USB to simultaneous data (or I2S) board

Now it is decided, we will release this project. At this moment I brought the PCB layout to a close, and we will complete the design, produce required quantity, and ship all the orders in the late November or the first half of December.

The goals of this design remain as explained in the original topic. So, this board is essentially a USB decoding stage, which also provides clock signal for D/A conversion. It means that at one side it connects to PC via USB port and, by operating as asynchronous (master) USB Audio Class 2.0 device, it accepts up to 192 kHz sampling rates. On the other side it outputs “raw” PCM signal, in Philips simultaneous data, or I2S form, that can be sent to D/A chips. Also, it provides galvanic decoupling in between.

Here is the block diagram. Please click on the picture to see it full resolution.



Simultaneous data protocol can be used exclusively with TDA1541(A) (and with TDA1540 if it is 14 bit), and I described its advantages in this post. This protocol has separate left and right channels data lines, and since this board will also output inverted data signals, and will have two BCK and two LE (Latch Enable) connectors, it can be also a source for real balanced TDA1541A DAC.

Alternatively, the output can be more common I2S protocol, with additional 75 Ohm BNC S/PDIF.

Practically speaking, the major difference between two options makes the firmware. Either board can be also reprogrammed later, and there will be several ways to do this, and the most radical option would be to replace flash memory chip with another one. We can also ship additional flash chip with alternative firmware with any order. This chip is SOIC-8, supposedly not hard for soldering for any serious DIY-er.

The board needs one 8-10 VAC transformer i.e. secondary winding, to feed the audio clocks and reclocker. I have given up of providing the option of connecting the whole board supply to USB +5 VDC Vbus, as it would compromise the design, and is probably not what many owners will actually want to use anyhow.

The output can normally drive reasonably long cables, up to 50 cm or so, and several gates, but it is not meant to drive terminated lines. Model S MkIII users who possibly want to use this board with it are welcome to contact me on this matter. S/PDIF output stage is of course different, and it includes discrete buffer.

Clocks are AVX SMD cans, and frequencies are 22.5792 MHz and 24.576 MHz. For simultaneous data mode it is possible to use lower frequencies, and on demand we can install 11.2896 / 12.288 MHz pair instead, so please specify if you want these when placing an order. Please however note that the 22.5792 / 24.576 MHz are required for full hardware compatibility with I2S & S/PDIF option, namely with 11.2896 / 12.288 MHz an S/PDIF output will not operate at 176.4 kHz and 192 kHz sampling rates. Selected clock frequency is also available as master clock (MCK) output, which TDA1541A users normally will not need, unless they want to experiment by clocking it directly – the old non -A version via its “system clock” input pin (4), or later -A version via adequate frequency divider.

The other parts you will find on this (usual Audial white) 95×105 mm board are XMOS processor (128 pin), common mode passive supply filtering, Schottky rectifier, low noise discrete regulator, VHC logic, NP0/C0G capacitors for critical local decoupling, etc.



Driver for Windows 7/8/10 will be provided, whereas Mac OS X and Linux are USB Audio Class 2.0 compliant, so they do not need a driver.

The board is available as ready made module only, and price per unit is € 150. Additional firmware flash chip is € 10. Shipping and handling fee is € 17 (Airmail, registered and insured), and it is charged once per order, regardless of Qty. We accept PayPal and bank wire transfers.


Edit 30/10/2016: PCM1704 format is added as an option in the ordering form.

Edit 06/11/2016: Added instruction manual preview.

Edit 11/12/2016: The whole project took a few weeks more than originally planned, but the next week we will start producing a series of these boards.

Edit 05/01/2017: Added instruction manual revision 0. The boards are now shipping. Ordering form is now closed, but in the second half of January we will make somewhat more samples.



  Instruction manual rev. A




Related topics:

USB to simultaneous data or I2S converter, series 2

USB board Mk3 and AYA II revisited (proposal)



105 Replies

  1. 28-01-2017

    Hello Thorsten,

    Another run is just being finished. Most of these samples are actually sold “offline” in the meantime, but some, mostly with 11.2896 / 12.288 MHz master clocks, are still available.

