Location:
    Seattle Area
    Before losing power, I performed another experiment. I connected the grado headphones to the RME and then measured THD while playing 50 Hz and alternatively blocking the back of the headphone versus not. Blocking it reduced distortion by 10 dB! So it seems that this is indeed related to back EMF energy from the driver coming into the output of the amplifier. This is what Mathias is saying is happening. What he does not address is why Benchmark is not sensitive to it, and they are.
     
    Veri and Dro like this.
  4. SIY

    SIY Member

    Joined:
    Apr 6, 2018
    Messages:
    63
    Likes Received:
    35
    Location:
    Chicago-Milwaukee axis
    What does the impedance curve of those phones look like? I'm just not seeing this distortion rise at all with the APx1701 driving the Superlux phones that I have on hand.
     
  5. Blumlein 88

    Blumlein 88 Major Contributor

    Joined:
    Feb 23, 2016
    Messages:
    3,026
    Likes Received:
    1,145
    Superlux phones are too low end to show the difference. They are so low end not only will they not resolve distortion while listening, they won't even do it measuring with an AP unit. :D
     
  6. mindbomb

    mindbomb Active Member

    Joined:
    Nov 11, 2017
    Messages:
    124
    Likes Received:
    58
    I think Benchmark might have been trying to say that the output impedance specification might not reflect actual output impedance at low frequencies, and that's why some amps are more sensitive to it.
     
    Last edited: Apr 15, 2018
    amirm likes this.
  7. Dro

    Dro Member

    Joined:
    Apr 8, 2018
    Messages:
    35
    Likes Received:
    6
    "The main point of the paper is how to make standard op amps stable for capacitive loads, and protect them from oscillating. This, of course, is no issue if you use circuits that have been designed already not to act like this, and are developed specifically for operation with phones."

    I read that as Benchmark use the setup as described by TI to protect their amplifier from misbehaving (which eliminates THD), whereas RME do it differently, so you can still measure the THD fed in by the headphones.

    Maybe somebody who is good with electrical engineering in these types of applications could follow up in the RME forums? They sure seem quite responsive to inquiries.
     
  8. SIY

    SIY Member

    Joined:
    Apr 6, 2018
    Messages:
    63
    Likes Received:
    35
    Location:
    Chicago-Milwaukee axis
    Yes, that must be it.:)
     
  9. Rod

    Rod Active Member

    Joined:
    Mar 24, 2018
    Messages:
    173
    Likes Received:
    22
    That didn't take long. Already its sounding better. Had to turn the bass back down. OMG.
     
  10. Sythrix

    Sythrix Member

    Joined:
    Mar 8, 2018
    Messages:
    68
    Likes Received:
    24
    Is there anything conceivable that could be done post amp to prevent this EMF energy from coming back into the output? Something akin to a one-way street?

    It doesn't necessarily bother me at these levels, but it would be interesting to attempt and compare if a practical solution could be made available.
     
  11. pierre

    pierre Member

    Joined:
    Jul 1, 2017
    Messages:
    14
    Likes Received:
    10
    Location:
    Switzerland
    You could get a rme ufx 2 for 2k and get 8 outputs in 1 box.
     
  12. restorer-john

    restorer-john Active Member

    Joined:
    Mar 1, 2018
    Messages:
    149
    Likes Received:
    98
    Location:
    Gold Coast, Quuensland, Australia
    Some of the back EMF from headphones or speakers is returned via the NFB loop to the front end of the amplifier. Unless your amplifier has zero impedance and zero feedback for the buffer stage, I can't see how this contribution to THD can ever truly disappear. This effect has been well understood with speakers and amplifiers for many years.

    Many amplifier designers in the 1970s researched the effects of speakers and speaker cables and their contribution to THD and the results led to very fast, ultra wideband, monoblocks with very short cables to pursue the lowest THD.

    A fast feed-forward/direct distortion cancelling setup could IMO isolate the back EMF characteristics of the load from the amplifier and let it work in isolation?
     
