Ec-Designs Multi-Bit DAC

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by murphythecat, Dec 2, 2018.

  1. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
    I really do wonder if the issue with optical isn't that it's fundamentally flawed, but that they just refuse to advance the technology.

    High-bandwidth network connections run on fiber optics and it just seems strange to me that they would use it if it had the problems that TOSLINK has.
     
  2. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,775
    Trophy Points:
    113
    Location:
    Padre Island CC TX
  3. Clemmaster

    Clemmaster Friend

    Pyrate Contributor
    Joined:
    Sep 28, 2015
    Likes Received:
    3,268
    Trophy Points:
    113
    Could also mod it to use ST Fiber pretty easily using a ODL Series II receiver::1361AAC RX
    It requires more power than a regular TOSLink receiver, but that could be modded, too.

    The 1261AAC TX / 1361AAC RX is what Theta used to use, back in the days.
     
  4. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Home of Jiffy Mix
    Hi Soren,

    I do bow to your expertise, but could you explain the difference a little more.

    I was relying on Wikipedia (not always correct) and, in the SPDIF article, they said that the optical and coax signal were same, that is, 0v low, 0.5 - 0.6V high. Since they say that nominal high level for TTL 5V high, obviously they are incorrect in the SPDIF article because of voltage levels.

    In other words, to modify an optical input to coax, you would need to amplify the voltage levels. Do I understand this correctly?

    Jac
     
  5. soekris

    soekris MOT - Soekris Engineering

    Pyrate
    Joined:
    Mar 25, 2016
    Likes Received:
    954
    Trophy Points:
    93
    Location:
    Denmark
    The SPDIF receivers I know and use all have digital levels out, depending on their power supplies. To use Coax you need to amplify the signal to digital levels, you can google "spdif receiver schematic" and find those circuits, most of them use a couple of cmos gates.... Note that some SPDIF chip receivers accept both signal levels, check the datasheet for the receiver you intend to feed. A isolation transformer like a Murata DA101 should also be added.
    The data protocols are the same.
     
  6. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Home of Jiffy Mix
    A small comparison between the EC Designs MOS16 and the Metrum Flint (DAC1 version). They make a reasonable comparison because they are both NOS multi bit dacs and have similar enough pricing. Here in the US, I paid about $550 including shipping for the Flint and about $350 for the MOS16.

    Before I get into the dacs, the equipment that I used to compare the dacs consists of Linkwitz Lxmini speakers with the Nelson Pass analog crossover and 4 channels of My_Ref Fremen Edition amps. The source is an Rpi3 using Squeezelite software and feeding a DIYnHK USB to SPDIF converter. I did my best to match the listening levels, although I didn't have a meter available. Both the MOS16 and the Flint were fed an optical signal to keep the comparison equal. I also spent some time listening to the Flint using coax. That helped me understand the contribution of the optical input.

    EC Designs MOS16 dac

    My first surprise was how small the MOS16 is. I know, I read the dimensions in the factsheet, but pictures make it look bigger somehow.

    I am used to leaving my dac turned on. Something about digital clocks having variable jitter as they warm up. With the MOS16, we are running on battery and you have to turn it off to charge the battery. The tricky part is that the MOS16 puts out a pretty good spike when you turn it on or off. It becomes necessary to remember to turn off your amp before you turn off the MOS16. And vice-versa. This isn't a criticism, just information you need to know when using this dac.

    Dac Comparison

    I found my results interesting after reading again Purr1n's comments on the RDAC and MOS16.

    Flint
    • Nicely balanced frequency response. Actually, both of these dacs seem pretty neutral to me.

    • Stage is up front, close to the stage feeling. Stage width is good, beyond the speakers and good depth.

    • Inner dynamics are pretty good. You can hear when the fiddle player emphasizes a note with his bow. You can hear some difference in piano note loudness and a medium level of harmonic decay as the string rings.

    • Detail is pretty good although there is a little high frequency haze sitting on top of everything.

    • While its not the best high frequency sparkle and clarity, I kind of expect a little top end roll off in a NOS dac.

    • There is a minor sense of compression, both for high/low frequencies and dynamics.
    MOS16
    • frequency response matches the Flint.
    • Stage is further back with equivalent width, but a little less depth.

    • The battery power seems to provide a very black background. There is an overall sense that this is a quiet dac.

    • Inner dynamics suffer somewhat. I am stating too strongly, but it feels like all the notes are at the same loudness.

    • There is a sense of compression at both low and higher frequencies, in particular the highs.

    • Compared to the Flint, resolution suffers because the compressed sound muffles the soft sounds. The harmonics of string decay is clearly compromised.

    • High frequencies are further rolled off compared to the Flint.

