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

Hi, I haven't a problem with my two layers smd version where all the digital came from miro board and all the rest analog and decoupling I redesigned. You can load the gerbers I gave here in the thread (not the miro's link in his blog spot, but a further post, I don't have a link sorry, but someone asked me here so perhaps he can give it). It has two smd op amps stage. Load the gerbers it works and you can redisign from there.

I made a 4 layers without noise as well (much quieter than a 2 two layers for obvious reasons but it is tricky) but not public domain (for obvious reasons as well). No digital noise either.

I have a brief look at the picture you gave and I see many crosstalks and coupling everywhere and "isolation of AGND DGND a little odd w/o the "moat" to be so effective everywhere and even usefull sometimes.

But whatever a 2 or more layers, you should not hear with your ears "noises". The layout of the shift registers can be tricky too.
 
Last edited:
I tried changing the GND connection in many ways, but nothing improved the problem I asked.
If it was a GND crosstalk problem, at least when I tried some of the connection methods, this "digital noise" should have improved. But it didn't!

Sorry for my poor English.

This "digital noise" is a bit like distortion in the high frequency part. At lower volumes, it makes the high frequencies very sharp and very obvious.
 
Last edited:
Is there a way to determine how far in class a an opamp is tuned?
I understand your suggestions of a discrete opamp.
I see it more as an experiment from the "standard configuration". So don't want to deviate to much from it to keep it simple.

You can see in somme datasheet, some op amps having big quiescent current, typical to be near from Class A for some.

Frankly, adding such a biasing will more harm than help on the ready made pcbs here and adds stray capacitance (and it is a very fast op amp that needs quietness) and strange feedback in an transimpedance op amp scheme, There are more to hunt with the LM6171 first.
 
Friends. Sorry, I'm not sure if I made a stupid mistake.

Just now, I improved the "digital noise" problem I mentioned by adjusting the AD1862 "OPTIONAL TRIM ADJUSTMENT"! !

I always thought that this adjuster would not bring such obvious "low volume distortion", but the facts tell me that it does! It can significantly improve the "low volume digital noise" I mentioned, at least it is completely gone now. The sound of NOS AD1862 is very real, I like it!

I am very curious, how do you adjust this potentiometer? Use a distortion meter or an oscilloscope?

22.png



11.jpg
 
It migth or not according how it is made, you have to look at the shematic when present. It can be also something like VFA that has partial CFA topology according what I read, but I dunno. My ears are my guide, measurements come only after the design respect the rules, then I refine. 5 iterations of my ad1862 pcb to come there... Diminishing return, but it is our jobby, isn'it.

Btw, that is the input stage we focus more here in transimpedance schemes and noise suceptibility.
 
Last edited: