AYA II 2014 DIY edition talks

AYA II 2014 DIY assembly manual is now done and sent to the users, and all the necessary data is there.

Still, some things in this project might not be straightforward. And some parts are harder to find, and not all sources are reliable. Also, many users will want to abandon suggested parts, and replace them with ones of their choice. Furthermore, this project includes some optional features, which use can require more detailed instructions.

This page is here to help AYA II 2014 DIY users avoid pitfalls, so we see as many boards as possible turning into the nice DACs. So, everything related to the AYA II 2014 parts, functioning and possible tweaks can be asked, discussed or shared here.

262 thoughts on “AYA II 2014 DIY edition talks”

  1. Thank you for your help and this nice & kind shared project.

    Have a question please :

    For the TDA shunts could Rubycon caps advised to be swaped with the old Sanyo organic semi-conductor Oscon SP serie (or the normal serie ?) if no rubycons owned ?



  2. You are welcome.

    Yes, OS-Con SP usually works well in digital and mixed signal circuits, and it definitely worked well for me with TDA1541A.

    However, you will still have to use something appropriate in the analog output stage, and OS-Cons don’t qualify here. In the years after Black Gates, I was using Rubycon ZA, and now I think they were one of two best local decoupling caps for the analog audio circuits, on pair with BG N/NX. Maybe not that full bodied as BG, but greatly balanced, and maybe with even better texture. Too bad they were also discontinued, due to their RoHS incompatibility. Rubycon however makes ZLG as a replacement, and even if they cost notably less than ZA, they indeed offer similar performance.

    Still, I will admire any other good solution, but please note that local decoupling caps in the analog output stage are very… touchy. And these Rubycons are… thankful. 😉

  3. Hello Pedja. I have some BG NXhiQ 0.1uf. Would it be wise to use them as an alterantive for the smd pps caps?(leads would be cut short).I know that probably it isnt a good choice but i would like your opinion please. Also, is the tolerance for 1nf and 2.2nf caps important?

  4. Hi does anyone know a honest source for the TDA 1541A, or do have to take my chances on EBay ?

  5. Antonio María Ferragut

    Hi Pedja,
    I guess the use of two 47uf 6.3V is because the no existence of 22uf value to use as a single capacitor. (I´d like to use the pair of BG NX 22uf/6.3V I have in my tray).
    Also, I will use BC550C and BC560C (Is it the higher HFE, The better?)and LM431 instead TL431, because I have them (are exactly interchangeable?). By the way, I know that package is since 29.09.2014 in Spain, but I haven´t received it yet.

  6. @ Nikos:

    BG NX Hi-Q 0.1 uF – sure. That was what I was using in the first AYA, later also 0.47 uF as they were the same size, and also what I was using to modify some CD players, at the time they were available. It takes some time to sort everything at SMD footprints and keep leads short, but it is possible, and everything will work great.

    Both 2n2 and 1nF act as low pass caps, but their cut off frequency is not any critical, and can be notably shifted both up and down. The 2n2 works with OPA861 emitter input impedance (10 Ohm) so its -3 dB point is in MHz range, but practically it is here rather to provide low impedance path at high frequencies to TDA1541(A) output. The 1 nF bypasses I/V resistor so its -3 dB point is somewhere at 100 kHz, but again with notable margin both ways. Apart from acting as low pass, this cap is here also to stabilize I/V stage (stability is however not that critical with OPA861, as with some other similar circuits). Sonic effects of these caps will be way more determined by their quality than by their nominal capacity tolerance.


  7. @ Paul:

    I would rather avoid posting online actual recommendations. However I will try to post some tips on this matter next days.

    @ Antonio:

    Silmic are unipolar caps, and this way they act as bipolar combo, so they can take either positive or negative voltage at their ends. Required capacitance is defined by downstream device input impedance. So with say 100 kOhm input impedance amplifier, something like 0.1-0.2 uF can be enough. Consequently, with 10 kOhm anything more than 2 uF will suffice, and if you want to use it with amplifier having a front end like Audial Model A, you will need all these 22 uF, to achieve linear frequency response right from 20 Hz.

    Of course, it is no problem to use Black Gate, I still like N/NX.

    Yes, basically the higher hFE, the better. And yes, you can normally use LM431.

    And let’s hope you will get the shipment soon.


  8. OPA861 is regularly available at most major electronics parts vendors, such as Farnell, Digi-key and Mouser.

    Suffix -C marks later revision of PCM270x family that brought two changes (Windows logo certification and volume control behavior), but none of them is of interest in this application.
    So you can use either version.


  9. Thanks Pedja for the elaborate answer. Turns out that i have only 12 BGnxHIQ 0,1uf. I’ll go the SMD PPS route then

  10. Hi Pedja, Thank you for your reply above. Have an other question please :

    I plan to use C-core transformers but have no sucess sourcing a 16 V transformer (neither IE or torroids) for the -15V ! Does the shunt reg of the -15V rail work with a 18 V transformer ?

    Maybe someone can advise me a place in Europe sourcing such 16 V secondary transfomers if 18 V is a problem ?

    I found several 47 uf standard BG in my Tools box I will // for feeding the 861 OPA 🙂

  11. Yes, 18 VAC will be fine. This voltage will move the voltage at C3 close to its nominal 25 VDC rating, but it will cope.


  12. Regarding TDA1541A:

    Firstly, there is no place that I could direct you to, with guarantee to get good sample. With some care and a bit of luck you can find one though, for reasonable price.

    Known vendors very rarely sell this part. They face the problem of not only bad series of discontinued parts, with some or many faulty samples, but also the problem of completely fake parts. Donberg was one of these, and some yeas ago I got indeed completely fake batch from them. TDA1541A markings looked relatively fine, but it was who-knows-what inside. However, Donberg is reputable vendor, and was exactly great example of good attitude, and I was fully refunded for my order. A bit later they started selling good samples from my favourite Taiwan 1997/1998 series, however 40 EUR was significant price for this part at that time.

    (These days however I see these series offered even for double this price. It might turn out that my preferences helped someone to earn money this (easy) way, which certainly wasn’t my intention.)

    And here is also what I want to point out. Vendors are normally not in a position to check the stuff they sell, not only every sample, but often they can not check any. But even if you will not be at real big loss if you miss your money for one bad chip, vendor’s readiness to act accordingly is of paramount importance, as it is not only guarantee for your money, but it also pushes them toward more reliable suppliers (and that is how they avoid their loss). This applies to both regular vendors and eBay soloists.

    So, two things can be a problem with TDA1541A:

    1. Faulty sample. Many TDA1541A sold today are apparently coming from originally abandoned series, that didn’t pass quality control. The chances are that all the chips will work, however some with audibly deteriorated performance (noise, distortion), at either one or at both channels. Also, some can suffer from problems with power up: right after the power up they distort, and a few minutes later, when they warm up, they work fine. Which of course also is not acceptable.

    2. Totally fake sample. In other words, DIP 28 package marked as TDA1541A, with unknown content. So far I’ve had no problems in terms of damaging the rest of the unit with such pieces, however if you power up the unit and your chip fails to send anything to the output, I would suggest to observe supply as quick as possible.

    This all might sound scary, but so long as you need one or two pieces only, it shouldn’t be still a rocket science to buy TDA1541A. Request the picture of the chip from your seller, and compare it – a lot of pictures of many different series of TDA1541A are floating around, and you might be even able to see this way if it is genuine and good or not. As usually, common sense helps, and a bit of caution in actual purchase should do the task.

    BTW, for this reason I would also recommend installing socket for TDA1541A. As time goes by I am more and more convinced that the brass is not good material to route the audio signal, however sonic shortcomings are way lower than practical benefits. Besides, socket will make any possible future experiments with different TDA1541A easy.

  13. Hi Pedja,

    thank you for the heads up on obtaining (or trying to…) genuine Philips TDA1541A chips. It’s interesting that you mentioned Donberg since I got one of my chips from there. Not having dealt with the issue of sourcing rare DAC chips before and after searching on the internet for images I wonder whether there is a safe way (relatively speaking) in identifying whether a TDA1541 is fake or not, apart from measuring of course.
    I know the lettering differs on some fakes, but could the grain on the housing be an indication, or any other hints?
    My second chip was a donation from one fine gentleman in Australia, and the outer surface on this one is a much finer texture than the other one…

  14. When I bought a whole bunch of S1’s I thought they were fake too due
    to it’s printing but after further investigation & not to mention I
    send 2 to Lesha for his parallel DAC all was confirm fine. Now off course
    the question becomes will it work after all these years of storage. LOL
    Btw Pedja just to confirm for the output of the transformer, does all the
    supplies use the CT tap as ground reference ?


  15. @ Oliver:

    Donberg had good samples after I warned them about the problem, and it was back in 2007.

    Chips can not be faked in a way the transistors or capacitors are faked, or at least I am not aware if they can. So, for me if chip is operational it is not fake, and if it is fake it will only have package and markings in common with real thing, so it simply won’t work.

    Yes, different series do have somewhat different look (both regarding the surface / tone / grain, and print), but within the series it should be consistent.

    @ Jaffrie:

    No worries, chips, especially these chip are eternal. 😉

    Each secondary winding should be completely separate with respect to the other ones. Only rectified voltages are referenced to the ground, and this is done on PCB.


  16. All the oscillators suggested in the parts list are low jitter, with overall figure of 1 ps or less, in SONET range (12 kHz – 20 MHz). In the audio, we still want low frequency jitter (1/f corner) to be low too, but manufacturers don’t always publish this data, or complete phase noise graphs.

    So, you can possibly still find some better performer. I didn’t try this one, but as far as I can see it won’t fit into AYA II oscillator PCB footprint (5 x 7 mm).

    Also, please note that even though the low frequency jitter might always matter, major contributor to the overall jitter figure will here remain PCM2706 itself. Its jitter however by its most part can be considered benign – please see part 3 (USB decoders) of this article.

  17. Dear Pedja
    Can I replace C302 Rubycon 100uf/16V with the Panasonic FM 100uF/50V ?
    Can I replace C401~C404 Rubycon 100uf/16V with the Panasonic FM 100uF/50V ?
    Can I replace CC305 C308 Rubycon 100uf/16V with the Panasonic FM 100uF/25V ?
    Because that ripple current and impedance of Panasonic FM 100uf/50V(25V)
    are similar to the Rubycon ZA 100uF/16V(6.3V).
    Could you please tell me that sounds like Rubycon ZA?
    I also can buy the Nichicon MUSE KZ 100uF/25V.
    But I do not know which choice is better?
    Thanks a lot.

  18. Dear Pedja
    Can I replace the J113 with the 2SK170?
    The IDSS of my 2SK170 are about 2.5~3mA.
    Thanks a lot.

  19. @ Denny:

    Rubycon ZA and Panasonic FM have different sonic properties. I was using FM for local decoupling everywhere in the AYA III (which was basically S/PDIF only version of AYA II), because ZA were scheduled for obsolescence back in 2008, and I just didn’t know how the things with their availability will turn out. FM are good for their dynamics and three dimensional presentation, and tone is clean, but they can be too bright and unforgiving – many systems in fact favoured AYA II tone and texture.

    This applies to these caps at all the positions you have listed, especially at C401-C404.

    I never experimented much with Muse, and I can not really give you advice on them.

    Yes, you can use 2SK170, but please note its different pinout (gate and source should be swapped).

    @ Nenad:

    This part is also normally available at most major stores. I am not aware of replacement.

  20. Hello pedja,

    I saw some BC547-C in the shematic are different than the BOM for the same emplacement : what is the good one please ? Do you favor a brand (on-semi, etc ?)

