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

[woodturner-fran]

This is how to break and wire-jumper in the v2.0 PCB

Thanks Miro. So if someone still wants to have a go I'll have a few boards for free within the EU. PM me if you are interested.

 

[sworder84]

I got a chip but he's dead also.
I don't know it is original or fake. I only saw a few photos on internet.

 

[EUVL]

Through hole Version 2.
DC stable to +/-3mV, which is equivalent to +/-1µA with 3k Riv.
And that is with a total bias of 35mA.

Beta tester already lined up.
Publication only follows thereafter.




Patrick

 

[Vunce]

Very cool Patrick

 

[baggerbole]

Do you mean this one? Diyiggy

 

[jpk73]

Thanks
Yep I remember your post about that. When you say PC it is a big one or laptop through usb ?
Wonder how such cips compare if the PC is a reclocked isolated Rpi aka Fifopi or Allo Kali...Or Kusy board the same?
Cool work you did

Thanks! It was a PC, but the DAC was extremely well isolated and reclocked.

You can do that from flip-flops, but you must generate the LRCK directly from the master clock (like the HC4040 does) ... it divide the MCLK (or presious crystal clock) directly and generates the proper LRCK.
If you are referring the old (jittered) LRCK for generating new LRCK through flip-flop, you will get jittered LRCK because one flip-flop works only like a valve controlled by the MCLK (if jitter in old LRCK is larger than one MCLK cycle, it will be transfered to the new LRCK --- that is why you need the clean generation based on division instead of referencing the old LRCK).

Wow, larger jitter than MCLK, OK that is a case I never thought of! Does this really happen...?

 

[Vunce]

Do you mean this one? Diyiggy View attachment 1059588

I believe diyiggy is referring to Miro’s first pcb version.
https://www.diyaudio.com/community/attachments/image005-jpg.841454/

 

[diyiggy]

Yep, I find the decoupling of this one the best of all. The output of the chip to the oap is also straigth forward and the decoupling cap 47 uF filter is near from the pin chip which is (theorically speaking) important. On the oap both 0.1 uF are close the power pins (permitt a hack a la Vunce with smd cap or mine with a little wire to the gnd via with 0805 CDE FCA 0.1/16V ). And this 2.5 mm pitch for the oap lythics and the gnd pins of the lythic and the mkp very close .
But for me ufl pads and close I2S source evem more important so I would like to test the jlsound stacked or direct ufl pcm option with 2" uf-l wire. Just curiosity as I will not do a JPLCB order to fullfill this curiosity.

 

[Vunce]

Vunce Hack

 

[diyiggy]

Through hole Version 2.
DC stable to +/-3mV, which is equivalent to +/-1µA with 3k Riv.
And that is with a total bias of 35mA.

Beta tester already lined up.
Publication only follows thereafter.




Patrick

A cool tht for papies . Zetex bjt ?. Damit... 22 uF Nich FW ? Better something else cap...of course my pref only.

 

[diyiggy]

Vunce Hack

Well done...C29 mkp is indeed a little too far according some oaps... but the mkp vias could still serve with a higher 1 uF mkp or a SAL lythic or as well a solid modern tantalum... ears and the system one has having being the final judge...no universal solution as usual. But definitly a good idea than ceramic for the HF decoupling very close of the pin with 0603 casing

 

[EUVL]

Before I decide what cap to use where, I first consider its function.
Horses for courses.

Cheers,
Patrick

 

[diyiggy]

Me too, it doesn't arm though to avoid some caps... but well had you not chosen as wel the bipolar Nich ES too here...in the past iirc . C mon if a Fw drop it to the bin...I have nothing to sell.

 

[EUVL]

And do you know that tantulum caps fail by short circuit ?
To minimise the chance of failure, you need to use 5x the voltage rating.
I have managed to avoid them for the last 15 years.


Cheers,
Patrick

 

[EUVL]

"it doesn't arm though to avoid some caps..."

I have used Nichicon FW in other applications with very good results.
I don't have problems with them.

But of course everyone is free to pick his own preference.


Cheers,
Patrick

 

[diyiggy]

Tantalum are not today those we used 15 or 30 years ago. They are also solid tantalum that not fire easilier. Reverse voltage still a problem though as lythics btw. I do no like too much them for analog, after all still a tantalum kathode. But each hifi syst is different and they can help sometimes.
Fca caps as SAL are not tantalums. In my syst Fw has a soft veil that soft everything and has not real transparency...Ymmv as you said. But the Nich ES for direct DC coupling I do not give some credit to Nichicon as far the sound is the first factor. But I like their gold letters on casings...
I like though the acrylic pen cap on your board though.

 

[EUVL]

Another 2 finished with very repeatable results.


Patrick

 

[EUVL]

Before anyone rushes to get matched complementary JFET pairs, please read this carefully :
https://www.diyaudio.com/community/...lsk170-jfet-pairs-grade-b.321649/post-7035859
Guess how I found out.

So my advice -- don't buy any LSK170/LSJ74 matched pairs unless you know for sure they are matched at the same voltage as your intended application.
Simply "matched" can mean a lot of things.




Patrick

 

[Vavilen]

Hi guys! How to find the connector on Mouser? What are the keywords?

 

[Vunce]

Here you go:
Molex 22-23-2031