    • Overall, it is a very pleasant dac to listen to but suffers in direct comparison to a little bit better dac.
    Optical vs Coax

    The Flint gave me the opportunity to directly compare coax with optical. Just a push of a button and the mode is changed. That made me curious, how much of the difference is due to the optical input. This is particularly important to me because I have figured out a way to feed coax to the MOS16 with minimum modification.

    Going from optical to coax was a clear improvement. Inner dynamics and detail improved noticeably. Now you could hear the attack of the bow on the strings or the piano note. You could also hear singer breath and multiple harmonics of a string ring out clearly. I believe the term is “more air”. You could also separate and locate softer instruments in the background. That helped stabilize the image and give the musical experience a kick toward reality.

    I look forward to seeing what happens with coax feeding the MOS16. The optical Flint was about in the middle between the coax Flint and the MOS16. The hope is that a coax MOS16 would be where the optical Flint is today or better.

    PS No funny extra wires inside this one. Murphythecat must have gotten a prototype.
     
    Last edited: Mar 21, 2019
  7. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    It's a DIR9001. I tried to get some really detailed pictures, if you're interested. I was curious myself what it would take to swap in a coax input. AFAIK, the DIR9001 is fairly friendly towards both types of inputs.

    I highly suspect that toslink is holding back the DAC's true performance. I've read some feedback that the ECdesigns player makes a big difference in how the MOS16 sounds, but that's a tall order to pay for a device that is toslink only, might not work with other DACs, and seems to have such a simple method of managing your music files, to the point it becomes an extreme hassle. Seriously, I can have 99 "CDs" on a USB stick with up to 9999 songs, all of which much be numbered and in WAV format? Give me a break...

    Meanwhile, the Eitr or DigiOne Signature are both cheap (and cheaper than the ECdesigns player), already sound fantastic, and utilize relatively common, proper ways of managing a music library.

    That said, I'd be open to try the ECdesigns player to see if it really made that big of a difference, were someone to send it to me while I am borrowing the MOS16. Maybe I'd change my tune, but...my magic 8 ball says probably not. I'm also still just mad that I can't use coax, because I have a great coax source, and toslink blows.

    The best I can do right now is hook my coax source up to a $15 coax->toslink converter (using a LPSU I had lying around, to remove poor variables from the equation)...which I have my doubts is really up to par. Nor do I have a good USB port for the ECdesigns toslink stick, mostly because trying to make USB good is about as worthwhile as trying to make your shit edible. (Barring a few exceptions, but I'm commenting on the extreme nervosa around USB cleanup/reclocker devices and the like.)

    I'm also iffy on the battery. I do have to imagine it helps the sound. But I've heard and measured plenty of squeaky clean DACs without having to rely on batteries. Please, give me a damn power jack and let me pick my own PSU if I really want to go crazy with clean power. This whole having to charge it thing (which you can't use as a power buffer while listening, nooo), even with 100 hours of play time, is just another thing I DON'T want to have to worry about in my life.

    Sound impressions in the DAC blind test thread coming up soon. :)
     
  8. Avinunca

    Avinunca New

    Joined:
    Feb 22, 2019
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    EU
    I have the UPL16 and MOS16. Great combination with very unfatiguing and detailed sound. I don't find the usb stick storage an issue. I am also impressed with the Dac connected to my MacBook via the UTOS. Playing Kate Bush's live album I was very impressed with how alive and solid she sounded. The Dac is better than my Mani MB but have nothing else for comparison. Charging the batteries is no problem. For the money I think this is a winning combination. The only thing lacking for me is a second optical input on the Dac but John Brown did suggest https://www.amazon.com/Digital-Tosl...d=1553686050&s=gateway&sr=8-8#customerReviews
     
  9. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Home of Jiffy Mix
    I am waiting on the mail for the bits to convert the MOS16 to coax. My approach isn't for everyone because the mod kills the warranty.

    Similar to Hands suggestion of the DigiOne, I am using the Pi2Design 502 DAC that has BNC coax, toslink, and AES output (via TRS). I am going AES (TRS/XLR cable) to a $30 AES to SPDIF converter (Canare BCJ-XJ-TRC) to coax (BNC/RCA cable) to an RCA jack replacing the toslink receiver. The advantage of this approach is that the AES signal is 4.5 - 4.75V which is what the receiver chip is looking for from the toslink. I agree with Hands that most of the receiver chips are OK with 0.5V from SPDIF, but I felt like having the board see minimum change was worthwhile.

    EC Designs is said to have chosen Toslink for isolation. Ironically, the AES to SPDIF converter gives me isolation through a transformer.
     
  10. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,775
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Looking forward to the experiment!
     
  11. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    FYI from sticking a DMM in the MOS16: VCC to ground is 2.9V off the optical receiver. Optical receiver output pin going to RXIN on DIR9001 is 1.4V to ground. These will probably be a bit higher with a fully charged battery.