    There will be probably no another online ordering form, so it is the best if you also place your order offline. You can use either the website contact form or my e-mail address.


  2. 29-01-2017

    Hi Pedja
    Is it possible to use a socket for the firmware chips for those of us who might well swap DACs around regularly? There are SOIC-8 sockets on eBay very cheap. They have DIP8 legs but I guess they can be bent to conform with the SOIC-8 pads on your USB board. Can you recommend anything better?

    Do you think these will work?

    I also wonder if there will be any sound quality problem using a socket? I imagine the flash just loads on power-up into the XMOS so I cant see it being an issue?

  3. 29-01-2017

    Hi Pedja,

    thank you for all the work it sounds wonderful!

    just finished my prototype setup in combination with aya2 board, the simultan mode shows the best sides from 1541 🙂

  4. 29-01-2017

    Hello Hugh,

    Yes, that is correct, flash chip loads the firmware to XMOS chip at power up, i.e. at the moment the USB part of the board gets the +5 V from USB line.

    There is however a problem with Ebay link, as it apparently does not open any particular item.


  5. 29-01-2017

    You are welcome, Klaus.

    I agree, it is the simultaneous mode that gives all the transparency from TDA1541A.

    Also, this board is in my view one very nice and successful project, and I don’t think the source for TDA1541A, especially if it is driven by 11.2896 / 12.288 MHz clocks, can get much better than this.


  6. 01-02-2017

    Hi Pedja,

    It looks very good. In looking at it further you have also provided the inversions of left and right channels which I imagine would provide differentials from the same DAC for each independent channel.

    By the way I converted my Marantz CD-94 to NOS mode and redid all the analog circuits. It sounds very good even against a DCS machine. Hence am looking forward to getting one of your boards to connect to a USB port.

  7. 02-02-2017

    Thanks, Gerrit.

    Ordering details are just sent to you by e-mail.

  8. 09-02-2017

    At this moment two completed boards are available for shipping, both with 11.2896 / 12.288 MHz master clocks.

  9. 11-02-2017

    Can you hold one for me? If you still have one I will send you the money for it.

  10. 11-02-2017

    Yes, one sample will be put aside for you.

  11. 15-02-2017

    Hi Pedja,

    If the board is still available,
    I want to buy one.

  12. 15-02-2017

    Hello Katsumi,

    The answer sent by e-mail.


  13. 15-02-2017

    Also, a notice regarding possible additional inquiries.

    There are already more requests than available boards, so for now I can not accept additional orders.

    A few pieces remained incomplete from previous run though, and I will let you know if and when some of these are completed and available.

  14. 16-02-2017

    Hi Pedja,
    I went through PayPal and payed for mine a few days ago. I hope everything went OK. Can you provide a tracking number?


  15. 16-02-2017

    Hello Gerrit,

    I have just checked, and something apparently did not work well, as your payment still did not appear at this side. Please send me your payment ID (you can send it to my e-mail), and I will check this with PayPal support.


  16. 16-02-2017

    I was having trouble sending you a payment and didn’t verify with PayPal. In checking there was no payment made to you. Can you provide the name I should making out the payment to again please.


  17. 16-02-2017

    Audial account address at PayPal is (Audial d.o.o.).


  18. 16-02-2017

    Hi Pedja,

    Thanks I have connected USB simultaneous to aya11 and sounds great, I want to ask you about Dem BCK because I need to use 2,I noticed that there is 2 BCK on the PCB so could I connect one to TDA and one to Dem using both BCK J206 and J207?

    Hope you can understand my question.


  19. 17-02-2017

    Hello Peter,

    Supposing it is the AYA II 2014 that you use, so long as you use one AYA board, one BCK connection will be fine. And there, at AYA 2014 board, you can hard-wire the BCK signal from relay output to send it to the DEM clocking circuit input (Q208).

    If you want to use USB board with two AYA boards, then for that second board you should use the second BCK output. And then at second AYA board you make the same connection for DEM as on the first AYA board.


  20. 17-02-2017

    Hi Pedja,
    I cleared things up with PayPal as some form of security check that caused them to deny my payment without informing me. Anyway, the payment now shows up in my email so I hope everything is now fine.


  21. 17-02-2017

    I am confirming it is fine this time.