    Did you try for curiosity the TDA shunts without their feed-back resistor . Some says instead the measurement, non feed back give good subjective listening results with just around 0.2 uf at the output (eg. 0.1 after the shunt and 0.1 as local chip decoupling )?

    If no ZA/ZLA are sourcable easily and some OSCON SP organic are owned but few : would you advise a swap with an OSCON for better the input cap shunt or the output one ? I talk only about the TDA’s shunts !

    I would like testimonie for a good result elswhere with United CHEMI CON npcap – PSA serie polymer for dac chips. better for me than the brighty Panasonic SEPC/F or too sweet Nichicon (but maybe the Nichicon PLE serie) : e.g. a UCC PSA serie before the shunt and a simple Panasonic SEP (not SEPC/F) after the shunt worth ihmo a try instead simple Panasonic FM if no Black Gates or Rubycon or discontinued Sanyo SP serie are owned for the TDA regs ! It’s a simple idea as I didn’t try it with this DAC especially.

    thanks in advance, kind regards, Eldam

  21. Hi pedja. Im back with another weird question. I have been reading around the net about capacitor values in pins 13 & 18. These are the MSB pins if im not mistaken. Some people have used a high value capacitor there. Would it be insanity to use 22uf Black Gate NX hiQ in Msb pins? I have 4 of these so i am thinking of using 2 as coupling caps and 2 at the MSB pins

  22. Eldam,

    Indeed, parts list shows 546. My apologies for possible confusion, here is the explanation. This BC546-550 transistors series includes three parts specified for different voltages (546, 547, 548), and two somewhat lower noise parts (549, 550). They all in practice have about the same sonic properties, even when used at signal path. And here they are at less critical positions, so you can not really go wrong using any of these. The same applies to their 556-560 PNP counterparts.

    Regulators used for TDA1541A have no feedback, and resistors you are probably referring to (like R303/R304) only provide quiescent current to Darlington transistors.

    These regulators, and I mean regulators as such, so without output capacitors, have linear output impedance of about 0.5 Ohm, up to a few MHz, before its inductive part (rise) takes the place. You can indeed lower the output capacitance to less than 1 uF, and still be sure that its impedance will be nicely integrated to the regulator output impedance, with no peak. (There is no need to use two such caps, because regulators are physically close to the chip, and each regulator feeds only one IC supply pin.) Still, by using something like 10-22 uF you lower the overall impedance in the 100 kHz – 10 MHz range.

    Rubycon caps are regularly available at Farnell and Digi-key.


    That is correct, these pins normally shunt a higher current spikes than other DEM pins do. That is why higher cap value can be beneficial, and in the first TDA1541A DIY DAC that I have published in 2003 I was using here 2u2. Differences however were not that huge. And, as you may notice, I designed my later TDA1541A DACs DEM capacitors setup with tight layout in mind in the first place.

    If you have another pair of BG NX at hands, I would rather guess that you will get more audible differences by using it to make what BG called “super-E caps” for output coupling, to possibly fine tweak the DAC to your liking. BTW, please also note that, even if BG NX have no polarity, they have “directionality”.


  23. Thank you for the hint about BC transistors. I was frighned to do a fatal electrical mistake !

    My apologize for my poor understanding about the shunt regs shematic. I was talking about R304/R308/R312. I believed they was here for less noise but believed also it could sound subjectivly good or better without it because I saw some near shematic elswhere in an analog stage ! Forget it, it was a naive question from me.

    Is my understanding correct about what you writed about BG NX-serie caps above: if 2 just owned the best use should be C409/C411, because the texture they allow at the output (and “directionally” is minus to positiv for the RCAs plugs ? Am I correct ?

    I maid the mistake not to buy a unique transformer because I wanted C-Core. So I will have at leat three primary as I will have 3 Transformers (with sheilding and a wire between primary and secondaries)… Do you advise a ‘Scoobidoo wirering (=wirewound wirering)” with the primary cables of the transformers before the AC plug of the cabinet ? (to minimize the default you talked about ?)

    At least : is no crystal needed if I use an external SPIDF source and don’t use the embeded USB of the shematic (don’t need USB because my streamer is a Squeeze Box Duet: SPIDF) ? (sorry for 2 cents question, my understanding is the crystal is here for USB but I’m not sure !)


  24. Thanks Pedja. I never thought of Super-E configuration at the coupling caps. Im going to definately try this, although two of the four BG NXhiQ caps were used in another project, so their leads were cut to a length that i cannot distinguish which lead connects to outer and inner foil. I’ll just do some experimenting to which one sounds best. If i use smd PPS caps in MSB pins and the rest DEM caps are BG NXhiQ will i be ruining the sound? (ok i admit it, i want to use these BG’s 🙂 i bought them a long time ago and never got to use them)

  25. @ Eldam:

    If you don’t want USB front end, then you don’t need anything shown at page 1 of schematic drawing, in other word you don’t need parts marked as 1xx.

    @ Eldam and Nikos:

    Regarding my findings on BG directionality and (anti)paralleling, please see:

    At N and NX series, shorter lead is at the side of the cap marked by NP or NONPOLAR at its sleeve.

    It really is not my intention to turn you off using BG for DEM. As I said, the first AYA (2006) was done that way, and it was just fine.



  26. Hello Pedja,

    Sorry for the two cents question : could the I/V resistor value be reduced to 1K ? Is the value in this position than the quality of the resistor ?

    Do you advice a reference or type better than an other for this position?

  27. Yes, you can reduce it safely to 1k, with no major change in performance. However please also remember that the I/V resistors value directly determines the output voltage, and with 1k you will get something like 1.4 V RMS (4 V p-p).

    These are the most critical resistors in the DAC, so if you have your favourite resistor type, you should use it here.


  28. Thank you Pedja.

    this answer makes my day… 🙂 This voltage should be enough in my system. Though I will try with a less good 1.5 k I have for a POC…to benchmark both but not with the same voltage… The resistor I want to put is sourcable in 1k or 2k and my poor understanding is the I/V is better with less ohms if possible according the shematic involved.

    Populating finished… soon 🙂

    best regards

  29. A 2k I/V resistor will give a hard time to OPA861 chips with their +/- 5 V supply. The chances are the I/V stage itself will still work somehow (“collector” output), but second OPA861 which is the output buffer (“emitter” output) will be right at its clipping point.

    So, yes, it is better to use 1k, and the output voltage will be sufficient in most systems.


  30. Nenad,

    when you are also audio kutak see my reply to the post in the GB thread on Diyaudio. 😉

  31. Also, on another note…

    I recently performed the correction to the board (Errata 1) using a very sharp Xacto pencil knife instead of a file (I don’t have such a fine file).

    Insert the blade (preferrably a pointed one) from the backside of the PCB and drill the knife between your fingers (rotate the blade). Within seconds the via is disrupted. Double check with your DMM!

  32. Hello,

    I found some Rubycon ZA on Ebay but don’t know if they are genuine. However have some genuine Black Gate standard to compare and check the quality of the Rubycon SA.

    Does anyone know about the break in period of Rubycon ZA ? Same as Black Gate ? 100 hours, more ?

    Thanks in advance


  33. Problem with BG burning in is their weird burning in curve. They sound fine when new, and get better and worse for a couple of times, until they settle to what they are. The process may take a few weeks.

    Rubycon ZA also require some time to give their best, but their burning in time is shorter, and they behave quite nicely in comparison.

  34. Hi Pedja,

    I have been checking the DC at the output of the reg shunt on the output caps without the TDA1541 :

    Have 30 second after switched on : 2x 4.8 for the analog stage

    -5.3 and 5.5 v for the TDA but -20.9 for the -16 v rails

    for this last rail my traffo is a 18 V and is 22 V AC before the bridge and 20.9 DC on the output caps before the TDA !

    All the BC transistor used are 550/560 C series.

    According to you, what do I wrong ?

    Also for the bais with the trim pot about analog stage, do I need a very precise voltage tester or a normal one is enough ?

    Maybe someone to help me here without disturbing Pedja for these few things please ?

    Anybody has a voltage control shematic on several points of the board ?

    Thanks a lot in advance, Regards,

  35. Hi, just to say I checked the zener on the -15v rail ! Pfffff, my electronic parts shop gave me a 18v zener instead 16v ! But don’t understand why I measure 20.9 V ar C302 ? What is the minimum voltage value (traffo checked unloaded) do we have before the rectifier bridge ?

    Also : on the analog regs : my red leds don’t shine ! Where is the cathode on the printed logo of the pcb : the plate side or the bumped side of the logo please ?

    Sorry for my two cents

  36. Dear Pedja
    I want to use TAKMAN carbon film Resistors for all the carbon resistor parts.
    I want to use Audio Note Tantalum resistors for I/V resistor parts.
    I want to use TX2575 resistors for signal resistor parts.
    I want to use Nichicon MUSE KZ capacitor for c302 c305 c308 c401~c404. (Because I can not find the Rubycon ZA or ZLG)
    Are these choice suitable for AYAII?
    Thanks a lot

  37. Hello,

    Put finally a zener 16V on the -16V… I have this time -18 V on C302 !!!! Alwas 2 v above the value of the zener diode !!!

    Is ist my 18 V traffo ? Is is the BF45A?, The BC559 ? (résistors were checked and are ok).

    Anybody ahve an idea ?

    @Denny : I’m not Pedja but you should follow its BOM. But believe it’s ok with the takman at the places you talk about(I surmise you will hear no difference with a simple carbon here)But I would swap the TX with the tantalum each others ! Foil are good in I/V, not speacially this one. As said if you go on Farnell you will find some Rubycon ZLG, some SA can be sourced on EBay but avoid chinese shop, too much risky… If you go on Farnell tou can also buy a Rhopoint 816G 1k ohms if you can live with 1.5V at the output… most of us can live with such a value.


  38. Eldam,

    Something might be wrong with Q302 or Q303. They should drop the voltage, not increase it (I mean in terms of negative voltage). Please check the voltage at the emitter of Q302.

    At PCB, looking from the top, D401/D402 anodes are left, and cathodes are right. The rule is simple, LEDs either do or do not work. So if they do not light, they are connected in the opposite direction.

    Conventional Voltmeter can do everything you need to complete the unit.


  39. Denny,

    Even though I believe that the officially recommended passive parts are ultimately not that hard to source, I wouldn’t like to discourage using different parts or combinations of parts.

    In this case, yes, you can get good results with parts you listed, but it is important to make sure that you are familiar with their sonic nature, and you are using them… well, in a controlled way.

    Yet, using Takman at all positions is really redundant.


  40. Hi,

    I’m afraid that my skill is not good enough: just after

    For the -16 v : I have -17.2 at the output of the diode on the emitter of the bc557… and know maid a short with the soldering pen, so have probably to change the darlington, the 245 jfet, the bc557 and the zener again !

    for the 5V rails I have a too important voltage at the switch on : 5.5 v & -5.6 V !

    Where am I wrong : does the diodes have a too low reverse wattage ?
    the zener 16 v is a Motorolla BZX 85C ! Are the 245a jfet bought on a shop could be fake !

    What is the voltage we have to measure after each of the 3 diodes ? Could I just solder the zener & a BF245A to check the voltage or have I to solder also the BC557 to check the voltage (do the bipolars involved in the voltage drop out?)

    Sorry, think I had no luck during my assembly of the board !

    If any could advise me please, as all the components are already sold… I hesitate to beginn again from scratch !


  41. eldam,

    i dont understand: do you have new “white” PCB and BF245 or
    old green PCB and BF245?

  42. Nemad,

    BF245A = J113 ; this is just I found some BF245A off the shelf, and as I’m nostalgic, I bought it ! but results are strange with my voltages !

