Good day,
In another DAC thread here on DIYA member Soundcheck mentioned this DAC and qualified it as surprisingly impressive, or so. Even I am a very satisfied owner of a sligthly modded excellent AYADAC with TDA1541 S1 dac and I was currently playing around with IANs newest thingies, I thought I need this dac. Not at least that I consider Soundcheck being a very critical mind.
It helped that this is more a budget dac for the raspberrypi.
As soon it arrived it was plugged into my DCDC converter less Pi and yes, it really surprised me. Even more, when I put the FifoPi board in between.
What a threedimensional soundstage, what an involving sound. I rarely "hear" my speakers and with this source they completly dissapeared. The room expanded far to the rear. The word 'real' came into my mind. I started to understand Abraxelito and his being dedicated to the TDA1387 chip.
Both my other dacs, especially my AYADAC, are excellent dacs and in any case more detailled and have some other preferrable things, still this one is very special in its own way and I would like to reveal more of its potential and this is why I opened this thread.
I would like to share my findings and receive info from members they already went down the road with this dac or its simpler, but otherwise similar siblings that are only to get from Asia.
My plans so far are:
*reading more about Abraxelitos dacs.
*Bringing the curcuit to paper in order to understand this implementation. Has anybody done this?
*I'd like to have it run in synchronous mode instead of ASR. Is removing the VHC574 the only way?
*I'll change the I/V resistor to lower value and a better resistors. First lower and then better. Rhopoint maybe since I am happy user of them in my AYADAC.
*Thinking of an active I/V stage, if this is the outcome from reading Abraxelitos thread.
*Later I want to try Ivans transformers I got for the TDA1541
Maybe something else.
Please, let us keep this thread to mainly the TDA1387 in this application and having some input from Braxis findings.
And, let me mention, I am slow and like listening to music. So the whole process might take some time.
Cheers, Ernst
In another DAC thread here on DIYA member Soundcheck mentioned this DAC and qualified it as surprisingly impressive, or so. Even I am a very satisfied owner of a sligthly modded excellent AYADAC with TDA1541 S1 dac and I was currently playing around with IANs newest thingies, I thought I need this dac. Not at least that I consider Soundcheck being a very critical mind.
It helped that this is more a budget dac for the raspberrypi.
As soon it arrived it was plugged into my DCDC converter less Pi and yes, it really surprised me. Even more, when I put the FifoPi board in between.
What a threedimensional soundstage, what an involving sound. I rarely "hear" my speakers and with this source they completly dissapeared. The room expanded far to the rear. The word 'real' came into my mind. I started to understand Abraxelito and his being dedicated to the TDA1387 chip.
Both my other dacs, especially my AYADAC, are excellent dacs and in any case more detailled and have some other preferrable things, still this one is very special in its own way and I would like to reveal more of its potential and this is why I opened this thread.
I would like to share my findings and receive info from members they already went down the road with this dac or its simpler, but otherwise similar siblings that are only to get from Asia.
My plans so far are:
*reading more about Abraxelitos dacs.
*Bringing the curcuit to paper in order to understand this implementation. Has anybody done this?
*I'd like to have it run in synchronous mode instead of ASR. Is removing the VHC574 the only way?
*I'll change the I/V resistor to lower value and a better resistors. First lower and then better. Rhopoint maybe since I am happy user of them in my AYADAC.
*Thinking of an active I/V stage, if this is the outcome from reading Abraxelitos thread.
*Later I want to try Ivans transformers I got for the TDA1541
Maybe something else.
Please, let us keep this thread to mainly the TDA1387 in this application and having some input from Braxis findings.
And, let me mention, I am slow and like listening to music. So the whole process might take some time.
Cheers, Ernst
Greetings ernesternest,
Do you think this board deserves a trip vs an AYA DAC already tweaked ?
If yes I could be tempted for some diy with it. I suggested Abraxalito to put 3 I2S uf-l plug on his tda1387 kit but he didn't answer about that !
Cause first thing will be to know the gap between the Audiophonics board and Abraxalito's gap !
Has the Audiophonic's a good slam, tight bass and a good tonal balance according to you ?
Anyway, thanks for the tip
Do you think this board deserves a trip vs an AYA DAC already tweaked ?
If yes I could be tempted for some diy with it. I suggested Abraxalito to put 3 I2S uf-l plug on his tda1387 kit but he didn't answer about that !
Cause first thing will be to know the gap between the Audiophonics board and Abraxalito's gap !
Has the Audiophonic's a good slam, tight bass and a good tonal balance according to you ?
Anyway, thanks for the tip
I've just done a relatively 'quick and dirty' modification to an AudioPhonics board for another interested DIYer. I soldered 16 extra DAC chips on top of the 8 already there to give a total of 24. I removed all the output filtering and also the VHC reclocker and oscillator. I fed the output current of all those paralleled DACs into some small (EP17) ferrite transformers and then after that, a CLC filter followed by a resistor for passive I/V. The last stage being an SE MOSFET classA buffer.
