Oh shit! SINAD can suck my ****s

Discussion in 'Measurement Techniques Discussion' started by Marvey, Jul 2, 2021.

  1. Hartono

    Hartono New

    Joined:
    Jul 8, 2021
    Likes Received:
    2
    Trophy Points:
    3
    Location:
    Currently in Singapore
    Does these later test done with only one driver connected ? or both L&R channel driver connected to the amp ?
     
  2. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    Both. L and R signals and measurements are run separately to avoid effects of crosstalk.
     
  3. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    This is mounted on my head (not free-air). Fs is around 73Hz. Even if free-air, Fs doesn't change much, although bump goes down a bit. Chalk it up to production variances and silent revisions. It's all close enough. 6H choke - haha.
    upload_2021-7-14_9-18-23.png
     
  4. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,652
    Trophy Points:
    113
    Location:
    On planet
    Version 3, will implement this one.
    HD650 simulator v3.png

    Correlation between HD650 LCR Z measurement and the circuit simulation:
    HD650 LCR Z vs Sim Z v3.png

    I have two HD650 headphones and the LCR Z measurement is very close between them.
     
    • Like Like x 8
    • Epic Epic x 1
    • List
  5. Hartono

    Hartono New

    Joined:
    Jul 8, 2021
    Likes Received:
    2
    Trophy Points:
    3
    Location:
    Currently in Singapore
    if both driver are connected to the amp during the whole procedure, I think the situation I described in post #116 is still happening here, albeit at much reduced level because the shared (L&R) return ground wire from the Y splitter is much shorter.

    Not to derail or not respect the original intent of these tests, just trying to get to the real cause here. Really appreciate all the effort for the testing.

    Maybe it is not back EMF but rather the second channel amp is interacting and cause this distortion, if that so then all 3 pin non transformer/balanced stereo headphone are affected.

    The only way to be sure is to have only one driver channel connected during the testing.
     
  6. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    Now that I feel I can get consistent results:
    1. I'll give it shot with only one driver connected and driven.
    2. Will try other Y connectors to see if there is any difference.
     
  7. NekoAudio

    NekoAudio Acquaintance

    Joined:
    Apr 18, 2016
    Likes Received:
    50
    Trophy Points:
    18
    Location:
    San Jose, California
    Home Page:
    I guess you'd need to have the other channel's wiring completely removed from the equation, meaning a custom-made cable?
     
  8. nobodyNoot

    nobodyNoot New

    Joined:
    Aug 30, 2020
    Likes Received:
    6
    Trophy Points:
    3
    Location:
    Eastern EU block
    Home Page:
    For anyone who wants to quickly double-check their math on these filters, here's a website I found a while ago with a whole bunch of useful tools:
    http://sim.okawa-denshi.jp/en/

    Good luck!
     
  9. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    Thanks for link!

    I don't know if you saw @atomicbob's post but he already validated and tweaked the RLC-RL-R circuit in SPICE.

    LTSPICE does circuit sims. The only downside is a crappy UI which hasn't changed since the days of MS-DOS.
     
  10. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,652
    Trophy Points:
    113
    Location:
    On planet
    breadboard proto validation

    HD650 impedance measurement with Dayton Audio Test System v3
    01 20210702 HD650 impedance DATS L.png

    LT Spice simulation for HD650 impedance simulator
    02 HD650 simulator v3a.png

    HD650 impedance simulator breadboard proto measurement with Dayton Audio Test System v3
    03 HD650 simulator proto.png

    measurement setup
    04 IMG_2907_small.jpg

    Closeup of HD650 impedance simulator breadboard proto
    05 IMG_2909_small.jpg

    Next step is to create a two channel hard proto.
     
    • Like Like x 8
    • Epic Epic x 4
    • List
  11. wormcycle

    wormcycle Friend

    Pyrate
    Joined:
    Aug 13, 2016
    Likes Received:
    1,506
    Trophy Points:
    113
    Location:
    Toronto, ON, Canada
    I wonder what people way more knowledgeable think about this video. I do not know if he gets everything right but it is well made, and I definitely learnt from it
     
  12. Ox Cart

    Ox Cart Facebook Friend

    Joined:
    Mar 13, 2018
    Likes Received:
    202
    Trophy Points:
    33
    Location:
    Austin, TX
    I didn't find the argument convincing. As i understand it, the idea is that not all distortion is created equal, so boiling it all down to one number misses details about the distortion profile. The video then gives the example of -50dB distortion at higher order harmonics. The problem is, with super high SINAD, none of the higher order harmonics is anywhere near -50db, so it's not necessarily audible.
     