    I found also some Allen Bradley at Hifficollective UK and some Rubycon SA on EBay and some Rhopoint for I:V at Farnell like some Rubycon ZLG as well to check if the ZA sourced on EBay are genuine !

    voilà ! 🙂

  43. You can use BF245 instead of J113, but please note their different pinout. You have to mount BF245 the opposite way, so rotated for 180 degrees with respect to the position shown by PCB silkscreen.

  44. Thank you, this why I had problems with voltages… so the new white pcb is not the same surface printing here than the green one !…

    I thought you maid a same pin out replacement of the bf245a with the J113 !

    As we talk also here about alternative parts : J133 & BF245A : same same but different ? (I mean in sound quality)

  45. “(I mean in sound quality)”

    When you start to worry about the sound quality of EVERY part involved you will never be finished… 🙂

  46. Hello Oliver my Buddy fellow of the other bar ;

    Ah, it remind me the true story of Stradivarius : to go fast, he used to take the material in him own wooden floor saying “same same” !

    And this other story of Cromagnon cooking a Mamouth above the fire : “and if we invent electricity?”…”oh no fire it’s enough for cooking the Mamouth”… But the slow cooked Boeuf Bourguignon is quite good also, no ? 🙂

    Agree with you as I’m not Jeanne Calmant, nore I have the skill of Stradivarius!

    Here I just ask for this Jfet as I just learned than sometimes going fast can be too much fast! And I ask to someone who had the patience to improve him own DAC for ten years now for the better!

    I believe everybody may profite of patience and wisdome of a good “artisan” who take the time to improve always like Boule was with furnitures e.g. !

    This is just because I have already the BF245A and maid a “too fast” mistake not checking the datasheet in relation to pins arrengement with the “J133″… and I have two boards :). Notice in the posts above I follow exactly the BOM but for the I/V as no advice but the one we prefer (dixit Pedja Rogic himself)… well here I choosed a Rhopoint 816G sourcable at 1K ohms at Farnell !

    Sorry for the total Off Topic, Maybe here my question doesn’t Worth in relation to thios Jfet in the place it has in the layout, but has I don’t know and have the choice of those two ref, I ask 🙂

    Did you notice the taste of frenchs for long litterature 😉

  47. You sure should make a second (third) career as an author… 🙂

    On the topic: I have not started soldering yet but am still collecting parts.
    Over Christmas I will start populating the board and then go from there.

    My collection of TDA chips has grown to 5 with one more coming (the last three taken from defunkt CD players and all being Non-A :-()

  48. Ouarf, yes as comic author maybe !

    Pfff if we have now to care on each part 🙂

    Don’t forgett to remove the 470 pf cap if you use the “non-A” (dixit A.E Van Vogt)

    Sorry Pedja, Oliver is a friend in an other bar 🙂 but indeed have the interest to show we have to take care when a single part is changed for an other ! 🙁

  49. It is no problem, Eldam. 😉

    “As we talk also here about alternative parts : J133 & BF245A : same same but different ? (I mean in sound quality)”

    In places they are used in the AYA II, you can consider them the same.


  50. Hello Pedja.
    How important is it to use carbon resistors in the power supplies section? Should they be carbon film or carbon composition? is there any current production model/brand that can be recommended here? How about metal film (talking about power supply specifically). Will MF resistors change sound signature of the DAC noticeably?

  51. Dear Pedja
    Is there any current production model/brand that can be recommended in I/V and signal path (R415~R420)?
    Thanks for tell me where can buy the Rubycon ZLG.
    Thanks a lot.

  52. @ Malvin:

    Power supply resistors are conventional parts. They are not that critical, but recently I tend to avoid metal film at general level.

    @ Denny:

    Allen Bradley are long time out of production, however you can buy them still at PartsConnexion and HiFiCollective. They both ship worldwide, and offer different payment methods.

    You won’t find many manufacturers still making carbon resistors, and Takman would be one of them, but I’d still recommend AB. If you want to spend more, you can look for Rhopoint wirewound resistors, and limited selection of these is available at Farnell.

  53. Dear Pedja
    I will buy the AB carbon resistors in R415~R420.
    Should I use the 0.125W 0.25W 0.5W 1W or 2W?
    Thanks a lot.

  54. These resistors will run a few mA at less than 3 V peak, so this is really low power application. I am using mostly 0.25 W, however 0.125 W will be also fine, and you can use 0.5 W too. You won’t be able to normally mount 1 W and 2 W.

    BTW, revision A of assembly manual will follow shortly.


  55. if someone needs PCM2706, i have one more piece,
    wrong purchased at Farnell.

    please, PM me on the “diyaudio com” nick: audio kutak
    (free shipping to EU)

    9 days have passed and the resistors did not come from
    hificollective_uk?? How much should wait until the goods arrive?

  56. @ Kutak : One week for me by air mail…GB to Fr.

    @Denny: as writted somewhere above, ZLG are sourcable at Farnell.

  57. Aya II 2014 finished 🙂 ! Looks great. Thank you Mr Rodgic.

    – Outputs voltage are OK.
    – Burning caps phase now… with genuine BG N 33 uf/16V at output (the only ones I have !).
    – Setup of the trims phase… Not sure I really understood how to match the two trims pot ?? Ah, have to re read the manual and have concentration to avoid shorts 🙂

  58. Oups : sorry for the wrong writted “d” in your name, pardon my dyslexia…, so thank you Mr Pedja Rogic…

  59. No worries about that. 😉

    To tune the offset, please first make sure that TDA1541A receives some clock signals. Most simply, connect CS8414, and in case you don’t have a scope to verify that it works fine, you can use Voltmeter and its DC range, and measure TDA1541A input. You don’t need S/PDIF source actually connected, CS8414 will send out the clock signals without any source connected to it, as explained in the assembly manual at page 11. You want something like 2.5 V at TDA pins 1 and 2 (because clocks duty cycle is roughly 50 %), and 0 V at pin 3 (data – because you need digital silence).

    If it is like that, you can be practically sure that CS8414 works fine, and TDA1541A receives adequate signals to trigger its outputs correctly, to middle point, which you now want to set to 0 VDC.

    To do that, you should measure DC across I/V resistors, and use trimmers to set it as low as possible.

    If you use USB input, you need everything connected, and player best set to pause.

    And please don’t forget to set TDA in I2S mode, by the jumper or switch connecting its pin 27 to pin 28.

    (BTW, some will recommend to connect bit clock to TDA1541A pin 4 too, by tying up TDA1541A pin 2 to pin 4. This is what I was also regularly doing in my previous designs too – just in case – however it wasn’t possible to do that with this PCB, as we need this PCB to possibly use TDA in simultaneous data mode too. Still, the PCB has a provision for jumper between pins 2 and 4, and you will find it marked at the PCB top side by exclamation point, but I actually didn’t find it necessary, at least not with TDA samples I have.)

    And let us know how the things are going on, please. 😉

  60. Yes, let us know how things are going on… 🙂
    Unfortunately here is no space to put pictures, but they sure would be nice from Eldam’s build and all the others that hopefully will come (mine including ;).

  61. Hi Pedja,

    I also have a technical question about the output caps C409-412.
    Somehow I am not very intrigued by the idea of passing the audio signal through electrolytic capacitors…. Maybe I am wrong and I sure have the Silmic’s in my parts drawer, but:
    what choices do I have here? What is the absolute minimum recommended value for the caps to work properly (frequency cutoff etc)?
    I do have BG 470uF 16V left over where I could spare 2, or go film cap with a bunch of ERO KP caps (the blue ones) in // to get, like, 4.7uF…?

    Hvala vam mnogo!

  62. Hello Oliver,

    Capacitor as series element effectively acts like high pass, with cut off frequency (- 3 dB) at 1/(2*Pi*R*C), so required capacitor value depends on the impedance environment. In this case the impedance boils down to the (pre)amplifier input impedance, because the DAC output impedance can be mostly neglected, relatively speaking.

    So, with 10 kOhm (pre)amplifier input, 1 or 2 uF will suffice, and with 100 kOhm yet 100 or 200 nF will do. I am personally lately using amplifiers with low impedance input, less than 1 kOhm, and that is where a couple of tens of uF is necessary.

    Using more capacitance than required per equation above can be beneficial for moving associated phase shift further from audio range, but then again the phase shift is not that audible at the bottom end, either.

    As for using Silmics here, do not be afraid of them. 😉 Many years ago I have given up of using polypropylenes for coupling purposes, in favour of BG N/NX, and I consider Silmic good replacement in this place. Polystyrenes are my favourites among film caps for their natural tone and reasonable cost, but they also got mostly obsolete, and are limited to low values, or in other words to high impedance applications.

    Grüße aus Belgrade!

  63. Peter,

    In solely digital circuits (C105, C205) – yes.

    C303, C306 and C309 might not be necessary at all, and this will be explained in the upcoming assembly manual revision.


  64. Hi Pedja,

    With both 4.9 V reg and 1 K ohms resistor for I/V :

    Have matched the ouput of the I/V resistor: with red voltmeter on I/V output and black voltmeter pin to Gnd pin : -0.76 V for each I/V resistor output after trimming

    Is it ok ? or doeas I miss something ?

    In a manual for beginner something which helps a lot is the voltage on each pins of the pcbs writted on a photograph of the bare pcb.

    But have to say it’s really easy to make shorts with measurement when power is on: made a short again on -16 V (without the tda luckily). So re desoldering operation of the transistors and diodes of tyhe -16 v rail for me tonight ! Pfffff 😉

  65. Offset should be set inside 1-2 mV, and then it may drift somewhat, but again in mV range.

    What did you measure at TDA input pins, does it receive proper signals?

  66. Where does I put the voltemeter pins to measure ? Was I’m right where the description above ? May you please give me an example of what is a measure on each canal after trimming for my understanding ? Does the two canal be matched with each other: so -0.68 V on each is perfectly matched or does the offset is in relation to a zero Gnd ? Do I have something like 0.02 V on each canal ? Sorry for my poor understanding !

    -16 V fixed. I didn’t understand than when the tda is in place we have at the – 16V output reg something like – 3.2 V without music played !

    pins 1 to 3 are 1.8 V from an external board via uf-l in simple I2S mode for test without music sent. But the board outputr a clock signal (w8804 with external crystal). This board is just for test as I plan to go after with simultaned mode (in uf-l yet.)

  67. Yes, that is where you measure and trim the offset, black lead to Gnd, and red lead to the other side of R415/R416.

    I also noticed that you have wrong, yet perfectly matched offset at both channels. This points possible problem at D/A part. Especially because its supply voltage drops when you put the chip in, and it shouldn’t happen under any circumstances.

    Do you have any other TDA chip to try?

  68. Hummm, have a buzzzzzzzzz which is comming from the speaker with or without the external board for input signal dwitched on ! On the Left & Right speaker with both rcas wire or just one tryed on both canal….

    Do I have a problem with the opa861 ? are they sensible to body’s EMC ?

  69. Yes I have other tdas to try ! This one came from a Philips tda player which was shorted on the main common cap (6800 uf) and the ROM collapsed… so maybe the digital sound circuit as well !

    I be change the tda chip and come back !

  70. I experienced several OPA861 faults, so they are definitely not indestructible. However, let’s see what happens when you change TDA.

  71. Humm : speakers are buzzing without or with tda, but before all i have a buzz when the dac board is powered off as well!!!! (and of course the buzz is greater when i switch on the dac : yet with or without tda ?

    Could it be the vias of the revision note 2 ? I fixed it but maybe is there some coupling yet with few capacitance between the two sides of the vias ? A two cents idea of me surely ????

    Jfets of the trims ? they are both b245a in 180° as you explain to me.

  72. Buzz is fixed : my fault because a bad Gnd contact between the chinch body plug and its circle cable plug around it (need of course a good contact with the screws of the chinch…).

