@ happylistening,
Why would you add a board more for more complexity and less good result ? You have all you need in one and compact board ? Left & Right already ?
Sometimes it's better to stay on the design even if frustrating than to try to improve for worse.
It's a design as it is that is working perfectly fine and sounding good with the right dip 8 oap and a good I2S source like Wave I/O, JLSounds in I2S mode, SD card to I2S, all good made enough asynch which by principle are a sort of a recloking knowing 100% proof FIFO don't exist yet or avaliable easily but in theory.
We all wanrt to improve things to be a part of the story and to be a little a XXI century huan being, but again sad truth is it's not so easy to tweak hard a DAC (well sometimes it's so badly made than you have no choice like a stack of good ideas that sounds bad at the end but let you beleive you're in the hype while your ears cry the opposite, but again it's not he case with Miro1360's board that is simple, but well made board little pcb that borned well ! ).
You should have more margin by playing, imho & ime to focus on :
-quality of the digital frontend seen as a whole (from the clock till the outputt)
- Powersupply seen as a whole (from the energy source to the load)
- Passive parts locally to adapt to your tastes or avoid main mistakes (oscilation and so on) with PS association and/or adapt to the rest of you whole system as a final sounding (EQ, sort of)
Why would you add a board more for more complexity and less good result ? You have all you need in one and compact board ? Left & Right already ?
Sometimes it's better to stay on the design even if frustrating than to try to improve for worse.
It's a design as it is that is working perfectly fine and sounding good with the right dip 8 oap and a good I2S source like Wave I/O, JLSounds in I2S mode, SD card to I2S, all good made enough asynch which by principle are a sort of a recloking knowing 100% proof FIFO don't exist yet or avaliable easily but in theory.
We all wanrt to improve things to be a part of the story and to be a little a XXI century huan being, but again sad truth is it's not so easy to tweak hard a DAC (well sometimes it's so badly made than you have no choice like a stack of good ideas that sounds bad at the end but let you beleive you're in the hype while your ears cry the opposite, but again it's not he case with Miro1360's board that is simple, but well made board little pcb that borned well ! ).
You should have more margin by playing, imho & ime to focus on :
-quality of the digital frontend seen as a whole (from the clock till the outputt)
- Powersupply seen as a whole (from the energy source to the load)
- Passive parts locally to adapt to your tastes or avoid main mistakes (oscilation and so on) with PS association and/or adapt to the rest of you whole system as a final sounding (EQ, sort of)
Last edited:
Can u share your experience on any worsen consequence to separate channels by its own boards and its own power supply?
Thanks
Thanks
Well, it matters because opinion is just that, opinion.
As I'm working on a project that deviates from Miros original I'm curious to learn the basis for your statement;
or for that matter;
as perfection suggest something cannot be improved?
If my project is a blind alley it would be good to learn from you about where I may have taken a wrong turn?
As I'm working on a project that deviates from Miros original I'm curious to learn the basis for your statement;
or for that matter;
as perfection suggest something cannot be improved?
If my project is a blind alley it would be good to learn from you about where I may have taken a wrong turn?
Yes, but I'm not sure what you're wanting to acheive ? It will be wizer to say whay you want to double the boards ? Well maybe you did I don't remember it.
But from my opinion and experience and ears, even to reduce the I/V Z, it makes no sense... at least as I understand it the way you expressed it. My simple warm advice is you are complicating things. Is it not already separated ? Would we have to talk about PS fully separation just for it ? But for what improvment ?
short answer: my experience is it's a wasting time and at best you can worse the performance. It's just an opinion as far I don't hear myself what you acheive but that opinion is took from my experiences that are shared with a lot of sharing with my opinion after checking the experiments I want to share with you for all you confort 🙂
But from my opinion and experience and ears, even to reduce the I/V Z, it makes no sense... at least as I understand it the way you expressed it. My simple warm advice is you are complicating things. Is it not already separated ? Would we have to talk about PS fully separation just for it ? But for what improvment ?
short answer: my experience is it's a wasting time and at best you can worse the performance. It's just an opinion as far I don't hear myself what you acheive but that opinion is took from my experiences that are shared with a lot of sharing with my opinion after checking the experiments I want to share with you for all you confort 🙂
@NauticBoy,
Frankly, I don't care Sorry, sad truth, do what you want. My words were not for you but to avoid to some to waste time and money where it doesn't worth.