    The board will start its way tomorrow, and I will send you a tracking information.


  22. 17-02-2017

    Hi Pedja!
    The boards have arrived in good codition!

  23. 17-02-2017

    Thanks, Thorsten. I hope you’ll have them working soon.

    @ Gerrit: Your board was shipped today, and tracking information has been just sent to you by e-mail.


  24. 03-03-2017

    Hi Predja
    Why does the manual advise not installing the board on top the AYA board?

  25. 03-03-2017

    Hello Brent,

    To make the emitting between the circuits as low as possible, it is recommended to put them all into the same (ground) plane. Or, in different words, by mounting USB board above the DAC board, you help it emitting down to the DAC board. Things would be somewhat better if ground plane on this USB board was consistent, but that is not the case.

    You can think about this just like about putting i.e. elevating antenna above the ground (plane): the higher it is, the better it emits (or receives), and vice versa.


  26. 18-03-2017

    Dear Pedja,

    If the board is still available, I want to buy one please.

    Thank you,

  27. 18-03-2017

    Hello Antonio,

    I am sorry but all the boards are now sold out.

    However I am still regularly receiving requests, and it looks like this project still draws the attention. So we might rerun it in some similar form once again, later this year.


  28. 21-03-2017

    Hi Pedja,

    if you sell the boards again, please let me know.

    Best regards

  29. 22-03-2017

    Hello Tobias,

    I will do.


  30. 09-04-2017

    Hi Pedja,

    I have a problem no sound, the computer is showing audial usb audio in volume control but no sound, I checked connections and OK USB cable ok power on ok 2 green on Pcb ok everything else is working, i uninstalled the drivers and reinstalled them with same results i am also getting playback error on foobar. tried Ian’s PCM to I2s board and aya 2 is working ok so something is not right with simultaneous PCB. its been working fine up till now, its powered by the transformer I bought from you for aya i checked input voltage and it is withing specs.


  31. 10-04-2017

    Hello Peter,

    I have sent you an e-mail, so we can investigate this more effectively.


  32. 02-05-2017

    And now it is official, the second run of this project will be released soon.

    It will have practically the same architecture, but there will be also some changes.

    The first, the USB side of the board will now also have its own local supply (instead of using Vbus). The board will thus require transformer with two secondary windings, one for USB decoding part, and the other for clocks and reclocker. This way it will be fully “self powered”.

    And second, in addition to U.FL, a PCM / I2S output will also have a PCB provision for BNC connectors, so the board might be more convenient for use in its own chassis.

    Which brings us to another news: other than as a module, this version will be offered as ready made unit, in its own chassis, and with BNC outputs (unless U.FL is specifically requested). It will have regular Audial logo, but please note that it still will not be sold on regular basis, but will be also only one time offer.

    More details and schedule will be announced next week.

  33. 03-05-2017


    I don’t know if it’s possible but could it have also an uf-l I2S inputt (for instance to use with a Pi III or Odroid C1+ boards) ?

    (you should make also a mailing list, there are problems with automatic notifications, I just falled on this good news by luck as I check sometimes for diy modifications of the users for my own AYA board)

  34. 03-05-2017

    This board was meant solely as USB front end, so making conversion between different raw PCM formats would be beyond its scope.

    I have checked, and your mail address is now subscribed to this topic. Please let me know if automatic notifications still do not work for you.

    Also, if you are subscribed to the newsletter you will be notified about new board that way too.


  35. 27-05-2017

    Anyone use the board with the Logitech squeezebox touch, especially with the “Enhanced digital output”? I have this setup used with my Amanero and sounded good, but now the Logitech doesn’t known the DAC 😳

  36. 11-06-2017

    How hard would it be to convert the original series 1 board to a series 2? would it be a simple as adding 5v+ to the USB header pin?

  37. 13-07-2017

    Hi Pedja

    If not using on-board DEM circuit will C213 and C214 have to be removed?

  38. 13-07-2017

    Hello Peter,

    Yes, that’s correct. If you use TDA1541A internal oscillator for DEM, which is driven by the capacitor C324, you have to remove C213 and C214.