When feeding a transformer single-ended there's a need for a compensating current as trafos don't much like DC. So I arranged current sources on the secondary side to compensate for the primary current offset. Overall the sound was really enjoyable, though to my ears, running true balanced into the trafos is an improvement.
When feeding a transformer single-ended there's a need for a compensating current as trafos don't much like DC. So I arranged current sources on the secondary side to compensate for the primary current offset. Overall the sound was really enjoyable, though to my ears, running true balanced into the trafos is an improvement.
Good morning Diyiggy,
Let me answer your last question first. Yes, I think it has slam, tight bass and a good tonal balance. And it has good mids and to my surprise it has good heights.
Short OT, my AYADAC ist currently out of order, as I damaged Pedjas USB board and it needs repair, which will take some time. This led me to test the AYADAC with the FifoPi and Ryans i2s2pcm board which was working perfectly and anyway I was tempted to try something different. So I ordered this board.
I have already given a short impression of the sound and I want to add, that maybe 30 years ago I had a system with a just good CDP, a simple DIY hybrid (E88CC imput - IRF(9)530 output) amp and Fostex Sigma103 horn speakers. This simple system let me have a view into the rooms where the music was recorded or at least I thought about it in that way. Then I tried to modify the amp, made a mistake and I never was able again get this amp to sound like it was. No idea why.
Strangly this specific DAC upon my Pi in my current system (Bastanis speaker- they are able to do what the S103 horns were able to do but on a bigger scale, USSA-5 amp, really great, Raphael tubepreamp - old, simple and pretty good) reminds me of this old system and pulls me into the music.
AYADAC is more refined, more detailled, somehow cleaner and IANs9038 is even more of that and anyway I think it will give me great pleasure trying to keep the dacs basic abilitys while trying to refine it.
Is it worth it? For me absolutely yes, and it is rather cheap fun.
Cheers, Ernst
My first post in the forum, so be all greeted from the north of Spain.
Ernesternest, me too I am so impressed by this DAC that I was about to start a thread just like you have just done.
Abraxalito, thank you so much for all the work you have done and documented, I learned a lot.
I have two boards of this DAC, one of them I intend to keep as passive I/V, and the other one I plan to go active.
On the passive one I have made some minor modifications, lower and best quality I/V resistor, short the 2,2uF cap (this removes bass boost), and exchange the output caps for some film ones. Nontheless, the biggest improvement come from powering the DAC with the Salas Simplistic regulator.
Attached is the passive I/V and filter with the mods I made. The ling8DACfilter is the work of abraxalito and what I intend to do in the active unit.
Ernesternest, me too I am so impressed by this DAC that I was about to start a thread just like you have just done.
Abraxalito, thank you so much for all the work you have done and documented, I learned a lot.
I have two boards of this DAC, one of them I intend to keep as passive I/V, and the other one I plan to go active.
On the passive one I have made some minor modifications, lower and best quality I/V resistor, short the 2,2uF cap (this removes bass boost), and exchange the output caps for some film ones. Nontheless, the biggest improvement come from powering the DAC with the Salas Simplistic regulator.
Attached is the passive I/V and filter with the mods I made. The ling8DACfilter is the work of abraxalito and what I intend to do in the active unit.
Thank you ernsrernst for the quick review...very impressive for a little board with a basic supply. When we see what such people like T Loesch from AMD dit with the tda 1387...one may want to work on it... I would like to see Abrax board to have ufl inputt for usefull front end tries...soldering may change the quality at each swap...hard for a proper front end benchmark on those smartda1387 pcb...
I'm gonna to try Andrea Mori's good clocks on my Aya when the new oscillator is ready...could be nice to try it with tda1387 as well...fun!
Cheers
I'm gonna to try Andrea Mori's good clocks on my Aya when the new oscillator is ready...could be nice to try it with tda1387 as well...fun!
Cheers
I can't quite figure out the need for minature coaxial connectors on I2S. Could you please explain why you feel they're necessary @diyiggy? Soldering joints isn't an issue that I can see - if you feel solder joints change significantly in quality each time they're made then put forward your case please 
Hi Abrax...did you finally try the Panasonic caps ?
Well, it's usefull to have each time the same length of i2s cables, to have them coaxial for a better and individual current return path...no doubts when connecting several front end which are an important side of the whole result.putting ufl since the pcb drawing gives the insurance of a good individual return path for each of the 3 signals till the inputs of the dac chip.
You can easily swap front end boards to check dac boards mods versability, chaging for example the quality of the soldering past is heareable with good hifi.