    The chinese spidf test board i have output: 2.5 v on pin 1 then 5V to each pin 2 & 3 ! Hurk not sure the tda support such voltage at i2S input ! I can’t find the info. So I stop the experiment before having a clean i2S input !

    Well the second TDA give the same – 0.7 V at I/V ! (have 3 left TDA to check ! )… but of course will not check them with the chinese spidf board (Sure electronics spidf to i2s test board).

    To be continued !

  73. TDA1541A was made in time when 5 V logic was common, however 5 V at pin 2 means that you don’t have any (bit) clock signal there. It is probably just pulled up, and that is why TDA can not trigger the output properly, and its output offset is hence not controlled, and can vary in different ways, from sample to sample.

    (On the other hand, a 5 V at pin 3, i.e. at data line (all “ones”) could serve for trimming, with 2’s complement it is actually only one bit (about 30 micro V) close to all zeros.)

    Still, voltage sagging at TDA pin 15 was not normal. Does it happen with second chip too?

  74. I just saw I was wrong on the way I took measures !

    I took it on output of the i/v resistor and an other vias gnd elswhere !
    Tomorrow is an another day : will take the measure on the two side of the I/V resistor…

    But think the LRCK (WS)of my chinese i2s to spidf test board is certainly not compatible with pins 27/28 shorted.

  75. Hi Pedja,

    With all my dac chips and also without: I have between the two pins of the 1k ohms (instead 1.5 K) I/V resistor :0.83 at minimum trim or 4.2 at max trim. (the 1 nf polystyren place is not populated. With an other tda :0.9V to 4.4 V. Each time there is no drift.

    With the TDA1541 in place, voltage without spidf injection :

    pin 1: 0.5 V
    pin2: 0.7
    pin3: 0.7
    pin 27/28: 4.3

    With spidf injection

    pin 1: 1.4 V
    pin2: 1.4 V
    pin3: – 2.3 V
    pin 27/28: 4.3 V

    But diferent results with an other dac chips !

    spidf on :

    pin 1 to pin 3: 4.3 V (no negative voltage on pin 3 this time !)

    I have two questions please :

    what about this i/v measurement ? opa861 are broken or its jfet ??? Any way to check the opa861 with a voltmeter ? (regs are ok at 4.9 V)

    Not sure my Sure Electronics spidf test board is bad ! So is there a good place to source a genuine TDA1541 ? Donberg Ireland as per Oliver input ? But I would like to fix the problem of the I/V too big voltage before to check again the dac chips (5 are from shortcuted cd players, but for the ROM, not sure the dac chips are all broken.

    What did I make bad on the Aya analog stage about the i/v according to you ?

  76. Offset trimming operates as expected, so output stage is probably fine. Actual voltage is wrong rather because TDA1541A doesn’t receive proper clocks, and zero at data line. There is simply no way for TDA to pull proper (“zero”) output current without such input signal.

    I still can not tell if there is or not the problem with TDA part (4.3 V looks low), however I would firstly make sure that the source works and feeds TDA properly.

  77. Thank you for the hints,

    4.33 V at pin 27/28 is with all the TDAs with or without signal injection on pins 1 to 3.
    Regs are 4.7 & -4.8 V & – 14.5 V at the beginning… which seems correct values (BC550/560C instead BOM)just after a switch on.

    This external spidf was just to test in standard I2S as I’m waiting for a DIY simultaneous board via uf-l with an USB input “Amareno like” board before it(the goal was to try the simultaneous mode as your flight ship DAC Audial S allows but I can’t affoard)

    Have to try the Cirrus spidf chip instead. Is the PCM USB chip subjectivly better or does the sound quality spidf with this Cirrus USB chip result of the source streamer quality (I have a simple SqueeeBox Duet for SPIDF)?

  78. Voltages at TDA1541A supplies are usually 4.8 – 4.9 V for nominal 5 V, and -15 V can be between -14.5 and -16 V. Voltage however shouldn’t sag that much when you put the chip in, so that is why things here have to be checked too. I will send you an email on this matter so we could be maybe more effective.

  79. Dear Pedja
    Could you please tell me that can I replace 6*9VAC(0.7A) with one transformer 2*9VAC (50VA)?

    Could you please give me more information about the “J202″and
    S3″(How should I finish that in the pcb)?

    I really want to know how to build the switch (USB/SPDIF) like this picture.





    Could you please tell me which transistor (designation in manual) need to be matched?

    Thanks a lot.

  80. Dear Pedja
    Could you please tell me that how to adjust the VR401 and VR402?
    Thanks a lot.

  81. Dear Denny,
    why would you want to change the multiple secondaries transformer for s single secondary one…? Do you want to feed all the AC points in parallel then? IMO not a good idea and surely not the intention of Pedja.

    As for your other questions have a close look at page 16 of the manual… you can see the pinout of header J202. The switch S3 switches the TDA1541A between I2S and Simultanious mode, again see the manual (page18).

    Dialing in the DC offset–> see manual.

  82. Hi Everyone,

    I’m starting to populate my board and wondering how others dealt with R213, one side is thru hole, the other a pad. I’m guessing I could just use a thru hole resistor and tack one side to the pad, yes ?



  83. Denny,

    The input switch at original AYA II was a classic small toggle switch. It can be either on/off or on/on (and you use only one input), single pole is enough, but you use a double pole too. Oliver already answered your other questions.


    Yes, and that is in fact how it is supposed to be done.


  84. Thank you Oliver and Pedja.
    Dear Oliver
    Could you please tell me that
    Dialing in the DC offset–> see manual (in page??)

    Dear Pedja
    Could you please tell me which transistor (designation in manual) need to be matched?
    The attach file is my AMP schematic.(LM675-Gainclone)(Cs=1000uF/50V Panasonic FR)
    Please give me any suggestions about this amp? (especially Ci)
    Should I need to change the capacitance value?(C409~C412 47uF/6.3v )

    Thanks a lot.

  85. Dear Denny,

    see page 11 in the chapter “Making it done”. It is quite short and I have not done it myself, yet. See also Eldam’s questions and Pedja’s replies a few posts up 🙂

  86. Denny,

    You can use transistors “off the shelf”, there is no need to match any.

    Regarding your amplifier, it looks like its potentiometer setup is somewhat non-standard, but if drawing is correct, then its lowest possible input impedance is Rin (6.8 kOhm) in series with paralleled connection of VL+R0 (34.87 kOhm) and Rg (22 kOhm), so roughly 20 kOhm in sum. At this impedance, anything above 1 uF will be enough for DAC output coupling (C409 – C412).

    Also, since your amplifier has input capacitor after the potentiometer, you can also think about omitting caps at DAC output, however that would imply extreme cautiousness with output wiring, because errors there might fry OPA861.


  87. Hi Pedja,
    I having no luck sourcing a transformer here in the US with 7 secondary windings, and I would like to follow your advice and not just parallel the power from one transformer. Do you have a transformer that I could buy from you ?



  88. Hi Everybody,

    I’m nearing the completion of stuffing my board and have a few questions. 1.) I’m just planning to use I2S only and want to jump S3, am I right guessing a jumper from the center hole to the end hole next to the last pin on the TDA chip ? 2.) In the BOM It calls for LEDs D204, D401, D402 (red) and D101,D203(green), I found D401 and D402, but for the life of me I can’t find the rest. Have they been deleted from the board ?



  89. Here is a shot of the bottom, I put some of the film caps on the bottom because they fit easier.

  90. Paul,
    the other LED’s are not on the board and to be placed in your future case… see the schematics page 15-16 or so 🙂

    Good work with the soldering… I have finished mine as well but have not yet tested the power supplies for lack of time.

  91. Dear Pedja
    Can I replace 9VAC transformers with 12VAV transformers?
    Should I need to change others parts?(like 1500uF/16V,120uF/16V –> /25V?)

    Thanks a lot.

  92. Paul,

    It is good to see it progressing.

    Please however take note on PSU resistors, these look like 1/8 W parts. It is relatively low voltage drop there, and these resistors power rating should not be high because of dissipation, however they have to convey current spikes exceeding 100 mA, which is why I recommend more robust parts here.

    Yes, you are correct regarding the jumper.


  93. Denny,

    Yes, you can use 12 VAC. In this case capacitors C1, C2 and C4-C7 will be at their limit if rated at 16 VDC, so it would be good to use 25 V parts there.


  94. Dear Pedja,
    while my AYA board is sitting on the workbench 99% finished I do have a question: I am planning to use the board with a WaveIO USB-I2S converter (that I am already using in another DAC).
    Do I have to use pads J305-307 (already populated) or pads J301-304 to hook up the WaveIO? And since the designations of pads are different… what goes where?
    Data to I2S Data, BCK to BCLK, WS to LRCK (?) and DataR to MCLK then (the latter ones always being the connections on the WaveIO)??
    Your help is very appreciated here!


  95. Hello Oliver,

    So long as you intend to use only I2S, you can use either of two external inputs. The only advantage of J301-J304 is that it can be used also as simultaneous mode data input.

    Yes, LRCK (left/right clock) should be connected to WS (word select) input. The terms LRCK and WS are often used interchangeably, and denote the clock that “tells” when binary word starts and when it stops, i.e. when data belongs to the left and when it belongs to the right channel. This is because in I2S mode data is time multiplexed, and carried by one line. This clock frequency equals sampling (or oversampling) frequency.

    In simultaneous data mode, data for each channel is carried separately (and simultaneously), and in this case pin 3 receives only left channel data, while pin 4 is used for right channel data. That is also why in this mode we don’t have “word select” or “left/right clock”, but the same line is used as a “latch”.

    TDA1541A does not require master clock (MCLK) or system clock, however early, non -A TDA1541 pin 4 was used as a separate input for clock that triggers its output in I2S mode. In fact, when you look at most implementations of both -A and non -A TDA1541, including those done by Philips (or me for that matter), you will notice pin 4 shorted to pin 2 (BCK). Normally however TDA1541-A in I2S mode does not require any connection at its pin 4, because its output is triggered by BCK.

    (BTW, you will notice that this PCB actually has two pads that make possible to jumper pin 4 to pin 2, and it is marked by exclamation point.)

    Either way, TDA1541A can not use master or system clock anyhow, so you can simply disregard this output from WaveIO.


  96. Oliver,

    A photograph 🙂

    Do you plan to use the isolated (non-ufl) or the non isolated (uf-l) of the Wave I/O ?

  97. Damien,

    no WaveIO yet…
    I have not decided yet what to use.. maybe I am trying both ways when I have it working 😉

  98. Another question just to be sure…

    Are the two spots marked with an arrow in my pic supposed to be connected when soldered?
    On my board they sure are because they are VERY close together…

  99. Yes, these pins are already connected to these pads.

    Regarding your previous picture, I see that you are using S/PDIF front stage, and u.FL connectors at the same side. In this case it would be indeed more practical to use J301 – J303, so the DAC normally has two inputs, switched by S2 (S/PDIF / u.FL).

    Great parts, BTW.

  100. “In this case it would be indeed more practical to use J301 – J303”

    Thanks for your reply. I am pretty much set on using I2S only (and probably some Fifo boards later to make this DAC capable of playing high res content in simultanious mode)… so in this case I can still use J305-307?
    Also… for some reason I only ordered one relay from Mouser and populated the S/PDIF side. When I don’t use the S/PDIF input (nothing connected), can I jumper the second relay as per your description in the manual?

    Yesterday I powered the DAC up the first time (after I installed really nice WAGO terminals in light grey for the AC connections 🙂 )… and I seem to have all the voltages needed. The voltages around the TDA are just a tad low but I really didn’t let it warm up much (-14,92V etc).

    So now I have to wait for the WaveIO (sans daughter board)!


  101. You are welcome, Paul.


    Yes, you can leave one relay as is, when unpowered it will work as a jumper for S/PDIF stage, and use wire jumpers across the other one, as explained.