    The DIR9001 specifies it accepts up to a 5V TTL signal on RXIN. That’s the max, not the requirement. Clearly the DAC isn’t going that high.

    I am having a hard time figuring out with certainty if the DIR9001 works with a usual 0.5v SPDIF signal or if it needs a higher voltage. One note in the data sheet says an external amplifier is needed on RXIN if used with a coaxial transmission line. But I’m not well versed enough in all this, and reading electrical specs on data sheets, to know if that means what I think (i.e. you need higher input voltage for DIR9001 than SPDIF provides by default).
     
  12. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    IT'S ALIVE! (Before anyone gets scared, I did NOT modify the MOS16 in any way. For testing, I used a light touch with some wiring on the toslink receiver input pins/through holes to verify.)


    Here is one way you can get your MOS16/24 working with SPDIF over coax via a TTL signal. It is not theoretically perfect in all ways.

    You will need:

    - 5V PSU (I imagine you could find a way to piggyback this with the USB 5V charging). Try to get something decent and clean, but don't go crazy here.

    - The "independent SPDIF board" as seen on this page:

    https://translate.googleusercontent...700253&usg=ALkJrhj4bMy85PuavTI5hRLslWytzFKdnQ

    Which you can buy here, as one example:

    https://www.ebay.com/itm/SPDIF-TTL-...m4b4487c3c0:g:sAoAAOSw6Tdbm3aH&frcectupt=true


    - Coax jacks, wires, etc etc. You'll need to know how to solder.

    - DC jack + wire for the board, unless you want to hard wire the PSU.

    - A case if you want it to look official. Or, in theory, you could probably find a way to install it inside the DAC along side the battery were you to make a custom plate for both. Will touch on that in a bit.


    What you need to do from there:

    1. Wire up the 5V DC jack (or hard wire PSU). Plug it in and make sure the optical jack light comes on.

    2. Wire up a coax jack of your choice into "J1" (bottom left). Per the schematic, this input has a 75 ohm resistor on it, so I would recommend a 75 ohm BNC jack if possible.

    Note: All the other inputs/outputs I mention I don't believe adhere to the 75 ohm spec, so it probably matters less. (As I understand it, TTL, since it's at a higher voltage, doesn't necessarily require adhering to a tight impedance spec? Either way, those of you OCD about impedance matching will have more work to do across the board.)

    3. Wire up a coax jack to J3 on the other side of the board.

    4. On the MOS16/24, if looking from the top down with the front of the DAC facing you, you'll see 3 pins/through holes by the toslink receiver. In order from left to right, these are IN, GROUND, VCC (with IN being the one closest to, say, the battery).

    5. Wire up a coax jack to the IN and GROUND through holes. Don't do anything with VCC. You don't need to remove the toslink receiver, but know that if you don't, you'll need to find a new place for the coax jack. Then again, you'll have to find a jack that either fits in the PCB where the toslink receiver used to sit or get creating with mounting it.

    Note: You can also use the OUT/GND pins from the toslink transmitter on the converter board. Not sure if that would change performance/sound.

    6. Plug it in, and that should be that.

    The little SPDIF converter board I linked takes an SPDIF input and converts to AES, SPDIF, TTL, and Toslink. Pretty handy, actually! Just supply your own 5V PSU and other parts.

    I was lucky enough to have this random part around for some project I never actually ended up starting a couple years ago.

    I obviously haven't done anything other that test and see if it produces music, which it does! No idea if the sound quality changes.


    One variant of this, which I briefly mentioned, was installing this converter board inside the DAC. While that would defeat the utility of having an isolated digital audio converter box, maybe you'd prefer the seamless look and having a very short TTL wire run?

    Instead of installing an output jack on the converter board, you could install the board internally and use very short wires to IN/GND. The SPDIF input jack would then be considered the DAC's input, so you'd still need to figure out a solution for that. You would most likely need to modify the plastic battery mounting plate, or make a new one, to hold both the board and the battery.

    Since the MOS16 uses 5V from USB to charge, I believe you could find a way to piggyback off that. I imagine you could set it up so the board draws power regardless of the front switch position. You could go full blown crazy and start sorting out ways to use an LPSU for this purpose.

    It's more work overall that way, but a path you could take nonetheless.
     
  13. jexby

    jexby Posole Prince

    Staff Member Pyrate Contributor
    Joined:
    Sep 28, 2015
    Likes Received:
    8,145
    Trophy Points:
    113
    Location:
    Exit stage left....
    dude is this a DIY project or a product humans can actually use?
     
  14. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    Doesn't DIY imply it's a product a human can actually use once they've done it themselves?

    How many of @MisterRogers DIY endeavors have benefited you personally? ;)

    Plus, ain't nothing wrong with tinkering for the sake of learning on a cold Sunday!
     