    Last edited: Apr 15, 2018
  13. SIY

    SIY Member

    Joined:
    Apr 6, 2018
    Messages:
    63
    Likes Received:
    35
    Location:
    Chicago-Milwaukee axis
    Two questions, then:

    1. Why would the back EMF generate harmonics when measured across the transducer? I have not seen this phenomenon when testing (competent) amps with loudspeaker loads. As well, given the moving mass of headphones (low) and the not-spectacularly-high impedance peak at bass resonance, I wouldn't expect that the back EMF would be super high.

    2. Why am I not seeing that same effect with a different amplifier having the same source impedance?

    I'm going to try a few more variations and post results. I'm still unsure if this is an amp, Grado-specific, or a test setup issue and would really like to pin that down. Needless to say, if there's anyone in range who has an RME and/or a set of Grados that we can put on my test bench, the reward will be knowledge and a bottle of something excellent.
     
  14. SIY

    SIY Member

    Joined:
    Apr 6, 2018
    Messages:
    63
    Likes Received:
    35
    Location:
    Chicago-Milwaukee axis
    And a little more fun. I increased the source impedance by putting 2R1 in series with the headphones, then increasing the drive voltage to maintain the same 630mV across the phones. Two second sweep time, just to get the frequency resolution a bit higher. Lo and behold, I did see a greater amount of bass distortion in the electrical waveform, but still far less than Amir was observing. And still, at the same rated source impedance as the RME, the distortion was lower yet.

    FWIW, the distortion was unsurprisingly dominated by 2nd and 3rd order.
     

    Attached Files:

  15. Rod

    Rod Active Member

    Joined:
    Mar 24, 2018
    Messages:
    173
    Likes Received:
    22
    I used these settings from another forum but its not the same headphone that I have. I noticed it sounded smoother right away. But does someone have settings for the HE-400i for E.Q.?

    Hifiman HE-560
    B 1, G +4.0, F 120, Q 0.7, SHELF
    B 2, G +6.0, F 1.90k, Q 1.2, PEAK
    B 3, G -7.0, F 4.0k, Q 3.0, PEAK
    B 4, G -6.0, F 10.5k, Q 3.0, PEAK
     
  16. watchnerd

    watchnerd Major Contributor Beer Hero

    Joined:
    Dec 8, 2016
    Messages:
    3,115
    Likes Received:
    540
    Location:
    Seattle
    I got my ADI-2 Pro at Sweetwater, no issues.

    (I've also bought a lot of other gear from them, too)
     
  17. amirm

    amirm Founder/Admin CFO (Chief Fun Officer) Staff Member

    Joined:
    Feb 13, 2016
    Messages:
    11,042
    Likes Received:
    2,918
    Location:
    Seattle Area
    I don't but remind me in a few days and I will play around and report back. :)
     
    Rod likes this.
  18. DonH56

    DonH56 Major Contributor Patreon Donor Technical Expert

    Joined:
    Mar 15, 2016
    Messages:
    1,435
    Likes Received:
    787
    Location:
    Monument, CO
  19. Dro

    Dro Member

    Joined:
    Apr 8, 2018
    Messages:
    35
    Likes Received:
    6
    @amirm have you measured the noise level of the IEM out with no signal?

    I am fairly certain I can hear a tiny bit of hiss with my Andromeda when no music is playing. I think that is to be expected even with the very good specs of the IEM out. Over 135 dB SPL at 1 Vrms is loud. I listen at -30 dB and due to replay gain I basically never have music go above -5 dBFS.
     
  20. dc655321

    dc655321 Member Patreon Donor

    Joined:
    Mar 4, 2018
    Messages:
    57
    Likes Received:
    14
    The Andromeda are ridiculously sensitive IEMs. When I had mine, I could hear the screen flicker noises from my iphone se when scrolling quickly.
    OTOH, the Andros were completely black when I paired them with the SMSL iDEA.

    Great earphones... sometimes I miss not having them any more :-(
     
Loading...

Share This Page