    And yes, for I2S only, a J305 – J307 will do, however any time you connect an external I2S source you will have to disconnect the output of on-board S/PDIF receiver.

    For that reason, even though it might slow you down somewhat at this time, mounting J301 – J304 (and the other relay) could give you more convenience later. It will be also good because you are planning to try simultaneous mode at one point.

    A -14.92 V at pin 15 can be considered precise. 😉

    Viele Grüße

  102. Pedja,

    whenever I make it to Belgrade again I will have a Jelen/Law/Apatinska pivo with you 😉

    Will get some more U/FL connectors…

  103. Dear Pedja
    I want to use the external clock.
    I just need to remove C324 and put the clock signal into the “DEMCKT in”?
    Thanks a lot

  104. That is correct, Denny. And you will also need to connect appropriate Q208 output to R218.

    Still, if you have any doubt about this section, you can start with typical C324, and make external DEM clock circuit ready for use by populating all the parts shown at schematic sheet 3 but C213/C214. Once you make sure all works well, you can remove C324 and connect external DEM clock by soldering C213/C214.


  105. Hello Pedja,

    Does anybody try the Ian’s I2StoPCM for inject simultanous mode ?

    Have yet some problem with 0.8 V (instead the normal around 0.0025 V)at the pins of the i/v resistors desptite the Wave I/O (connected and playing via its driver on a W8)through the Ian’s board into the AYA II 2014.

    So I putted a strap for simultanous mode on the AYA II 2014 board.
    The regs are ok with 4.8 and -4.7 & – 14.7 for the TDA-1541A & 4.8 v for the the oaps.

    What sould be the normals values at Pins 1 to 5 of the TDA-1541 while having into it a I2S signal (streaming music) please ?

    I would try also the Wave I/O I2S mode (without the simultanous board)? Oliver, may you inform me please about your pins voltage measurements when your Wave I/O will be connected ?

    Regards (I hope to be near to let it sings :). Sorry for the non conventional question as far from the stock conf for what the Aya 2 where re-print. Hope someone follow the same path as me to help to fix it.)

  106. Errata : some good news ! We sould have always re read the manual ! As I am in simultanous mode the rules to check the voltage across the I/V résistors is not the same as I2S mode !

    So I have something more conform now : a drift showed aound00.20 mV on the screen. But many drift between + 00.1 mv and – 00.4 mv

    Sorry to re ask about how to process : when the TDA is running for 10 minutes (so enough hot):

    Should have I music streaming (or bitclock is enough?) to fix the drift with the pot ?

    Should have I a damned zero on the screen ? or a drift between zero and -0.20 mV only !

    Sorry for my noob-ery 😉 I don’t post picture with wave I/O and I2StoPCM with the AYA as off topic despite the gentle ufl connectors possibility offered by Pedja.

  107. And and and…. It sounds Hurra !

    Very huge and great sound and it is just from the non isolated side of the Wave I/O (no isolation nore simultaneous ! )

    THANK YOU VERY MUCH Pedja…. I can tell you you’got the good ears 🙂

    PS : I have the DC blockers Black Gates N series at the output : 33 uF /16 V… Greaaaaat !

  108. Well I finally got the hardest part done, the case. Because I’m a poor married guy I can’t afford to buy a ready made case, so I need to DIY. Here is the front

  109. Well here is the front this time. I have almost everything I need to finish now and I’m just waiting on a few more parts

  110. I do have one question about J201, do I have the numbering correct in the picture? I made up a small daughter board for the indicator LEDs and want to make sure that I plug them into the right pins.


  111. Thanks for the pictures, Paul. Keep us updated, please.

    Numbering goes in pairs, so 6 = > 2, 2 => 3, and 5 => 9.


  112. I have a problem with the negative power supply for the opa861 op amps, there is a short circuit across q402 and C403 resulting in no negative power supply for the op amps. All the components are in the correct places and are the correct components.all other power supplies check out ok.

  113. Thanks for posting, Peter.

    Please note that you have to cut two through hole pads connections, according to the errata information sent by e-mails on 23 and 30 October (the second one applies to the problem with output stage negative supply). Please let me know if you missed these e-mails, and I will resend them to you.


  114. Hi Pedja,
    Sorry can you please email them again, looks like my old emails have been archived.

    Thank you


  115. And now one nice thing, especially because many AYA II 2014 builders asked for source for reliable and genuine TD1541A.

    Vallo from Tallinn, Estonia, has some quantity of these chips and he was kind enough to send me the batch. I have checked it, and this is definitely genuine and reliable series. Vallo has no time to sell and send the chips one by one, but he would kindly donate some of them, and all he wants is to make them go to those who really need them.

    These days however I do not have much time either, but I am ready to send some of these chips to the volunteer(s) who will forward them further. Thus, at this time 15 TDA1541A pieces are offered to AYA II 2014 builders at no price. Instead, sending back the picture of completed and/or working AYA II 2014 DAC will be more than enough.

    As for performance, these chips are marked R1, so you might think that their linearity error is inevitably 2 LSB. However datasheet claim of linearity error of "< 2 LSB" ("less than") should be taken literally, and my measurements of these chips show that their measured performance is on pair with 1993 made samples I used in the late (after-run) AYA II units, and in completed AYA II 2014 boards, which had no suffix, and are hence "< 1 LSB". Still, it doesn't mean that these series are also sonically entirely the same, in fact my earlier finding that each series of TDA1541(A) has somewhat distinctive sonic signature (and each still being the good old 1541(A)) confirmed this time too.

  116. Hi Pedja,

    I could handle the US distribution of chips to builders if no one else is doing it.


  117. Dear Pedja
    Is there any current production model/brand that can be recommended in 11.2896MHz crystal?
    Thanks a lot

  118. I’d like to have at least 1 of these precious chips (bought 2 AYA ii boards previously.. ). I’m in the US. Thanks

  119. Sylvain Bergeron

    Hi. I can distribute chip in Canada, using affordable Canada Post shipping. I would like one chip for me if possible. Thanks

  120. Anthony Graham

    Hi Pedja,

    I could handle distribution in Australia. One chip for me personally.



  121. Richard (rab)

    Another two for Australia please, for the two boards I bought; Anthony could send them to me.


  122. If I was counting correctly, at this moment ten chips here, plus two offline, so three pieces are still available.

    I hope we could do this by one shipping for USA and Canada, another one for EU (now there is Audial partner in Zagreb, so I could probably use them to forward Marijan’s sample), and if we could by any chance do one shipping for Australia and Taiwan we would be done.

  123. Denny,

    You won’t need such a crystal for basic AYA II operation. Anyway, big vendors still sell them. You should look for low ESR / high Q part.


  124. So, all the samples are reserved now.

    Let me check all the locations, please, so we can make a plan for shipping.

  125. Hi Pedja,

    If there are any still left I would like one chip, bought 2x PCB, I’m in Australia also.


  126. Hi Pedja, just wondering if i could still buy AYA II toroid transformer from you? i realise now that i should have when i ordered the PCB’s.


  127. I will ship a few more chips than originally planned, but we will stop here. So please, no more requests.

    @ Peter:

    There are no transformers on stock anymore. But maybe I can do something in this regard, especially if you are agreed to forward the other chips for Australian users to Anthony? 😉

  128. Anthony Graham

    No problems my end either. Peterma, perhaps you may even like to distribute them in Australia? Either way is fine with me. I’m just glad to have access to the chip.



  129. I think we solved shipping to USA and Canada.

    Also, we solved shipping to Australia, and if anyone from Australia is willing to forward one sample to Taiwan, we are done with that part of the world too.

    And we still have to solve shipping to EU. It would be great if EU users can organize forwarding, so I can send all samples (four, excluding Marijan’s one) to one address.

  130. Pedja, when you have someone wanting achip (on top of what you planned).. I will donate mine to this person 🙂

  131. Pedja
    I try to finish this DAC project.
    It can detect signal when use S/PDIF in.
    It can not detect signal when use USB in.(try to switch “S2”)
    I try to check every part on PCB.
    Could you please give me some suggestions.
    Thanks a lot.

  132. It can not detect the “USB audio device” on my computer.(WIN7 64bits)
    I also try to change other computer.
    Thanks a lot.

  133. I try to use S/PDIF and USB at the same time.
    I try to turn off/on the S2.
    I find that the D203(SPDIF LED) is always “ON”.
    I find that the D101(USB LED) is always “OFF”.

    I try to ONLY use USB.
    I try to turn off/on the S2.
    I find that the D203(SPDIF LED) is always “OFF”.
    I find that the D101(USB LED) is always “OFF”.
    Thanks a lot

  134. The problem is apparently with USB input stage, however I can not spot any problem from these pictures for now. (Actually, you shouldn’t use both C324 and C213/C214 at the same time, but that would be the other problem.)

    I have sent you an e-mail, we might more effective that way.

  135. Hi,

    I believe, despite the Pedja’s advice not to go without DC caps, I fryed the opa861 at switch on or off ! (after putting the output stage with no dc caps and removin the 100kR !

    Mostly because I wanted to keep my BG N caps for the -16 V rail and still don’t like the swap with all the others DC caps !

    Any way to check with a voltmeter if I have the first two opa861 still alive ?

    Does in the future a little Lundhal traffo (have one on hand in a cd player diyed !)can avoid the “plop” problem in a futur or still the windings let travel the “plop” which break the oap861 at switch on/off in relation to the load (the next device : pre or amp) ?

    They told me : RTFM !

  136. Well, it happens. Then you fix it and then forget. 😉

    If you can run sine wave, just put Volt meter to AC range, and check what you have at Q409 and Q410 pins 8 (also points where you trim DC offset). It will show you if Q409 and Q410 work or not.

    If you use real output transformer, with low DC resistance windings, then it will present low DC resistance path to ground, so without caps OPA861 may be at risk.

  137. Yeap it happens more easily I believed.

    Thank you Pedja,

    You know how are french…stubborn sometimes !

    Ah this time, I will cut carefully the legs of the opa861 before desolder it. Flux, and desoldering bread 😉

    Finnaly I saw I haven’t anymore value on one of the I/v resistor: : so on this channel both oaps must be fried !

    On the other I have something, but it distorse at ears, well better to be sure and also remove the two oaps !

    My main problem is to find something wich is giving an as good sound signature than the BG N (mate & neutral to my ears)at DC blocking ! Two Silmic II in reverse are good but not enough good for my personal taste in relation to the BG N ! (for the highs especially!)

    With 47 K at input of the pre, i would maybe try some non mettalized film cap for speakers or some BG NX (as the BG N are not sourcable anymore at desecent price ! )

    And a box not to touch it anymore 🙂 . Hey the AYA 2 2014 was playing nearly every day since it was finished !


  138. Hi Eldam,
    for DC blocking test a triple: Nichicon Muse BP with a small Styroflex and MKP, that’s my favorite.


  139. Hello Everyone,

    I finally finished assembling my AYA but I’m running into some issues. Here is a pic of the installation.


  140. Here is a close up of the board. I’m getting correct AC from the transformer, the power light comes on when I power up, but I can’t get the s/pdif to lock, didn’t try USB yet. I measured some voltages on the pins of the TDA 1541A, here they are :
    1 : 2.39 vdc 15 : -14.71
    2 : -2.35 16 : -11.01
    3 : 0 17 : -11.03
    4 : 1.52 18 : -6.55
    5 : 0 19 : -6.68
    6 : 0.01 20 : -4.61
    7 : -3.46 22 : -3.04
    8 : -3.86 23 : 0
    9 : -3.91 24 : -0.43
    10 : -5.18 25 : 0
    11 : 0 26 : -4.87
    12 : -7.51 27 : 4.9
    13 : – 2.2 mV 28 : 4.9
    14 : 0

    I’d appreciate any help to point me in the right direction



  141. Paul,

    Supply voltages are fine, but something is apparetnly wrong with TDA. Can you pull it out of the socket, and check what you have at BCK line (pin 2), so without TDA in place?