Sorry, no boom list to share for saving your time. I'm helped here I think more than you did, always nicely. I can understand my english maybe too basic and sometimes could give you the envy to figth a little, but at this point I would say, you should stop or I call Harry ! 😀
For what it worths in less compact and because I'm a nice guy still though I'm going to catch the EUVL syndrom if you continue: the design from a link you kindly share (thanks for the effort) is not working but for a TDA1541A, you DO NOT need imho the sk170 while of course I could be wrong but prove me the opposit and not your 'pinion here for the 1 mA of the Ad1862A. But your understanding is certainly greater than mine and your experience as well.
FWITW, as you're wisch is to share, better for you to coppy the design made here at Diya in the Painkiller thread that had the authoriation of Pedja and the advices of some members that seems to hav great opinions and experience.
Hey, just an advice, of course do what you want.
Frankly, I don't care Sorry, sad truth, do what you want. My words were not for you but to avoid to some to waste time and money where it doesn't worth.
Sorry, no boom list to share for saving your time. I'm helped here I think more than you did, always nicely. I can understand my english maybe too basic and sometimes could give you the envy to figth a little, but at this point I would say, you should stop or I call Harry ! 😀
For what it worths in less compact and because I'm a nice guy still though I'm going to catch the EUVL syndrom if you continue: the design from a link you kindly share (thanks for the effort) is not working but for a TDA1541A, you DO NOT need imho the sk170 while of course I could be wrong but prove me the opposit and not your 'pinion here for the 1 mA of the Ad1862A. But your understanding is certainly greater than mine and your experience as well.
FWITW, as you're wisch is to share, better for you to coppy the design made here at Diya in the Painkiller thread that had the authoriation of Pedja and the advices of some members that seems to hav great opinions and experience.
Hey, just an advice, of course do what you want.
Last edited:
Hey calm down, why so defensive? I was looking to learn, not have a willy-waving competition.
I'll assume you have expressed opinion and make my own exploration.
I'll assume you have expressed opinion and make my own exploration.
I have a slightly different view of this @diyiggy - this DAC is sounding so good as is, it is worth exploring the options there are with it, and sometimes that means going down the wrong way, reversing and retracing your steps. You discover little if you don't try things.
The separate boards have a lot of merit IMHO - they are very small form factor, they stack very neatly with the JLsounds board, and there is a logic in doing the L/R split and reclock in the JLsounds card, rather than afterwards. I also note that Miro designed the boards with convenient pads that can be cut if you have such a source that splits L&R and you don't need the shift register.
So rock on I say. I will test out reclocking, and hopefully it doesn't make any difference and we will be saved some work and cost. If it brings some improvement, then even better. If using a JLsounds board with nautiboys boards sounds better than an amanero or diyinhk usb card and the shift registers, well wouldn't it be great to know that?
We are all friends and all here to learn.
The separate boards have a lot of merit IMHO - they are very small form factor, they stack very neatly with the JLsounds board, and there is a logic in doing the L/R split and reclock in the JLsounds card, rather than afterwards. I also note that Miro designed the boards with convenient pads that can be cut if you have such a source that splits L&R and you don't need the shift register.
So rock on I say. I will test out reclocking, and hopefully it doesn't make any difference and we will be saved some work and cost. If it brings some improvement, then even better. If using a JLsounds board with nautiboys boards sounds better than an amanero or diyinhk usb card and the shift registers, well wouldn't it be great to know that?
We are all friends and all here to learn.
Simple... cause you begun ! Do you really believe I can not read between the line. So from my seats, calm down yourself first 😉. You exactly disearved what you have and it was made with nice words enough.
Well, all my opinions is shared here to avoid some to make errors I have the feeling to have made before. But of course I'm all for it as far it's about learning.
Reading from the post 1) untill now, you should be honest enough to know I should not be asked if it's about opinion or experience. I can still make the diference between a calm piece of s..t suited into a said calm sun ray ans someone honest enough in its communication. Astra Zen Caca... sort of ! 😀 Sorry I'm poet !
I'm ok for entertainment and experiment, but when you input to some they are maybe making errors and you explained it already before... well it's a little disapointing no ? And yesy do what you want, I'm ok to be on your ignore list, not the first time you made your perfid Albion with me !
Last edited:
I'll second that! I like to try out new things. In the 3 iterations of my AD1862-pcbs, the use of different regulators etc. I've learnt a lot which will be useful in other projects in the future.
But my primary goal isn't a good sounding dac I want to become more skilled in my hobby 😉
@diyiggy
I don't get all of your meaning but let me just say that I have no wish to create conflict and that was never my intention - I'm sorry if you've taken my questions as disrespectful. At the end of the day this is just a hobby - the world is already a troubled place and here we have an opportunity to create an oasis of mutual cooperation and sharing but we should also be confident enough raise and face challenges.
Life goes on, let's move on.