  39. 15-08-2017

    Hello Pedja,

    I always seem to be auto logged out whenever I click on the Topics of this Board and the upcoming new USB Board.. Hmmm. I had to fill up the fields below and use the captcha code just to post here,

    Anyway, I just want to confirm if the v26 Windows driver is the 64 bit version, and v61 is the 32 bit version?

    Thank you.

  40. 15-08-2017

    Hello Frederick,

    I personally can browse the site, and post to the topics as logged-in user with no issues (and that is how I am posting this reply), and I am not aware of such a problem, but I will look at it. Hope your browser has cookies enabled?

    As for the drivers, both versions can be installed to both 32 and 64 bit Windows.


  41. 16-08-2017

    Hello Pedja,

    Thanks for the reply. I always use Explorer when browsinf the site, I will try Chrome..

    Anyway, got to try v61 yesterday and it worked well with J-river. Wasapi (direct connection) by default.. It sounds very very promising in Simultaneous Data Mode, even for just an hour to say the least =)…

    I tried v26 but there was an error message after finishing installation, so no sound even if there was a ‘handshake’. I’ll try to DL it again, maybe it got corrupted with downloading.

    Here’s a photo of the layout of the innards in a new chassis from Landfall Systems. All that’s left for my friend to install are the LED indicators on the front panel.


    Attached file:

  42. 17-09-2017

    Hi Pedja and others,

    Has anyone here tried Spotify streaming? I tried using it for background music listening, but it plays direcly to the laptop speakers instead of streaming to the USB Board. The driver is recognized by J-River MC21 and has no problem streaming to the USB Board.. Are there any settings I should disable/enable? Thank you.

    PS – I close down J-River before using the Spotify app so that there won’t applications running at the same time.

  43. 20-09-2017

    Hello Frederick,

    I am not using Spotify, but I’d guess that there is some sampling frequency conflict, or maybe even the problem with audio device selection?

    BTW, I hope the problem you had to post as logged-in user is solved. There was indeed an issue with this (related to site content caching), that could occur with some browsers.


  44. 20-09-2017

    Output format change is now possible by firmware update, which can be done from Windows control panel. Detailed instructions and most typical migrations files are already uploaded (requires log-in).

  45. 20-09-2017

    Hi Pedja,

    I have a temporary workaround for using Spotify by downloading WMD Driver (a virtual audio driver) in JRiver. It let’s you play other media apps through JRiver media center (with Audial USB Audio Device KS [WASAPI] selected).. So with Spotify playing I just have to have JRiver running too to be able to stream to the USB Board.

    The USB Driver that works without bugs with my OS (Win 7 Ultimate) is V 1.61. The other two drivers were buggy and always had pop up error messages. With V 1.26 I see that it could stream (D102 lighting up) but there’s no sound output at all. I tried a few sampling frequencies in the Control Panel >> Playback Devices > ‘Audial USB Audio Device’ (Properties) > Advanced, but can’t seem to find the right one for Spotify to stream by itself. This was before installing Jriver’s WDM Driver. I’ll be getting a new PC soon, so I’ll make sure to try all frequencies next time.

    Thanks for the new FW update with selectable output formats. I’ll need read up on it more..

    I have no problems logging in now since my last post.. Thank you.


  46. 01-10-2017

    So it looks like the problem is with routing between Spotify and the driver. You can use also dedicated application for routing, but please let us know if things will work better with another PC.

    And, one information for all having the boards with 22.5792/24.576 MHz master clocks. Now you can add 352.4/384 kHz capability, by firmware update. Please check out the download section.

  47. 02-10-2017

    Hi Pedja,

    Yes, I will try with a different PC soon enough..

    Thanks for the new higher sampling frequency FW update. Does the v 1.61 driver also have a DFU utility so I could revert back to the earlier firmware just in case? Only v1.61 works with my current OS without bugs (no sound output or error messages). Thank you.

  48. 02-10-2017

    Driver 1.61 does not include this utility, but it is now available as separate file, along with instructions.


  49. 03-10-2017

    Hi Pedja,

    Thanks for your efforts on this.. My older PC’s hardware limit is only up to 192Khz. Do most Windows 10 machines can do a 384Khz sampling rate? Thank you.

Leave a Reply