  142. Here is a close up of the chip. Voltaages with chip removed are as follows :
    1 : 2.38
    2 : 2.33
    3 : 0
    4 : 0
    5 : 0
    6 : 0.01
    7 : -0.15
    8 : -0.03
    9 : -0.01
    10 : -0.1
    11 : 0
    12 : -0.02
    13 : 0
    14 : 0
    15 : -14.64
    16 : -0.01
    17 : -0.02
    18 : -0.01
    19 : -1.05
    20 : -0.03
    21 : 0
    22 :-0.04
    23 : 0
    24 : 0
    25 : 0.02
    26 : -4.86
    27 : 4.87
    28 : 4.87


  143. For now I am not sure why CS8414 doesn’t lock, but without TDA the middle point of its BCK output is apparently fine.

    As for TDA, if you switch to standard DEM setup (so C324 in place and C213/C214 removed) it will be easier to narrow down the problem(s).

  144. Hi Pedja,

    Ok, I had thought that I had built DEM set up, I’ll double check and do as you suggest.


  145. Hi pedja,

    I just recieved the Toroid and TDA R1’s I must say they look brand new amazing never seen TDA’s like these, also can i get the names and address of all who ordered them so i can post them off asap?


  146. Hi Pedja,

    I removed C213 and C214 as you suggested, C324 was already installed, but under the board due to size constraints. I tested the DC on the board from AC1 through AC7 via R1 through R14 all were as they should have been. Rechecked voltages on the TDA 1541A pins with and without the chip in again. Much the same as before here is without :
    1 2.37
    2 2.33
    3 0
    4 0
    5 0
    6 0.07
    7 -0.01
    8 0
    9 -0.006
    10 -0.09
    11 0
    12 0
    13 0
    14 0
    15 -14.66
    16 -0.05
    17 -0.008
    18 0.007
    19 -0.052
    20 -0.056
    21 -1.002
    22 -0.042
    23 0
    24 0
    25 0.016
    26 -4.46
    27 4.88
    28 4.88

    Voltages with chip in to follow


  147. Here are the voltages with the chip in :
    1 2.38
    2 2.33
    3 0
    4 1.58
    5 0
    6 0.01
    7 -3.56
    8 -3.91
    9 -3.99
    10 -5.23
    11 0
    12 -7.52
    13 0
    14 0
    15 -14.66
    16 -11.02
    17 -11.03
    18 -7.38
    19 -7.51
    20 -5.47
    21 -5.01
    22 -3.96
    23 0
    24 -3.48
    25 0
    26 -4.85
    27 4.87
    28 4.87

    On power up the s/pdif led lights briefly then goes out. I also measured the pins on J202, pin 7(relay input) read 5v, pins 7 & 10 (relay output read 0. Tried hooking up a s/pdif source still no lock. I wonder if I have a bad TDA chip. Is there anything else that you can think of that I should check ?


  148. Peter, you have mail.

    Paul, there is definitely a problem with TDA. Can you recheck now with new chip, please?

    Also, before we try to solve possible problem with S/PDIF stage, did you verify your S/PDIF source to work properly (with working DAC)? In any case, S/PDIF lock LED should light regardless of input switching relay.


  149. Hi Pedja,

    I had the chance to try things again with one of the new chips that you sent out. Still not singing. With power on the power LED and D401 and D402 are lit. The S/DIF LED blinks on up briefly on power up then goes out. I also tried with a USB source with out luck. Both the S/PDIF source and USB work with other DACs. Here are the TDA1541A pin voltages :
    1 2.38
    2 2.38
    3 0
    4 1.59
    6 0.1
    7 -3.61
    8 -3.98
    9 -4
    10 -5.33
    11 0
    12 -7.65
    13 -7.56
    14 0
    15 -14.67
    16 -11.06
    17 -11.05
    18 -7.65
    19 -7.75
    20 -5.63
    21 -5.29
    22 -4
    23 0
    24 -3.47
    25 0
    26 -4.86
    27 4.87
    28 4.87
    By the way, I shipped all the other chips out today.


  150. Got the chip today. Thanks everyone

    On the other side, my board is mostly built. I found a non-working receiver to reuse.

  151. Power was ok, even with a mistake 150R instead of 180R at R408, R403(reworked later).

    other side

  152. It is surely great for the first build. In fact, if ABs and OS-Cons size was more appropriate, it would be great as such.

    TDA1541A DEM pins obviously don’t behave well with another chip too (pins 11 and 23). We will have to do some more checks, but it is probably better to do this offline, so I have sent you an e-mail.

  153. Once again, thanks to everyone who helped with TDA1541A chips distribution. We finished at 22 chips, and we are mostly done with USA and Canada, and half the way with other destinations. Oliver’s sample will remain with another AYA II builder from Zagreb, as he asked for it earlier, and chips were already there.

  154. Dave Downer

    Paul Norton, could you possibly contact me about your installation? I’m struggling some on mine and it looks very similar to yours.

    (831) 674-1895 Dave California (USA)

  155. Hi Dave,

    There have been many successful builds so far so I’m pretty sure I must have messed something up, or forgot something. Contact me at nortp2@gmail.com and we can go over our notes.


  156. Sylvain Bergeron

    I received the TDA1541A from Paul Norton in the US. I contacted the canadian guys. I should be able to ship the part next week. Take care. SB

  157. Hi all,

    It has been one year than Pedja Rogic re launched the AYA 2 with some layout improvement. The AYA2 2014 became my reference DAC till today. Still a lot of thanks for that !

    I worked on my own to find a setup which feet in my system and after a lot of work on acoustical jazz & classical musics with tunes I listen to since more than then years for all my setups & benchmarks I found something which imho Worth it and makes the AYA2 cross an important step (complex instruments like harmonica, piano or cello, bass don’t lie !). It was hard as I don’t have a reference pcb but as I work on the sames tunes since many years I have some habits with its !

    It’s not blessed by Pedja Rogic and one has to stay humble, but if your skill is good enough (because the pcb is fragile) here is my finding which add a Superior tonal balance and longer time on notes extinctions :
    on the -15V rail of the TDA1541 : swap the Panasonic 1500 uF by a Panasonic 1200 uf// on the legs with a Panasonic FM 330 uf (yes here FM one). But works only if you get out the 100 uF cap at the output of the shunt reg)
    You could say it’s just caps and very close… but it makes it improves it a lot, at least at home ! Of course I tryed many many things and Black Gate too !

    Just wanted to share this as it’s simple with no expensive parts. But if trying it: be prudent when removing the 100 uF cap : holes are fragiles : only for the experienced & skilled diyers :).


  158. Typo fault : the 1200 uf is a Nichicon FC (not FM)… this is crucial !

    So Panasonic FC 1200 uF // with Panasonic FM 330 uF

    (don’t forgett t choose the right voltage cap in relation to your traffo before : here 25V with Audial traffos or more if you have like me a 18V traffo ! Although the Panasonic FC are “sounding” better with 35v to 100V rating ! I use personnaly 50V Panasonic FC cap)


  159. Typo fault : the 1200 uf is a Nichicon FC (not FM)… this is crucial !

    So Panasonic FC 1200 uF // with Panasonic FM 330 uF (so FC + FM)

    (don’t forgett t choose the right voltage cap in relation to your traffo before : here 25V with Audial traffos or more if you have like me a 18V traffo ! Although the Panasonic FC are “sounding” better with 35v to 100V rating ! I use personnaly 50V Panasonic FC cap)


    Of course if you try it and think I’m bad, testimonial is welcome !

  160. Sorry to be long (don’t want to pollute as it’s just about tweaking)

    comments : the FM cap on the -15V rail can give a shorter (less powerfull low end) and one has to try before the FC cap alone (to understand this difference : listen to for example bass or piano pedal – low end).

    Than try to // the FM 330 uf cap : it will add a delicious mid & treble at the price of a shorter low end (but you have always this with the actual FM 1500 uF cap according to me : at least at home in my system : it’s not a truth writed in the stone)

    You can also try with a // FC 330 uF but the equilibrium is not the same but can give you some margin !

    And again this is without the output 100 uF cap of the BOM (which is important according in my tests).

    (adding // Panasonic FC caps on the reservoir caps could be an alternativ in settuping but very too subtle at the opposit of the -15V rail tweak)

    voilà !

  161. I would like to add some thoughts on Eldam’s posts.

    The first, please note that the local capacitor at TDA pin 15 was removed. So, there is no capacitor after -15 V regulator at all, regulator itself does all the work here. We have mentioned this option before, and here is some more information about it.

    This regulator has very high bandwidth, that reaches several tens of MHz. It means that, within this range its output impedance is about 0.65 Ohm (with 40 mA typically flowing through Q303), and its line attenuation is about 60 dB. Capacitor at its output takes the place at higher frequencies, and basically a few tens of nF is enough to work at 20 MHz or so and above. Reason why we use up to 100 uF here is to lower that 0.65 Ohm impedance at lower frequencies too. This capacitor also provides line attenuation of its own. That is the technical part.

    Now, for subjective experience. Everyone playing with different capacitors is probably aware of how important these local capacitors are. But how the things work here without this capacitor at all? It might be enlightening to hear how many things actually sound better without it, like instruments separation and definition (not in the “forensic” but in the musical terms). And there is more breath. Shortcomings come as certain lack of bass, and more forward presentation in a way I am not sure I like, and sometimes there is more grain.

    Now, the question is normally, could we solve these problems at some other point, say before the regulator, so we have the best of both options?

    We obviously can not solve the problem with increased impedance above several tens of MHz elsewhere, but we do have a smoothing (reservoir) capacitor prior to the regulator, and resistors that form common mode filter make the good basis for attenuation by this capacitor. So if shortcomings came because of the lack of line attenuation at HF, we can try to solve it here indeed.

    Eldam changed FM 1500 uF with FC 1200 uF // FM 330 uF (330 uF will normally act better at HF than original 1500 uF). And I do agree that such a combo provides more deep bass, which comes here really for good.

    So, removal of the capacitor after the regulator really opens up some options and possibilities. And it opens up even more questions as well. 😉 If one wants to experiment, here is the place to do that. I will personally have to try and listen to such and similar set ups for some time, to make more sense of all this.

  162. Hi Pedja,

    For a reason I don’t understand but I already experienced elswhere, Panasonic FC at equal capacitance value gives often a subjectiv lower end and an emphasis on high-bass (say mid-bass)the FM has not despite its better ESR (relativly close here in its voltage and capacitance values though). But FC serie also give a different general tonal color elswhere : medium and highs, I subejctivly found better here (at least in my system and with my personal tastes ). Is it due to the different tonal balance it brings because the added bass ? I don’t know !

    for AYA2 2014 owners interested to do it :

    If one want to experiment, maybe the 330 uF cap is a challenge : adding 330 uF Panasonic FM tamed a little the bass again while staying much better in this frequency range than the 1500 uf FM cap alone. Luckily the mid-bass energy is keeped and a sense of rythm is added because due this time to the low cpacitance value.

    330 uF with Panasonic FC should be experienced as well(have not on hands yet !)and certainly some other brands to find a better equilibrium like a “darker” soundstage maybe(words are always hard to choose when it comes to describe in a subjectiv way what is heard !). In another way, FM serie cap has good points for itself as a “airy 3D soundstage” here… at least subjectivly controlled by ears as a cap alone on its own has no sound :)!

