Cayin RU6 R2R Dongle Review (and Measurements: shit happens)

Discussion in 'IEMs and Portable Gear' started by purr1n, Sep 4, 2022.

  1. Azimuth

    Azimuth FKA rtaylor76, Friend

    Pyrate
    Joined:
    Nov 9, 2015
    Likes Received:
    6,862
    Trophy Points:
    113
    Location:
    Nashville, TN
    Home Page:
    It gets worse and worse.
     
  2. jexby

    jexby Posole Prince

    Staff Member Pyrate Contributor
    Joined:
    Sep 28, 2015
    Likes Received:
    8,146
    Trophy Points:
    113
    Location:
    Exit stage left....
    someone send another RU6 to purrin.
    for only the basic of tests to see if the channel matching and bad waves are consistent.
     
    • Agreed, ditto, +1 Agreed, ditto, +1 x 5
    • Like Like x 2
    • List
    Last edited: Sep 7, 2022
  3. gotflute

    gotflute New

    Joined:
    May 8, 2018
    Likes Received:
    32
    Trophy Points:
    13
    Location:
    College Station and Houton, TX
    <deleted... read what I wrote and it won't contribute to the conversation>
     
    Last edited: Sep 7, 2022
  4. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,934
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Further thoughts:

    Thanks for being mature and level-headed about this. It's entirely possible that I have a bad unit. In fact, I highly suspect that I do because of one channel's lower performance into low impedance 32-ohm loads. Not only are we talking about higher distortion but significantly lower voltage output when compared to a 100k-ohm load. No doubt this will affect the typical use case for the RU6 (e.g., Grados, Mini C/S, Drop Etys, etc.) and likely affected what I heard subjectively.

    However there are certain things that probably are unlikely to be better even if this was a bad unit:
    • The channel delay and it's associated jitter. This is because of the form factor. There is only room to have one resistor ladder. Thus there is no way to avoid multiplexing the L and R channels.
    • Subpar distortion. I bet the RU6 uses 0.10% resistors, which are not anywhere near good enough. Just a few resistors toward the ends of the bell curve of tolerances will screw things up big time. The way only to fix this is to massively increase costs by using expensive precision resistors and/or by measuring/trimming the resistor ladder for each unit.
    I think it's kind of cool that Cayin came up with this idea. However, some things should just never be made. I don't know how an audio engineer can call himself an audio engineer making something like this, unless it was the marketing guy who just made the call: I don't care if from an engineering point of view it's screwy, it's an R-2R dongle, so it will sell!
     
    • Like Like x 17
    • Agreed, ditto, +1 Agreed, ditto, +1 x 4
    • List
    Last edited: Sep 7, 2022
  5. yotacowboy

    yotacowboy McRibs Kind of Guy

    Pyrate Contributor
    Joined:
    Feb 23, 2016
    Likes Received:
    10,845
    Trophy Points:
    113
    Location:
    NOVA
    Home Page:
    Happy to send my unit that's currently on the Debbie Does 4 Dongles tour, once that's done. Or jump in line on the tour, if that's acceptable to the other donglodytes who've already signed up.
     
    • Like Like x 4
    • Epic Epic x 4
    • List
  6. neogeosnk

    neogeosnk Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    1,883
    Trophy Points:
    93
    Location:
    Los Angeles, CA
    Ur review is spot on what I thought about the unit. About two weeks ago I posted that I didn't want to do reviews of hardware I didn't like and the RU6 was one of the items I didn't want to shit on. But this is maybe why I "should" be posting my impressions.
     
  7. Clemmaster

    Clemmaster Friend

    Pyrate Contributor
    Joined:
    Sep 28, 2015
    Likes Received:
    3,271
    Trophy Points:
    113
    HiBy RS2 next?
     
  8. penguins

    penguins Friend, formerly known as fp627

    Pyrate
    Joined:
    Sep 26, 2018
    Likes Received:
    3,467
    Trophy Points:
    113
    Location:
    SouthernCA
    Missed the lag the first time I skimmed the measurements but after looking more closely and at the other measurements this time, agreed, I doubt this is a case of "unit got shaken" but more of a "cut some corners b/c circuit too big".
     
  9. limesoft

    limesoft Friend

    Pyrate
    Joined:
    Oct 4, 2016
    Likes Received:
    489
    Trophy Points:
    63
    Location:
    London
    These beautiful measurements show me how closely RU6 comes to a vinyl rig - it simulates all the good stuff like volume variability, distortion, warble, channel offset - generating beautiful sound. This dongle is purposefully designed to measure and sound like this, incredible engineering.
     