You can use iancanadas, Andrea Moris, Rogics and so on best diy avaliable front end.
It s good for you to benchmark your work with others dac board with the same exact connexion with the front ends.
Avoiding ufl is a false economy in our diy life.
Will be proud to try your dac if it had it...
Well, it's usefull to have each time the same length of i2s cables, to have them coaxial for a better and individual current return path...no doubts when connecting several front end which are an important side of the whole result.putting ufl since the pcb drawing gives the insurance of a good individual return path for each of the 3 signals till the inputs of the dac chip.
You can easily swap front end boards to check dac boards mods versability, chaging for example the quality of the soldering past is heareable with good hifi.
You can use iancanadas, Andrea Moris, Rogics and so on best diy avaliable front end.
It s good for you to benchmark your work with others dac board with the same exact connexion with the front ends.
Avoiding ufl is a false economy in our diy life.
Will be proud to try your dac if it had it...
I tried the Panasonics yes, nice but certainly no better than the Nichicons. But they were cheaper so a good call from you.
I am not a fan of having multiple large ground loops which is what happens when an individual ground return is provided for each of three logic signals. But seeing as you like those connectors what would you have against laying out a small PCB which takes in 3 uf-l sockets and puts out pins which mechanically line up with my latest paralleled DAC board?
I am not a fan of having multiple large ground loops which is what happens when an individual ground return is provided for each of three logic signals. But seeing as you like those connectors what would you have against laying out a small PCB which takes in 3 uf-l sockets and puts out pins which mechanically line up with my latest paralleled DAC board?
I certainly don t add a pcb more between a front end and a dac board.
The shorter the better as you know, for mono dual boards some options exists like Iancanada but not only.
I prefer to hace an exact impedance between boards ufl cables permitt...hoping 50 ohms is yhe impedance chosed by both the front and dac designers.
I don t understand what you mean by groundloops if each trace has his close return individual path , but with a single groundplane you may have mixed current between the 3 signals... is it that important...maybe yes maybe not. Anyway, if you talk about ground loop with the digital in of the dac chip... there are options.
Frankly I never get a good tonal balance with Nichicon caps, not in the whole digital side nore the analog side, nore ceramic btw... I m certainly biased.... or my hifi,...
The shorter the better as you know, for mono dual boards some options exists like Iancanada but not only.
I prefer to hace an exact impedance between boards ufl cables permitt...hoping 50 ohms is yhe impedance chosed by both the front and dac designers.
I don t understand what you mean by groundloops if each trace has his close return individual path , but with a single groundplane you may have mixed current between the 3 signals... is it that important...maybe yes maybe not. Anyway, if you talk about ground loop with the digital in of the dac chip... there are options.
Frankly I never get a good tonal balance with Nichicon caps, not in the whole digital side nore the analog side, nore ceramic btw... I m certainly biased.... or my hifi,...
Thanks Abraxalito.
I'll mod this in steps. Removing the VHC and the resonator is clear. Removing all filters might remove the magic I would like to keep. Unfortunately I think the pcb is of low quality and won't allow not too much soldering attempts. I'll listen to it after removing the reclocker and then try to remove the filters one by one, not to loose the magic.
Would it be possible to be more specific about your mods, or can I find it in your thread? Maybe as well in one of the later posts here. Havent had time to look into, as I have spent the day outside.
Cheers, Ernst
Hi Sandia, it was time for me to start a thread
I guess, I'll order a second pcb as well in order to have a normal to compare. Thank you for your input. I think at first I keep the bass boost, as my open baffles suffer from a slightly low bass from 100 to 300 Hz due to baffle step. I'll gonna adress this in my next active crossover... and take it out later.
I second that the psu crucial with such a dac and for the Pi. In my pi there is no DCDC converter and it is powered by 4nV and 0,8nV ldo's and the dac is powered by that as well. It makes a difference.
Thanks for the attachments. I'll go through it.
Cheers, Ernst
I would like to shift voicing proposals to a later point in time and I guess, uf-l sockets are not needed here as the front end is given.
I do have a hot air desoldering gun and will give it a try later this week in order to remove the VHC. I am a bit scared not to harm the TDAs. Probaly some shielding to be applied.
Thank you for your ideas!
Cheers, Ernst
I do have a hot air desoldering gun and will give it a try later this week in order to remove the VHC. I am a bit scared not to harm the TDAs. Probaly some shielding to be applied.
Thank you for your ideas!
Cheers, Ernst
Sorry, it was totaly off topic to talk about uf-l inputs!
Btw, what is th epurpose of the VHC574 and the TCXO? Is it a upsampler/reclocker ? Buffer ?
What is the I/V resistor value ? If you have no Rohpoint with the good R, maybe some SUSSUMU very low noise if the resistors are smd ?
