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

Is passive IV + tube buffer (no gain?) your preferred choice for the PCM56?
I have been using purely passive resistor IV (no trafo) with AD1862 and it sounds so good, the SQ trade offs of passive IV with this DAC are very low compared to any other DAC ive used... this explains why the trafo also sounded so good here compared to other applications.

Yes at that time I did not find anything that would work better. I wrongly called it 'buffer', it's a 12B4 with it's gain (a low mu tube, though. & low distortion, high drive and fast.. Very much an opamp, effectively.. :).)
Today I would do it with the OPA1656. So much so that I will do a redesign and try..
There is one catch: with the tube stage I have taken advantage of eliminating the built in bias current source in the PCM63(K)..I do remember that was a real step forward.. Previously tried a bunch of bypass options in that position, nothing was comparable to the no bias..
 
  • Like
Reactions: 1 user
@Vunce, I saw your post over in the Yarra thread and thought I would comment. I actually saw this thread a couple weeks ago and subscribed.

A couple months ago during the lockdown, I discovered another DAC that was based on the R-2R AD1865 chip. I found a guy in Malaysia who had adopted his larger design to the RPi form factor, and thought I'd try one. He had hand solder some a few years back and had one on hand. A week later it was in CT! :D

Some pictures of the RPi based board in my little streamer chassis. In the present configuration, I have it running on a single linear 5V PSU. When I fired it up I was pleasantly surprised by the overall sound. Especially, for a NOS chip. In roon it's very easy to group 2 DACs (endpoints) together and do A-B comparison of two different DAC, using the same source just by changing the pre input. Funny thing was, both me and my son couldn't tell that much difference in sonics between the two; my big multi-PSU system, and the little single power source AD1865! My ES9038Q2M system still had a slight edge though. Sometimes NOS surprises you!
 

Attachments

  • P1020216.jpg
    P1020216.jpg
    984.7 KB · Views: 519
  • P1020217.jpg
    P1020217.jpg
    885.7 KB · Views: 511
  • P1020220.jpg
    P1020220.jpg
    716.7 KB · Views: 483
  • P1020225.jpg
    P1020225.jpg
    866.4 KB · Views: 482
  • P1020227.jpg
    P1020227.jpg
    879.2 KB · Views: 473
well, if one buys anything like this from eBay/Ali it should be clear that there is NO quality control as such. and what kind of QC one can expect on second hand chips? that they are not dead at best. in fact I am pretty pleased I have 2 good ones now. and instead of QC hifi-szjxic offers "Gurantee exchange if it is fault." I have contacted them, let see.

I would be happy to see there are measurements of AD1862 from China linked to a particular supplier. There are none. so now I would say there is 50% chance to get good one from hifi-szjxic. who knows are there better suppliers?

to follow up the AD1862 from China story, I have to say that the communication with hifi-szjxic went great and the seller offered reimbursement on return and cover all shipping expenses. probably the best experience with eBay discussions for me so far.
still before the shipping the chip back, I decided to measure it again. doing this several times over and over I discovered that it needs kind of warm-up for about 30min to reach the expected distortion levels. so after this period the 3rd harmonic is still somehow highish, but below -100dB, and THD+N is about 0.00145, meeting DS specs. it also gets a little warmer(43oC) in comparison to other ones (38oC). hard to guess what's up, but the SQ is fine then. for the time being this behaviour is there on every cold start, but that's not critical I think.
so I decided to keep the chip and buy 2 more from hifi-szjxic.
 
Disabled Account
Joined 2019
Good. Reliable seller. At these low prices, shipping is almost as expensive that the goods. But if he proposed to get the monney back, I think it's ok. One can assume when it's cheap there is no quality control, however, one should not sell goods he doesn't know. The faulty chip is not only second hand, something happened : at the factory or when get off of a pcb if second hand. So for me whatever marked second hand, I will not go there as I'm unnable to check myself what I purchase.


All in one at the end if yu're happy and you get chips enough at the price you wanted it's ok.
 