    • Like Like x 10
    • Epic Epic x 8
    • Respectfully Disagree Respectfully Disagree x 1
    • List
  10. zottel

    zottel Friend

    Pyrate Contributor
    Joined:
    Jan 22, 2022
    Likes Received:
    1,275
    Trophy Points:
    93
    Location:
    Franconia, Germany
    Here's an image I found on https://cayin.com/produkt/cayin-ru6-usb-dac-amp-dongle/ (the website is in German, but scroll down, there are more promotional images in English like the one I'm attaching here, including one with a functional diagram of the dongle):

    [​IMG]

    So, yes they're using 0.1% resistors. But, if I understand correctly, there is no multiplexing of the R/L channels.
     
  11. limesoft

    limesoft Friend

    Pyrate
    Joined:
    Oct 4, 2016
    Likes Received:
    489
    Trophy Points:
    63
    Location:
    London
    Bit disturbing that the Hiby RS6 also uses 0.1% resistors at a much higher price. It certainly sounds better than the RU6 in every sense, perhaps due it's DARWIN architecture that has linearity compensation. Not sure how it measures, I'd be surprised if it's similar to the RU6.

    I can't help but wonder if better performance could have been had with better resistors... or if linearity compensation offsets some of the need for high accuracy resistors. Would be good if someone could explain in layman's terms some of the tech possible around the R2R circuitry to make it more performant (like Holo May etc)
     
  12. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,934
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    That advertisement is just embarrassing. "High-precision" is a relative term, but 0.10% (assuming it's even that or they didn't get a bad supplier using bogus parts) isn't sufficient for a DAC ladder. As I mentioned, Soekris offers a 0.05% for their entry level dam1021.

    What's the point of having a 24-bit resistor ladder when the RU6 only achieves 10-12 bits of linearity?
    upload_2022-9-9_7-36-55.png
     
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,934
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Take this example for this triangle wave crawl of the 9 least significant bits. I roughly drew what the signal is supposed to be, a triangle wave. What we can see is that the signal output is not accurate, does not track, what the digital codeword says the output is supposed to be. In this case, the error is consistent from wave to wave, so it's not noise.

    There are two ways to fix this: 1) measure the output and add trim resistors so the output is correct; 2) correct via the filter / upsampling, where we can use dither to achieve error adjustment / trim values under 1-bit.

    upload_2022-9-9_7-39-56.png
     
  14. limesoft

    limesoft Friend

    Pyrate
    Joined:
    Oct 4, 2016
    Likes Received:
    489
    Trophy Points:
    63
    Location:
    London
    Gotcha.. is linearity compensation essentially this? "measure the output and add trim resistors so the output is correct"

    I don't understand how Hiby RS6 can achieve the level of sound quality with 0.1% resistors - subjectively, as I don't know how it measures, just know that it doesn't sound like something's a bit wrong unlike the RS6 - no treble nasties, etc - it's not as smooth sounding as a well measuring Apogee Element but it's not far off, and it sounds resolving, natural, and generally correct to my ears
     
  15. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    Unfortunately that’s most products. dudes love shitty cut down versions of 80s and early 90s r2r dacs like this and 80s crap circuits with junk Chinese caps like rme.
     
    • Like Like x 1
    • Agreed, ditto, +1 Agreed, ditto, +1 x 1
    • List
  16. goodvibes

    goodvibes Facebook Friend

    Joined:
    Jun 21, 2017
    Likes Received:
    35
    Trophy Points:
    30
    Location:
    Chicago
    It's weird. When digital finally trumps analog, people go back to vinyl and when delta/sigma's are better or at least as good as R2R DACs, people go back to R2Rs, often without up/oversampling.

    Don't get me wrong about vinyl, there's downloads that will never sound as good as an original old pressing but recent releases etc?

    Same goes for R2R. It can be great but so can D/S so it's about execution. The type of prominent errors become meaningless once below an audible threshold. For me, the latest ESS DACs done right bridge that character gap nicely.

    A product like this, regardless of cost, always puts me off of a brand willing to sell it. Better killing it instead of release.
     
    • Agreed, ditto, +1 Agreed, ditto, +1 x 4
    • Like Like x 1
    • List
    Last edited: Sep 18, 2022
  17. limesoft

    limesoft Friend

    Pyrate
    Joined:
    Oct 4, 2016
    Likes Received:
    489
    Trophy Points:
    63
    Location:
    London
    ESS dacs nowadays are good, but I find it takes considerable expertise to whisper them to remove that ESS glare/brightness where even residual traces give a certain ESS trait which to my ears sounds little plasticky/fake and fatiguing (and no, we don't have measurements sophisticated enough that can capture this and it's not my mind playing tricks on me)

    I've only heard a few companies that can make ESS sound non ESS that's natural, direct and pleasing, it seems like a very difficult thing to do - one is Apogee (Element series and probably their other stuff too). I'll just name all the others ESSes that i've tried that can't quite get rid of this ESS flavor: Fiio M11 Plus, Earmen Sparrow/Colibri, Woo WA8, Matrix Mini - i, and many others. RME ADI2 FS (ess) does seem to have a whispered one but it has it's own set of other problems due to junky caps I assume, adding timbre issues.