I don't get all of your meaning but let me just say that I have no wish to create conflict and that was never my intention - I'm sorry if you've taken my questions as disrespectful. At the end of the day this is just a hobby - the world is already a troubled place and here we have an opportunity to create an oasis of mutual cooperation and sharing but we should also be confident enough raise and face challenges.
Life goes on, let's move on.
I've not had a lot of time free this evening but I did track down the 'Painkiller' AD1862 DAC thread. I've not had time to read through it but there is a reference to the Diamond buffer in the final post so perhaps there's some more info earlier. I've not tracked down a thread specific to Pedja's diamond buffer so far.
AD1862 PCB layout
AD1862 PCB layout
Thanks Fran, that's the same link as the one Patrick posted in post #1661 - I was thinking/hoping that there might be a discussion/construction thread on it as well.
As far I can say,
I have both the Painkiller ad1862 DAC and Miro1360 dac.
Both are very good sounding dac. And the difference are more about layout and decoupling as well, than the final I/V tpology between both (at close I/V method : i.e. opa861)
My understanding is one of the great advantage of the discrete diamond buffer - P. Rogic gave the authorisation to use with the PainkillerDac cause the op gently asked him- VS an aop is the low impedance seen by the DAC chip.
Something like this as my understanding is basic. Perhaps EUVL that have already shared a lot can simply explain if he has the envy/patience, this advantage over an oap. there are maybe some over of course as thermal. Which can be also not an advantage, if I understood finely enough cause al the diamond discrete transistors should be thermally coupled... which the layout of the Painkiller Dac in in first discrete part iteration does well cause parts are very close. And its 4 layers avoid the problem of difficult layout and crossed trace (but as Patrick inputed a point to point i.e. 3D on a verroboard permits that more or less).
So my point is why reinvent the wheel.
P. Rogic writted that his discrete design showed will barely beat an ad844 or after the even better opa660 (there are also the today 860/861 sucessor I mistake between both with but he always prefered the non feedback design if you have to choose the ref (i.e here the 861, I'm sure of that for fried dozen of it in my several dacs and what I said several time to Vunce in his opa861 gerber he nicely shared). In some private talks I had with P. Rogic, a guy that has nothing to prove and greatly share aside is commercial business with some diy boards with have the pleasure to help to launch again with the help and good idea of some members here, he is perfectly happy with the opa861 till today as I/V thingy.
Why ? Cause you have the difficulty of sorting out the discrete part to have a chance to beat the transconductance oap that has the same goal and share also a same thermal behavior as in a standalone package.
The opa660 or 661 are obsolete and the 861 greatly is the legacy for the best.
This oaps however needs good layout and dip to soic adaptataor and vice-versa are not ideal and waste their great behavior... they need extra close decoupling, often with small parts that are not the ideal as little class II ceramic in the analog path. So quiet ground, good powersupply are mandatory for such a chips.
compact board and not splitted board can have some advantage as well with groundloop, and sometimes matching two channels with all splitted for a said better symetry can be worse with no sonic advantage but perhaps if needed (a full symetric design from A to Z ? Is the I2S symetric L/R splitted ? The only chip I know that is accepting natively a Word Data Left and Right separate dchannel, i.e. not multiplexed is the TDA1541 in silmutaneous mode protrocol. And John from ECEDESIGN proved btw that it is better inside a standalone chip that splitted in two different TDA1541A... cause the matching between chips itselfs !
We already talked about that several time in this thread about the i/v stage and its layouts.
That's why I made the shortlist to hapylistening about, ime & imho, about the safer path of improvment.
There is also the possibility to enhance the previous level to adapt to the i/v chip chosen (I proposed that already) : cause again the layput is not for nothing in the result and electrical needs. I learned that from others here. I also gave the idea while not took the time to draw it to have a stage for the buffer for the ones who would need more : difficult load after, headphones, ....
But again, as I said, nothing bad to experiment and learn from it. Just here we share our experience and opinions and it should not be underestimate.
Of course their a barrier language too. Which is great with Miro1360, is that is Kiss, plays fine with the right choice of parts (well I could argue on the passive part list but I'm a fanatic) : it is smart, compact, free, and the guy sure has some patience I haven't 😉
hope that helps, for people that want to make choices and choose their trade-offs.
I have both the Painkiller ad1862 DAC and Miro1360 dac.
Both are very good sounding dac. And the difference are more about layout and decoupling as well, than the final I/V tpology between both (at close I/V method : i.e. opa861)
My understanding is one of the great advantage of the discrete diamond buffer - P. Rogic gave the authorisation to use with the PainkillerDac cause the op gently asked him- VS an aop is the low impedance seen by the DAC chip.