  13. wric01

    wric01 New

    Joined:
    Nov 18, 2021
    Likes Received:
    9
    Trophy Points:
    3
    Location:
    san francisco
    Right on, anything to Schitt on Audioscience is good for the audio community as a whole. Letting measured tools to do the listening instead of ears, everyone of those hearing impaired in audioscience refuse to use ears to measure.
     
  14. Resolve

    Resolve MOT: Headphones.com

    Joined:
    Oct 15, 2017
    Likes Received:
    128
    Trophy Points:
    43
    Location:
    Vancouver
    Hey, just saw this. If you check out the last section there it explains it. What you said is true if we're talking about just THD. But SINAD is THD+N, so it matters if the device is noise dominated or distortion dominated, and our critique there is that the threshold for 'good enough' is arbitrary - like there are situations where a fairly high score would still have audible undesirable results for the headphones you're trying to drive. So that's kind of why in my view it's pass/fail for noise, and highly dependent on the headphones you're using. I think someone mentioned that realistically there's probably a limit, given the kinds of headphones most people are trying to drive, and maybe we don't need to worry about the edge cases as much, and that's probably also true, but then it's like saying "here's our threshold for versatility", which is going to be different, and likely far more specific for each person anyway.
     
    Last edited: Jan 18, 2022
  15. Ox Cart

    Ox Cart Facebook Friend

    Joined:
    Mar 13, 2018
    Likes Received:
    202
    Trophy Points:
    33
    Location:
    Austin, TX
    Sorry, I'm not sure I follow. I was just critiquing the example of a -50dB higher order harmonic. While that can be audible in your demonstration, a 110dB SINAD device would at most have a -110dB high harmonic, which wouldn't be audible in such a demonstration. Where does noise come into it?

    P.S. I enjoy your videos and appreciate that you approach the hobby with inquisitiveness and enthusiasm.
     
  16. Ox Cart

    Ox Cart Facebook Friend

    Joined:
    Mar 13, 2018
    Likes Received:
    202
    Trophy Points:
    33
    Location:
    Austin, TX
    @Resolve to put it more precisely, I think you convincingly argue that

    1. THD+N is not a sufficient metric to properly compare low SINAD devices

    2. Better SINAD isn't necessarily going to provide audible improvement if the lower SINAD device is already audibly good enough

    However, I don't see how two high SINAD devices would be distinguishable based on noise or distortion profile, since at something like 110dB SINAD, neither noise nor any harmonics will be audible(?).
     
  17. Resolve

    Resolve MOT: Headphones.com

    Joined:
    Oct 15, 2017
    Likes Received:
    128
    Trophy Points:
    43
    Location:
    Vancouver
    Yeah so whether or not there are audible unwanted products in a device with a score of 110dB SINAD will depend on how sensitive the headphones are and whether or not it's noise dominated - oh and I suppose also the room masking, but yeah with highly sensitive headphones or IEMs, it's possible.

    But again this is where I imagine if we limited ourselves to certain headphones you could find a safe threshold there. But that's what I meant earlier that it becomes an arbitrary measure of versatility.
     
  18. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,772
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    So I have a question, sort of a trick, but really good question for you (and everyone else): what exactly do you mean by "110db SINAD" device?

    When you say "110db SINAD", what would be the stimulus signal's frequency, voltage level, load, and bandwidth? Also would or should we specify the gain of the device under test?
     
  19. Ox Cart

    Ox Cart Facebook Friend

    Joined:
    Mar 13, 2018
    Likes Received:
    202
    Trophy Points:
    33
    Location:
    Austin, TX
    Okay, I think I see what you're saying. The way that I would state it is that what matters is measurements at a realistic output level for a particular transducer. In practice, when driving a sensitive IEM, SINAD won't be anywhere near 110dB, at which point noise floor can become an issue, so it's useful to understand the noise performance at different levels.
     
  20. Ox Cart

    Ox Cart Facebook Friend

    Joined:
    Mar 13, 2018
    Likes Received:
    202
    Trophy Points:
    33
    Location:
    Austin, TX
    All good questions, and I guess the implicit assumption is that something with amazing SINAD at 2Vrms at 1KHz isn't necessarily going to have amazing SINAD across the audible spectrum at realistic output levels (depending on transducer). At that point, noise performance and distortion profile become more important again.

    EDIT - One honest question I have is this. Of those amps that have amazing SINAD at 2Vrms, are there any whose distortion is worse at lower levels, and can we even know given that noise is also worse at lower levels?
     

Share This Page