    * There's an auditory satisfaction and beauty with a good vinyl setup; organic, natural, easygoing, direct, dynamic, you name it - a more direct connection to the music without the ear fatigue - even a cheap one can't be easily matched by digital sources in this respect for musical pleasure. This comes more apparent after spending years listening back and forth many different sources, expensive and cheap: vinyl > r2r dac > d/s dac> r2r dac > vinyl.

    Which is why I'd advise to anyone who wants the best sound quality for their money and is ready to sacrifice a little practicality, even with limited funds, to get vinyl rig - even a cheap one like a starter Project Debut deck with OM10/20 + ifi iphono, simply runs circles around most digital sources - and it's not just because vinyl is produced better.

    A good R2R can come close as it sort of emulates vinyl - we're seeing bit of a comeback to R2R as people are probably realising the d/s dac fatigue and lack of musicality is real, it takes a good few years to re-visit trends and conclude that new is not always better. I'm hoping we will see more (improved) R2R in portable and value sources - with more sophisticated feedback networks, dsp, clocking and resistors it is possible to get pleasing results. Look at the new iPhones for example - tiny optical sensors, but with advanced and multi shot processing they provide image quality that's 80% comparable to large sensor cameras - if R2R can develop in a similar manner it might be possible to get good results even with 0.1% resistors - I think something like the Hiby RS6 might demonstrate this with what they call DARWIN for their topology.. I'm not 100% sure it's a super high performer but it certainly leaves something like the RU6 in the dust to my ears, yet they both have 0.1% resistors.


    * Sorry unrelated ramble, I'm getting old
     
    Last edited: Sep 19, 2022
  18. Ash1412

    Ash1412 Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    667
    Trophy Points:
    93
    Be careful with mistaking popularity for superiority. History has shown that it's not the highest quality tech that survives but the most accessible and cheap to manufacture. Dynamics killed out stats and planars due to ease of driving and easier cheap manufacturing. 35mm film can resolve 4K/7MP and more, but that didn't stop everyone from moving to digital even when sensors were only 2MP (not that sensors are cheap, but cost per photo is pretty much nothing vs film). LCDs were worse than plasmas image-wise when they started out but more efficient, lighter, smaller and could scale in resolution.

    @limesoft also used the multi-shot processing example, a great modern example, since the technology emulates more light/a bigger sensor by aligning and merging frames, akin to D/S using modulation to emulate the precision of R2Rs. This only became viable because small imaging sensors and phone processing power have been developing at exponential speeds (thanks Apple), enabling faster readout sensors and algorithms to obtain and process 5-15 photos into 1 seamlessly. It's not been implemented in cameras cause the readout speeds of older sensor tech can't hold up, and frankly the dedicated camera industry doesn't have the engineering resources. However, with phones, you still have to use thin and small lenses which then become the bottleneck (try Google or Apple RAW and see how much diffraction there actually is).

    Here are two talks by ESS engineers who work on the chips about D/S errors being audible (to certain folks) at seemingly impossibly low thresholds. They absolutely do exist as engineering challenges and are describable if not measurable with a bit of ingenuity:



    All engineering solutions should have their place. My analogy of that would be DS as a good soup/broth/sauce-based meal like a comfortable hotpot and R2R as your favorite deep-fried dish with tons of flavor. Portable resistor ladder DAC seems as good an idea as ordering boxed fried food. Difficult to keep from turning soggy and gross.
     
    • Like Like x 3
    • Epic Epic x 2
    • List
    Last edited: Sep 19, 2022
  19. goodvibes

    goodvibes Facebook Friend

    Joined:
    Jun 21, 2017
    Likes Received:
    35
    Trophy Points:
    30
    Location:
    Chicago
    I honestly still own Quad ESLs and a 60" Pioneer Elite Plasma.:D LP12 for analog.
     
  20. jexby

    jexby Posole Prince

    Staff Member Pyrate Contributor
    Joined:
    Sep 28, 2015
    Likes Received:
    8,146
    Trophy Points:
    113
    Location:
    Exit stage left....

Share This Page