Something like this as my understanding is basic. Perhaps EUVL that have already shared a lot can simply explain if he has the envy/patience, this advantage over an oap. there are maybe some over of course as thermal. Which can be also not an advantage, if I understood finely enough cause al the diamond discrete transistors should be thermally coupled... which the layout of the Painkiller Dac in in first discrete part iteration does well cause parts are very close. And its 4 layers avoid the problem of difficult layout and crossed trace (but as Patrick inputed a point to point i.e. 3D on a verroboard permits that more or less).
So my point is why reinvent the wheel.
P. Rogic writted that his discrete design showed will barely beat an ad844 or after the even better opa660 (there are also the today 860/861 sucessor I mistake between both with but he always prefered the non feedback design if you have to choose the ref (i.e here the 861, I'm sure of that for fried dozen of it in my several dacs and what I said several time to Vunce in his opa861 gerber he nicely shared). In some private talks I had with P. Rogic, a guy that has nothing to prove and greatly share aside is commercial business with some diy boards with have the pleasure to help to launch again with the help and good idea of some members here, he is perfectly happy with the opa861 till today as I/V thingy.
Why ? Cause you have the difficulty of sorting out the discrete part to have a chance to beat the transconductance oap that has the same goal and share also a same thermal behavior as in a standalone package.
The opa660 or 661 are obsolete and the 861 greatly is the legacy for the best.
This oaps however needs good layout and dip to soic adaptataor and vice-versa are not ideal and waste their great behavior... they need extra close decoupling, often with small parts that are not the ideal as little class II ceramic in the analog path. So quiet ground, good powersupply are mandatory for such a chips.
compact board and not splitted board can have some advantage as well with groundloop, and sometimes matching two channels with all splitted for a said better symetry can be worse with no sonic advantage but perhaps if needed (a full symetric design from A to Z ? Is the I2S symetric L/R splitted ? The only chip I know that is accepting natively a Word Data Left and Right separate dchannel, i.e. not multiplexed is the TDA1541 in silmutaneous mode protrocol. And John from ECEDESIGN proved btw that it is better inside a standalone chip that splitted in two different TDA1541A... cause the matching between chips itselfs !
We already talked about that several time in this thread about the i/v stage and its layouts.
That's why I made the shortlist to hapylistening about, ime & imho, about the safer path of improvment.
There is also the possibility to enhance the previous level to adapt to the i/v chip chosen (I proposed that already) : cause again the layput is not for nothing in the result and electrical needs. I learned that from others here. I also gave the idea while not took the time to draw it to have a stage for the buffer for the ones who would need more : difficult load after, headphones, ....
But again, as I said, nothing bad to experiment and learn from it. Just here we share our experience and opinions and it should not be underestimate.
Of course their a barrier language too. Which is great with Miro1360, is that is Kiss, plays fine with the right choice of parts (well I could argue on the passive part list but I'm a fanatic) : it is smart, compact, free, and the guy sure has some patience I haven't 😉
hope that helps, for people that want to make choices and choose their trade-offs.
Last edited:
So I tried reclocking tonight. The setup is like this:
SD card output (MCK, LRCK, BCK, DATA)> Potato 74G374 timed by MCK)>>DAC board (LRCK, BCK, DATA)
I found a slight improvement - tiny bit more resolution/detail and perhaps a shade smoother. I think it is worth trying reclock after the splitting L and R with this SD card player input. I know I could set up the JLsounds card to do it, but I'm trying to keep as many things constant as possible here.
It is only slight. so whether it is worth the effort is an individuals choice.
Fran
SD card output (MCK, LRCK, BCK, DATA)> Potato 74G374 timed by MCK)>>DAC board (LRCK, BCK, DATA)
I found a slight improvement - tiny bit more resolution/detail and perhaps a shade smoother. I think it is worth trying reclock after the splitting L and R with this SD card player input. I know I could set up the JLsounds card to do it, but I'm trying to keep as many things constant as possible here.
It is only slight. so whether it is worth the effort is an individuals choice.
Fran
@diyiggy
Can I ask what value I/V resistor you have in your 'Painkiller' DAC Diamond buffer - as you know the published article in the DIY Audio archive is for the TDA1541 and specifies 1K2 or 1K5. I'm assuming you have 'normal' output levels, circa 1.5-2V, from the DAC.
Cheers
Can I ask what value I/V resistor you have in your 'Painkiller' DAC Diamond buffer - as you know the published article in the DIY Audio archive is for the TDA1541 and specifies 1K2 or 1K5. I'm assuming you have 'normal' output levels, circa 1.5-2V, from the DAC.
Cheers