    The difference in internal construction makes these caps sounding different in a circuit ! More than the exactly 1200 uF value : this is the FC serie which is important as well while the lowest value stays important as Pedja Rogic explained objectivly (on a technical/scientific way). It was a targett than the two capacitors makes 1500 uF togethers to have a proper benchmark.

    Finding THE right 330 uF brand/serie would ask some works and a margin to match your whole system while the above given combo brings already a lot of pleasure… at least to me.

    Not having an other reference fix pcb for A/B tests, I wait as well the final confirmation of Pedja about the musical feeling about it ! A musical device is not immediatly percibtible like the introduced change is, so ask time to be checked seriously. And here I’m biased !


  163. Hi,

    Very interesting and thanks to Eldam and Pedja.

    I wonder if using a modest value of SMD poly-somthing cap as the local decoupler (couple of uF only) at the SMD bypass pads would be worth experimenting with.

    Thanks again.

  164. I powered up. No smoke. No sound.

    Before any of the chips went in, power was good.

    After the chips, one of the CS voltages was ~0.8V. Not good. Taking a chance, I replaced Q204. Same result. Without Q204. R205 showed ~5.6V. With Q204, ~1.8V.

    I am wondering now if it is a faulty CS8414.

    Strangeness on the USB side also. Computer recognizes USB, sends music. However, nothing on the DAC pins. Before the relays, signals look good on the PCM side of the resistors. On the relay side, of the resistors nothing. Did the relays get fried also?

    Any thoughts? suggestions?


  165. @ bn : sometimes the sound of the main output of laptop speaker has to be switched off (desactivation) on the Windows audio interface to have sound with an usb device (despite you see this usb is playing with the level graph on the screen of your laptop : Windows audio interface is a mess ! !)!

    @ Shane !

    If you keep the smd local decoupling at -15V output : your mid-bass will have more dynamic and “impact”, the medium more transcient and dynamic and transparency like the treble as well ! But the mid-hi- end could have a sort of light which could be fatiguing !

    That’s why I talked about careffully choose the 330 uF ! My conviction is that 330 uF would work better with a Panasonic FC but not a FM if the 0.1 uF cap is omitted ! But Pedja Rogic explained why the highs of the impedance curve are so important and the reason of a final caps after the shunt near the load.

    I use myself with good result 0.1 uf 0603 smd case class I ceramic from Murrata in a serie with lower piezzo effect (pcb cracking) with good results (as far as you select the good main caps before the shunt). My first priority was here to have the best inductance I can solder with my eyes 🙂 !

    I’m sure the dielectric of this local smd cap, even with low 0.1 uF has a sound influence !
    One can try : PPS caps; , bigger styroflex cap with tin foil. non metalize KP, etc…. Inductance will change but also the dielectric affect the final result ! A trade off has to be founded in relation to your whole system sound signature ! At least my humble vision of caps tweaking.

    The importance thing is the final decoupling cap doesn’t affect both the transcient of the shunt and the final tonal color : so this 0.1 uF (or more ? or Less ?) should have a very good inductance. As a good dielectric : Ceramic class I : most of the time avoided in audio because cracking on the pcb, less than class II X7R though); Styroflex; KP (I hate MKP but some rare good enough in speaker filter serie); PPS; acrylic (same spec as PPS, maybe a different “sound signature”); MKT (with 2.5 mm pitch leads); and maybe Vishay SAL-rpm (an aluminium but very low ESR)
    I don’t talk about Polymer but some coud certainly feet : Panasonic SEP (not sepc); UCC SA serie, But here the low capacitance value of 0.1 uF is with higher esr but with the new little can : inductance could be good (2.5 mm to 3.5 mm pitch between leads) !

    Choose your poison : ) Personnaly despite the bad inductance of a huge Teflon cap I bet for a very good subjective sound result !

    The wiser thing is to swap the after shunt 100 uF cap and keep your smd caps if you have it already… try my combo tip before the shunt, then listen to 🙂 🙂 🙂 !

    I hope Pedja will come back with more tests of his own and testimonials for a less biased advice (and with a technic point of view) than mines, but I can promiss than the difference is huge enough: I find the improvement to be an important step than owner should try !

    One can ask if the 100 uF could be re populated with a cap of the same value and same pitch leads because inductance but with the “speed” of the caps above ! In my opinion the dielectric is important too and few polymers sounds good to me. Lately I founded the beloved OSCON-SP (discontinued) to be a little too coldy here ! And finaly the swap of the 100 uf final cap to be a greater improvement than any caps, even BG N ! At least in my hifi system !

    Sorry to be long, but it’s about sharing 🙂

  166. Ah. There is music 🙂

    After many missteps, the culprit was the u.fl connectors. These were shorted. Once I removed them, I was able to play via USB. I still think I have a defective CD8414. I’ll try to replace it.

    Thanks Pedja.

  167. Thanks to Eldam for the reply, and to Pedja for accommodating these discussions.

    What you are suggesting is ok for me to try. As they say, nothing ventured is nothing gained.

    Best Regards.

  168. @ bn sw: Yes, the load on Q204 is a problem. If all the soldering is fine there then I would also suspect CS8414.

  169. Hi I would like to use external Dem clocking, I have followed the instruction manual regarding pin 5 link on the 74hc590 and removing cap C324 at pins 16 and 17 and feeding I2s in simultaneous mode with Ian’s I2s to PCM, I am a bit confused about the BCK input is there 2 BCK inputs, one to the 74hc590 and one to pin 2 of the TDA 1541a chip in DEM mode?



  170. Yes, you should always send BCK signal to TDA, because TDA always needs it to clock the data into its input latches.

    When using external DEM clock circuit, you should also send BCK to the HC590 input. So yes, in this case BCK signal goes two paths. Alternatively you can use MCK here, or even LRCK/WS if it was 176.4 kHz. But it is usually simplest to use BCK.


  171. Peterma,

    Did you use either of the sources on AYA for the I2S input to I2S to PCM board?

    Would you mind sharing with us how you did this ?


  172. Hi bn sw I don’t use the on-board inputs, I am using a DIYHK USB to I2S board with Ian’s I2S to PCM board in simultaneous mode using the On-board AYA DEM, very happy with the sound, just noticed how the treble sounds just so natural and also able to get 16bit 196khz is amazing

  173. Hi Everyone,

    After putting my AYA aside for a while after I could not get it working I picked it up again. I went over it with a fine toothed comb and still can’t get it to work. I am admittedly out of my depth with this project. Is there a good Samaritan in the US who could take a look at it for me. my email is nortp2@gmail.com. Here are a couple of close ups.


  174. Hi paul

    I don’t know if you have done this yet but you have to cut two through hole pads connections, according to the errata information sent by e-mails on 23 and 30 October (the second one applies to the problem with output stage negative supply).I forgot about the mods needed, I had this problem at first but ok since.


  175. Hi Paul

    Sorry I don’t know too much technically, but maybe you could establish firstly if the TDA circuit is working ok by trying a direct i2s input, and direct out bypassing the op amp section somehow into another op amp, my tech level is average sorry I cant be more help, fault finding is tedious.. good luck don’t give up this dac is just amazing.


  176. I followed the errata instructions and cut the traces as outlined. But I went back to double check anyway but can’t seem to find the errata pictures. Does anyone have copies ? I do have a way to try I2S input, I have a ministreamer I can hook it up and give that a whirl.

  177. Hi Aya team. 🙂

    It looks like there is some interest generated for the AYA boards.

    Would you be considering a run of group buy for the boards or kits in 2015 ?

    What the reasonable number of boards or kits to run this GB 🙂 ?

  178. Hi guys just thought I should show my finished AyaII it looks rough looking down but is ok when lid is on.

  179. Hi everybody,

    I would like to continue to share my last setup to push the limits of the AYA2 on the side of the passive parts, because also a fellow asked me yesterday what are my last inputs.

    So this is an erata of my last input about a setup for better low end and mid bass with noticeable better transparency in mids & treble.

    – remove all the 100 uf (even Black Gate N or NX if you have it as I tried also !): the 3 of the DAC output shunts and the 4 of the analog output shunts.
    Remove the 1500 uF by Panasonic FC for the -16V DAC and the two analog. It’s very important also, even if you think the actual ones are better on the paper. Take care to the pitch legs when order it (diferent sizes)
    – For ground decoupling near the DAC : just use the -16V smd pads with a value you will test yourself before 0.022 uF to 0.100 uF. Avoid ceramic X7R which add a sibilance for benchmark (by yourself) Pan PPS vs Ceramic NPO.
    – Keep the Allen Bradley of the BOM (Hificollective)
    – Choose for I/V : Rhopoint wirewound 516G at Farnell (ohms value avaliable a bit lower but good enough)

    I personnaly trimmed the oap at 0V and supressed the Silmic or BG N for DC coupling.

    All those tweaking are with external simultaneous mode and external USB to I2S boards.

    For a reason of ground decoupling I can’t understand, using only the -16v smd decoupling give a far greater result in mid-bass and snap which is bader with decoupling on the -/+ 5V (tried alone and with both with values between 1 nano yo 100 nano).

    On the -16 V, trying 0.022 uF to 0.100 uf woul help you to find the acurate setup in relation to your whole system.
    Bypassing the main Panasonic FC (again: important to swap the BOM with the FC here) with something extra bypass from 300uF to 1500 uF of a different cap can help to tune the sound on both the digital and analog side at the cost of less snap (that the FC adds) but with a softer mid-high ends : dépends on your whole system.

    You really have to try it,

    Good sharing, regards


  180. Hi Paul,

    I think the workarounds are not needed for the white PCB. I thought that was for the earlier rev. Please see PCB revision history in the assembly manual.

    I have only the USB input working so far. I’ll contact you over email.

  181. This thread has been dormant for a while.

    Well, after waiting for the WaveIO for quite some time I somehow lost the momentum to continue on my AYA DAC.
    Now I have picked up the work again and connected the WaveIO with its own power supply. Connection is via uFL cables, unisolated output to J301-303. Checked voltages all around, everything seems to be fine and both output stages responding nicely to the trimmers. After powering on I set DC offset to 0… after leaving it on for a while this changes quite a bit?

    I am running the DAC with the Spdif stage unpowered, so it probably needs the CS8414 for some signal. But I also have no sound.. 🙂
    WaveIO is detected by my computer, drivers are installed.

    For your information here the voltages around my TDA1541A chip:

    1 0
    2 0
    3 0.53
    4 1.65
    5 0
    6 0.01
    7 -3.52
    8 -3.8
    9 -3.8
    10 -5
    11 -5.2
    12 -7.22
    13 -7.1
    14 0.01
    15 -15
    16 -11. 2
    17 -11.1
    18 -6.9
    19 -7.1
    20 -5.2
    21 -5.0
    22 -3.86
    23 -3.8
    24 -3.4
    25 0.42
    26 -4.9
    27 5
    28 5

    Anyone knows where to look further?

    Thank you.

  182. Hi Oliver,

    Can you firstly check without the uf-l cables, the uf-l plugs you soldered on the pcb !

    Remove the dac chip and with the buzzer function of your voltmeter, check if you have any curent between the center of the uf-l plug and the pin hole which it supply on the chip socket (pin 1 to 3)

    WIth this USB to I2S, you must play into Windows audio interface for a while to have a valid output : can be a mess, I spend ten minutes to understand (in Windows 7 /10 : you must see the icône of the USB to I2S card with volume level which move !

    A good way to see if your tda1541 feeds the OPAs is to check the off set : look at this blog before : The TDA1541 must have a clock signal if I remember to allow AOPs to “see” the dac chip (no needs to play music from your laptop) and to allow you to see the low mili voltage you need to adjust near zero (Read the manual again).
    But I believe your dac chip doesnt see the USB to I2S pcb (I have the same). Please also check twice the outputs names : use first the non isolated ufl- plugs (btw it is the best outputs of the two !)