You could identify what caps are involved with the VHC574 and works on that to understand first the behavior of this chip on the sound.
You could bypass the TDA1387 caps with acrylic 0805 smd from Cornell... while the 20% precision makes it a little hard to acheive !
What are the model/brand of the main caps ? Does it worth to remove them and measure them in order to see if they have the same capacitance ? Worth a 2% precision lytics caps ? .... While the good sound is probably also a consequence of the stock caps mix ! I do agree with you on the need of two boards for benchmarking the mods !
I do think swapping the BP Nichicon Muse by an outputt traffo is a good idea ! And I will start with that to listen first the difference with the stock board
Btw, what is th epurpose of the VHC574 and the TCXO? Is it a upsampler/reclocker ? Buffer ?
What is the I/V resistor value ? If you have no Rohpoint with the good R, maybe some SUSSUMU very low noise if the resistors are smd ?
You could identify what caps are involved with the VHC574 and works on that to understand first the behavior of this chip on the sound.
You could bypass the TDA1387 caps with acrylic 0805 smd from Cornell... while the 20% precision makes it a little hard to acheive !
What are the model/brand of the main caps ? Does it worth to remove them and measure them in order to see if they have the same capacitance ? Worth a 2% precision lytics caps ? .... While the good sound is probably also a consequence of the stock caps mix ! I do agree with you on the need of two boards for benchmarking the mods !
I do think swapping the BP Nichicon Muse by an outputt traffo is a good idea ! And I will start with that to listen first the difference with the stock board
I haven't spoken about those particular mods on my thread, happy to expound on more details as you sound interested. It was kind of an experiment to see whether I could drive a transformer with a single-ended (rather than centre-tapped) primary and compensate for the DC. In that respect it succeeded but I still prefer to work with a centre tap wherever possible. In this instance though that calls for an extra digital inversion and separation of the data lines between the DACs into 'true' and 'inverted' phase, something that wouldn't be at all elegant on the AudioPhonics board.
The transformer, if you'd like to make it yourself consists of an EP17 core, 10k material. Primary 38 turns 0.41 diameter (two layers). Secondary 3800 turns, 0.05 diameter.
The output filter is made from a 200mH series element and 390pF, 340pF for the input and output capacitors respectively. The terminating impedance is 18kohm. This filter is designed to give NOS droop correction at 44.1kHz. I made the 200mH inductor on a P14 core with AL=160, 1118 turns of 0.07 diameter wire.
Details for the FET buffer can be found on my lingDAC thread.
There are two current sources providing the compensating current for the DC bias from the DACs into the primary. They sink ~140uA into each secondary winding.
Hello Abraxalito,
thanks therefor!
Ok, the difficulties I see, is making the transformers myself - never done. 3800 turns - don't think that would work - unfortunately. Do you know a source for that?
Fet buffer should be doable for me.
Inductor, probably the same as for the transformer, but I guess this should be easier to source.
Current sink: I haven't seen a current sink affecting the secondary, but why not. Is the reason for that curiosity or that one can use a weak CCS?
Cheers, Ernst
@ernesternest - if you don't mind paying for my time (the materials are insignificant in cost) I'll be happy to make the transformers and inductors for you.
Overall though I'd recommend not using the AudioPhonics DAC as the starting point as running balanced is preferable and rather tricky to do with that DAC as a base. Better use a PCB of my own design as that's natively balanced, with two arrays of 18 DACs on either phase. With balanced you'll not need any compensating current sources. The gerbers for that PCB I'll be putting up soon (hopefully within the next week, once I've built a prototype up and tested it).
<afterthought> It just occurred to me that if you had two AudioPhonics DACs then you could make a balanced DAC based on two, feed one true digital data and the other one inverted. Just an idea!
Overall though I'd recommend not using the AudioPhonics DAC as the starting point as running balanced is preferable and rather tricky to do with that DAC as a base. Better use a PCB of my own design as that's natively balanced, with two arrays of 18 DACs on either phase. With balanced you'll not need any compensating current sources. The gerbers for that PCB I'll be putting up soon (hopefully within the next week, once I've built a prototype up and tested it).
<afterthought> It just occurred to me that if you had two AudioPhonics DACs then you could make a balanced DAC based on two, feed one true digital data and the other one inverted. Just an idea!
Last edited:
@abraxalito, this is a great offer! and I'll come back to that, later. I'll first do a few steps, that keep the boards simplicity (which I like) - so I just change a few parts (resistors and filters) and remove the VHC. Later then I try the trafo I have. Depending on the results I'll come back to you in order to ask you for some trafos and maybe TDAs. Thank you for your ideas!
Btw, currently I just have 1 piece of this DAC, but I think I should get a second one in order to have a normal.
Btw, currently I just have 1 piece of this DAC, but I think I should get a second one in order to have a normal.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.