  15. jexby

    jexby Posole Prince

    Staff Member Pyrate Contributor
    Joined:
    Sep 28, 2015
    Likes Received:
    8,145
    Trophy Points:
    113
    Location:
    Exit stage left....
    1. guessing mods to a non DIY Super 7 may be relevant.
    2. sure, learning is always good. we'll just count spdif / toslink dicking around = to USB input tomfoolery.
     
  16. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    Blasphemy! SPDIF is inherently good. USB is inherently evil.
     
    Last edited: Apr 2, 2019
  17. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Home of Jiffy Mix
    Hands,

    Thank you for this. I was just starting to mess around with my voltmeter when I saw your post. I just confirmed your measurements and will move forward.
     
  18. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Home of Jiffy Mix
    I measured 2.9V also for VCC to ground on a fully charged battery. Apparently, they are limiting the voltage to the optical receiver.

    I found some different information in the datasheet for the DIR9001. The absolute maximum supply voltage for VCC (analog) and VDD (digital) is listed as 4 V. The maximum input logic high voltage is listed as VDD or a max of 4 volts. However, the Digital Input Voltage is listed as -0.3 to +6.5V, so a 5V input is tolerated on the digital input side.

    I am pretty sure that a 0.5V SPDIF coax signal won't work without the external amplifier. The minimum voltage to get a "high logic level" is 0.7 VDD. With VDD at the minimum 2.7V, you would need 1.9V to recognize a high logic signal.

    Its clear that EC Designs is using a near minimum input voltage, probably to save battery. I can also see why they chose the DIR9001. Its typical current requirement when on is 6 mA. Compare that to the AK4113 that Schiit uses for SPDIF input which uses a typical 26 mA. I guess when you start out to design something for battery usage, you make different choices than a wall plug power source.
     
  19. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    I take it you'll be using the XLR->BNC Canare adapter without the -10dB padding to ensure you get a high enough signal level?

    I spent the last couple nights doing some listening with the SPDIF converter method outlined above (which, BTW, retains the 2.9V VCC and has RXIN over TTL at 1.5V). Again, nothing modified on the DAC itself, just had the IN and GND wires seated in the through holes. 24 awg solid copper fits fine.

    I must say, it sounds considerably better than what I heard over the toslink methods I've tried thus far. I'll have to see what other toslink sources I have as well as see how the UTOS sounds when @murphythecat ships it to me. Now to scrabble together a decent USB port for the UTOS to make sure I'm not muddying the waters further.

    Again, I do not recommend anyone go crazy with power supplies, but there was a sizable reduction in 60Hz and harmonics in DAC measurements when powering my SPDIF->TTL converter with a cheap LPSU vs a generic switcher. Obviously the DAC itself does not exhibit 60Hz noise with the battery power and toslink, so it's an angle to consider if you want to use coax to send a digital signal to the DAC. It will carry some (likely inaudible) noise with it.
     
  20. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Home of Jiffy Mix
    Exactly.

    Some interesting early results. I wired my RCA the same as suggested above. At the moment, its hanging out the back of the MOS16. I did it this way in hopes that I could listen to both the optical and coax by just changing the cable at the dac.

    I am using the 502 DAC as a SPDIF/AES source. The 502 DAC has a jumper that controls the voltage of the coax output. Apparently, there is a AES unbalanced spec where the signal level is about 1 volt. At 1 volt with the coax connected directly between the 502 DAC and the MOS 16, the MOS 16 woke up, showed the 44 kHz sampling rate, and played music. With the output dropped to the approximately 0.5 volt SPDIF coax level, the MOS 16 shows two dashes in sample rate display and doesn't play music. From the specs, I would have thought neither would have played, but apparently, it less sensitive on the low end than shown in the specs.

    Moving on to the AES balanced output, plus impedance converter, I found the same issue as the low voltage. The MOS16 shows two dashes in the sample rate window and doesn't play music. I had previously tried the same RPi/502 DAC/ AES converter in my main system feeding my Onyx dac in coax. No problem, music played nicely. Pi2Design tells me that the 502 DAC puts out about 4.5V through the AES output and I used a converter that doesn't have the -10 dB pad. So it appears that the MOS16 is very sensitive to the signal that feeds it.

    I was able to confirm that I can play music through either the optical or coax cable (@ the 1 volt level) by simply changing cables. If in the end, I decide to keep the coax cable input, it looks like the RCA will fit in the same space as the optical receiver. I will just need to add a patch panel to hold the RCA properly.

    I will spend a little time in the next few days comparing the MOS116 through coax and optical. I will also listen to the MOS16 coax compared to the Flint coax.

    Anybody need an AES (XLR) to SPDIF (BNC) impedance converter? It doesn't look like I'm going to use it on this project. ;-)
     

Share This Page