  183. Thank you Eldam for your input!

    Will do as you suggest and try the connection between the uFL connectors and the DAC chip.
    As a player I am using Foobar with the ASIO out, so probably I won’t see any Windows output?
    Do you have the Spdif section powered while using an external USB to I2S device?

    Bonne soirée,

  184. I re read your first message…more carrefully !

    If you can trim the OAP, it means he TDA1541 sees the MCLK !

    If you see zero on pin 1 & 2 while music is playing : two possibilities :

    1- TDA1541 doesn’t see a signal on pin 1 and/or pin 2 : check the ufl plugs on the AYA2 (but the one for MCLK as you have a signal)to see if they see pin 1 and pin of the TDA1541 socket (cold joint, uf-l plug with a bad orientation).

    2- or your PC doesn’t send to the Wave I/O due to a bad utilisation on the Windows sound interface (this is a mess !)

    If after checking the center (hot point) of the UF-L plugs on AYA 2 to the dac chip socket and find no problem, you should ask to Lorien directly because you enter in the second issue.

    By default Wave I/O use USB input. So if you use SPIDF input on the Wave I/O you have to read its manual to use the dedicated jumper.

    You don’t need CS chips for SPID/F if you use the SPID/F from Wave I/O ! or do you mean you use the electrical relay ?

    Also : check twice you didn’t use on the AYA 2 the uf-l plug made for the simultaneouse mode (pin 4) with the TDA : this is why I told you to use the voltmeter to buzz between the uf-l plugs and the sockets to check it !

    With any source playing the TDA pin 1 & 2 should have no zero voltage !

  185. Eldam,

    some more good info!
    I have not checked voltages while music was playing, so that might explain zero voltage on pin 1 and 2?? Right?

    Will check that first and then inspect continuity of cables…


  186. Played music and measured voltages on pin 1 and 2… still 0.

    Then checked the connection from the uFL contacts to the chip… got continuity for J305-307, so that works as intended.
    Other contacts have no continuity but that is because I obviously set wrong jumpers on relay 2 🙁

    This leaves the WaveIO not doing its job… which is funny because I am using another one just fine (USB –> Spdif out).
    Have to check with Lorien.

  187. Maybe you have just a problem with the Windows interface ?

    Wave I/O are checked before shipping, so I assume if you have not a problem of configuration on it : jumper to work in TDA1541 mode or Windows interface problem !

    You can also check the three I2S continuity from this board to the TDA socket (not playing)by putting the buzzer in the front of the uf-l pad where the hot point is (so it check also the wires and check if a jumper or relay stop the signal if you work with it).

    You should also post two picture of your AYA2 : front and bottom !

  188. Hi,

    I just thought I should mention that I am using an Odroid c1+ connected with I2s and I am very happy with the sound excellent results better than the usb to I2s card I was using and cheaper to buy.

  189. OK, the story continues….

    My AYA DAC is still not playing music… 🙁
    According to the hints of Eldam I checked a few things and am at the following now:
    I have a voltage reading on pins 1 and 2 (2.38V). I have continuity from J301 – 303 to pins 1 and 2 on the TDA.
    I added a ground connection from the WaveIO main ground to the ground point near the chip (between LIO and RIO, Lucian’s recommendation). WaveIO is being detected by my computer, should be no problem since I have a second one in my Subbu DAC (that I just sold to France 😉 )

    Things I noticed: the output stage’s offset can be dialed to near zero and stays there. The transistors Q 402 and 405 are getting warmer than I thought they should… really warm to the touch.
    I checked all resistor values around and they are correct.
    Any hints on that? I appreciate all input since I want to get this thing playing!

    Thank you.

  190. Hello Oliver,

    Sorry to hear the problem is still not solved… especially because the board looks this cute. But I am sure it can be done. 😉

    The current sources for the output stage typically supply 55 mA (there should be something like 1 V across R401/R406), and about 35 mA of this current is distributed between four 861 chips, and shunt parts preceding these transistors. These transistors then sink the rest of the current, so it is normally about 20 mA, which at 5 V equals 100 mW, and which makes them barely warm, but not that hot so you can not touch them.

    So if these transistors are hot, and current sources are fine, then there might be some problem in the output stage. You can make quick check of output buffer chips. They usually have 2.5 – 3 mA (negative) voltage offset between the input and output. You can verify this by putting DC Voltmeter to mV range, and checking between R417/R418 (either side) and pads preceding the output coupling capacitors.

    Also, you can check what is happening with TDA and I/V chips, for the start by removing TDA chips. Assuming the TDA has proper I2S signal at its input, its removal should change the output offset for 3 V or so. You can of course check this at I/V resistors.

    Please let me know the results, and feel free to drop me an e-mail, if you believe it can be more effective.


  191. Oliver,

    I’m sad your DAC is not working already. I made also mistakes, fried some transistors and Pedja Rogic’ help was of a first importance to allow mine to work !

    If you want, don’t hesitate also to pm me via Diyaudio or my private mail you should have if I remember ! No problem, I have extra time and maybe in the several errors I made, some answers would help you !

    It’s very hard to debug with distance… but with photograph and re checking the key points (voltages) it may allow you to suppress the main issues ? I don’t know, but you’re welcome if needs an extra beginner point of view 🙂

  192. Houston, we have music.

    With patient help and the systematic routine of Pedja I was able to trace the reasons that kept my AYA DAC from making music.
    In the end it was just a faulty relay and two uFl-sockets soldered 180 degrees off (shame on me). On the way I also learned to take reasonable measurements with my new scope!

    Thanks Pedja!!

    Also merci beaucoup to Damien for his offer to help… will come back to you when I start modding. 🙂

  193. Good to hear that Oliver,

    If you use the Wave I/O in direct on the AYA, I have a good mod for the 2 analog main caps : Panasonic FC 1800 uF /35v (both value are important) that you can challenge with the FM 1500/16V of the BOM: different presentation… in relation to the rest of the hifi you have (it gives you a good choice). Of course if you have big Black Gate, jump on it ;)!

    (mods change in relation to each Front end)

    Yeap the buzzer is a very usefull tool to see the continuity of current ;)or open circuit !

  194. Had my AYA DAC playing for a longer period of time now. It has a sum of qualities that make it reproduce music on a very high level. At most how natural it sounds without loosing or masking detail!

    But after a while the quality of the upper bass region appeared too thick to me (on Hifiman HE500 headphones since I am a headphones only guy :)) I saw the other thread is basically all around tweaking, and I saved all important information for further ‘mods’. But for now I simply exchanged the Silmic output caps to my favourite old ERO KP caps. I still have many of them stashed away in my drawers, so I found two 3.3uF and put them in.

    They certainly do the trick because they retain the overall character of the AYA but are cutting away the belly of the upper bass and have no harsh character at all! Really pleased now with the DAC but might try some of the other mods.

    What is to be gained with leaving C402 – 404 away? For C407/408 I used old styrene from the beginning.

    Then I performed a little heatsink mod to the CCS transistors Q401/404 that were running a bit tto hot fpr my liking –> see image.
    Working perfectly!

  195. Hi Oliver,

    That’s a really pretty board, certain components like the ERO film caps pop out against the white board.

    I’m curious, what type of connectors are those for the AC wires that connect to the board? They look really neat.

    Thanks for sharing..

    BTW, on the subject of film caps. Has anyone tried high value 200V Multicap RTX styrene/tin foil caps? They can be had at a discount from certain sellers. How do they sound with the Aya II?



  196. Hi Frederick,

    thanks for your reply.
    Since I expected to work on the board I didn’t want to solder the AC wires. The connectors are from WAGO when memory serves me right.. actually nothing special but they go nicely with the white PCB. 🙂

    As for caps to try we are on subjective territory, so to speak. Everyone has his/their own perception of things, so judging caps in this position is most likely up to you.

    In my opinion bigger or more expensive is not automatically better, so I would probably start with a decent MKP cap (Jantzen?)first to hear the difference to the Elna Silmic and go from there.
    The ERO caps I know from other projects so I “knew” what to expect, more or less.


  197. Hi Oliver,

    Thanks for the tip on Wago… Yeah, caps are highly subjective. Lots of variables within a system to consider.. Thanks again.


  198. Hello,

    I have please a question about my AYA II 2014: If the I/V R is 1K instead 1K5, do I need to swap the 2N2 shunt cap after the trim pot to another value for filtering behavior (can it be hearable ?) ? 1N5 for instance, or it doesn’t matter at all ?

    Thanks in advance and sorry to dig out this older thread by an OT question

  199. Hello Eldam,

    It is good to know that the AYA II 2014 still keeps your attention, and is being tweaked.

    Those 2n2 caps are not related to I/V resistors. On the other side, if you decrease I/V resistors from 1k5 to 1k, you may proportionality increase C407/C408 capacitance, so you preserve the same low pass slope. This slope itself is however not really critical, so you can leave all the caps as they are, as well.


  200. Many thanks Pedja,

    It is still my every day DAC 🙂

    I ‘ve got a TAIWAN 98 TDA1541A chip… indeed very different than the others older one from Europe. I don’t know if mine is genuine though.

    best regards


  201. Long live the Aya ll . Really appreciate your effort in bringing
    this wonderful dac to us Pedja. Thank you
    Thie dac is highly tweakable minute changes brings different
    results. Like Eldam I’ve been tweaking Aya. Thought that I
    reached the peak of tweaking until tonite I tried something
    wihich really brought an improvement beyond my belief

    Thanks again
    PS wish that the rest of you guys will tweak Aya as well cause
    you really don’t know how great this dac is till you do so.
    Just red book play back will eat Sabre dac etc for breakfast
    No need DSD

  202. Hello,

    I would like to ask if anybody has experience with shielding a wood cabinet with steel sheet (magnetic).

    I know steel as made from iron has good properties for traffo or traffo compartment shielding (the iron attract magnetic fields so stop a part of it if between the source and the area to protect (here the AYA & its digital front end)

    I also read than if Mumetal was good at low frequencies (say few Hz to 20 K) because its high permeability, steel was better for Frequencies above 20 K Hz !

    Does it Worth to make a case around our external digital front end as the USB to I2S board for instance ? Or does the effect of magnetic field catching will worse the things if too near from the ICs?

    Have a 0.7 mm thick steel big layer lying around… and also a plane rod of the same material (4 cm large for 2 mm thick to putt between the traffos (emitive source) and the AYA & digital boards (area to protect…)… though it will be just a 4 cm eigth wall !

    Have to say my AYA II cabinet is from wood and not the usual amagnetic aluminium we can see (or copper Inside sometime)… so my questions (have no hum btw and all my traffo are 90° from each others on a line at 15 cm from the ICs (but the digital front end a little nearer!

    Thank you a lot if any ideas



  203. Hello Eldam,

    It all depends on what you are starting with. If you don’t have to block some really strong radiation, you might simply abandon shielding.

    Anyhow, regarding mumetal, as far as I know it is effective at higher frequencies too, even though the copper is still better. And it is non-magnetic. Shield box made of mumetal, on the other hand, will protect from magnetic field.


  204. Thank you Pedja for all those inputs.

    So I will try just to putt just a wall between the power traffos and the pcbs as I saw in your Model S pictures, although I have just steel, no copper or aluminium either.

    Maybe i will try a steel wall between each traffos to avoid coupling (?) as I see your two EI traffos have casing around their core.

    In case of no différences at ears, I will remove it (I don’t know if C-Core radiate the same ways as EI traffod. Btw have steel your two EI traffos to experiment with, though no Ground Earth for the secondary CT of the isolator transformer yet in my living room.

    I also abandon the idea of casing above the front end.


Leave a Comment


Scroll to Top