Hi Redjr,
Maybe you can try Ian's Fifopi. He will have the latest version Q3 soon.
I compared my AD1865 with and without the Fifopi. The Fifopi improvement is enormous! Also Get a pair of good clocks.
Chi0001 kindly gave me the AD1862 PCB; I am very tempted to build one.
Of course getting good quality/authentic AD1862 takes time.



@Vunce, I saw your post over in the Yarra thread and thought I would comment. I actually saw this thread a couple weeks ago and subscribed.

A couple months ago during the lockdown, I discovered another DAC that was based on the R-2R AD1865 chip. I found a guy in Malaysia who had adopted his larger design to the RPi form factor, and thought I'd try one. He had hand solder some a few years back and had one on hand. A week later it was in CT! :D

Some pictures of the RPi based board in my little streamer chassis. In the present configuration, I have it running on a single linear 5V PSU. When I fired it up I was pleasantly surprised by the overall sound. Especially, for a NOS chip. In roon it's very easy to group 2 DACs (endpoints) together and do A-B comparison of two different DAC, using the same source just by changing the pre input. Funny thing was, both me and my son couldn't tell that much difference in sonics between the two; my big multi-PSU system, and the little single power source AD1865! My ES9038Q2M system still had a slight edge though. Sometimes NOS surprises you!
 
Hi Nano35,

I currrent have my FifoPi in my main ES9038Q2M system, along with multiple PSUs and the sound is articulate and really gorgeous. I'm still improving it with little tweaks here and there.

I bought the AD1865 after hearing some good things about it over in the roon community forum, the RPi form factor. At the moment, I'm using the AD1865 in a secondary, non-critical system in my shop and sounds 'good' enough for now. I may buy another FifoPi from Ian at some point, but really waiting for his version of a AK4499 chipset implementation - whenever that comes along. He's also offering some new complimentary boards for his 'DAC stack', the BridgePi and ConditionerPi. Those interest me as well.

Too many goodies and not enough time! :D
 
Member
Joined 2014
Paid Member
“Too many goodies and not enough time! :D

Ain’t that the truth, redjr!
Thanks for sharing your setup.
:cheers:

Has anyone used the OPA860/861 Operational Transconductance Amplifiers with the AD1862 for I/V and buffer duty?
The OPA860 has the buffer integrated, so I’m thinking this could be simple solution with a few modifications to Miro’s board.

https://www.ti.com/lit/ds/symlink/o...ll-mousermode-df-pf-null-wwe&ts=1594306753798

https://www.ti.com/lit/ds/symlink/o...ll-mousermode-df-pf-null-wwe&ts=1594306798961
 
Last edited:
Member
Joined 2014
Paid Member
Thanks Miro for running a Sim with the OPA861:)
The +/-5v supply isn’t a problem.

For an even simpler option, the OPA860 has a buffer already.
If you have time, can you run a Sim with this OTA also?

Thanks!!
 

Attachments

  • D05F3E76-7F4C-4134-A14D-63CF40B06C2A.jpeg
    D05F3E76-7F4C-4134-A14D-63CF40B06C2A.jpeg
    62.5 KB · Views: 386
Disabled Account
Joined 2019
stay with the opa861 for a better result, you can divide by two the 1k5 I/V resistor if not needed 2V outputt. R7 is 150R, not 1500? A // styren with tin foil from LCR till 1nF is advised if you listene to NOS material lile RedBooks. The opa861 without buffer can drive a pre or an amp with just serie resitor.


Is it not better toputt a resistor between the opa861 outputt and the LME ?


You can also use a 861 reversed instead the LME for the buffer. the 861 has no feed back s quiet simple.


You can share the same 5V but local decoupling is important at the feet of the chip.
 
Last edited:
OPA860 with integrated buffer has similiar THD (about 0.05%), but the noise floor is worse (if the load is 1k5), THD is way worse for higher load (0.8% for 32R load, in LME49600 it still be nice 0.05%)
 

Attachments

  • OPA860-miro1360-THD.jpg
    OPA860-miro1360-THD.jpg
    251.2 KB · Views: 329
Last edited: