DAC AD1862: Almost THT, I2S input, NOS, R-2R

Over the years I have seen many designs of Discrete R2R DACs, with claims that they sound better.
I have seen next to no performance measurements though.

The fundamental design of a discrete R2R DAC is well explained here :
http://www.sonicillusions.co.uk/discrete_dac.htm

And most of the problems of such designs are well known, including the need to use 0.01% resistors, etc. :
https://www.diyaudio.com/community/...-magnitude-24-bit-384-khz.259488/post-3999275
https://www.diyaudio.com/community/...-magnitude-24-bit-384-khz.259488/post-4004298
https://www.diyaudio.com/community/...-magnitude-24-bit-384-khz.259488/post-4010487

So I very much look forward to you showing us something original and innovative, whenever you are ready to do so.


Cheers,
Patrick
Agree. The resistors is one of the big problems. Best option is still the old chips imho.
 
  • Like
Reactions: 1 user
I compared today JLsound Xmos and Amanero (Cetecj) on different tracks. It was Jesse Cook - Libre and Live Haggard - Awaking the Gods DVD.
Jlsound gives more "airly" sound with more open mid and high and Amanero gives drier and restrained sound. You can hear it clearly. That's why I told about studio and flat sound but with Jlsound transport the "more colors appears". I used same cables, I2S, etc. JLsound without power supply - I'm planning to buy (is there a reason?). I'am very happy with that Xmos :D
Photo of the audiophile debauchery:LOL:
 

Attachments

  • IMG20221128153105.jpg
    IMG20221128153105.jpg
    462.8 KB · Views: 138
  • Like
Reactions: 1 users
I use that in Pavouk AD1865 DAC. For PCM63 I wil choose LT1963. In principle, just avoid the LM317T or 7805 of unknown origin, buy only originals. The LM1117T proved to sound good, and there are others. Ebay, Aliexpress, but primarily avoid kit form, look only for PCB and buy quality parts.
 
  • Like
Reactions: 1 users
Over the years I have seen many designs of Discrete R2R DACs, with claims that they sound better.
I have seen next to no performance measurements though.

The fundamental design of a discrete R2R DAC is well explained here :
http://www.sonicillusions.co.uk/discrete_dac.htm

And most of the problems of such designs are well known, including the need to use 0.01% resistors, etc. :
https://www.diyaudio.com/community/...-magnitude-24-bit-384-khz.259488/post-3999275
https://www.diyaudio.com/community/...-magnitude-24-bit-384-khz.259488/post-4004298
https://www.diyaudio.com/community/...-magnitude-24-bit-384-khz.259488/post-4010487

So I very much look forward to you showing us something original and innovative, whenever you are ready to do so.

Not better than advanced designs, there are many very good :) Definitely not better than AD1862 or the like. This discrete one was a simple one just for a fun.
It can be improved in many ways. The only thing missing is time for that :cry:
Resistors can be 0.01% ... or existing DAC chip can be used instead of the most precise resistors and the most significant bits stays from R-2R (like 8-bit from resistors) (this is how D20400 is created).
Very accurate switches with low resistance can be used. It can handle the higher current better, like the TS5A4624 (only 1R on-state resistance). Plus a clean power supply and it will be even better separated from the digital interference.
SignMagnitude is interesting concept. One CPLD can do the job, but I like digital logic circuits and I created signMagnitude concept from that logic :ROFLMAO:
Maybe sometime in the future I'll do something better :)
 

Attachments

  • miro1360_R-2R_signMagn_test.jpg
    miro1360_R-2R_signMagn_test.jpg
    272.3 KB · Views: 99
  • Like
Reactions: 1 user