The Best DAC is no DAC

[koldby]

Yes really impressive what the LVC1G79 can do...

[coresta]

Originally Posted by laserscrape

Just seeing if anyone has any thoughts on this.

Attached are rough sketches of the 2 possible ways im going to try to build this power noDAC.
Chances are it will have unusable performance but it will be fun and costs so little.

In both examples the signals are "balanced" before the flip flops by an inverter and then inverting flip flops are used, first this keeps the signal path slightly simpler after the signals have been reclocked since non-inverting flip flops have a non-inverting buffer on the output (i.e 2 more inverters).
Additionally the single output flip flops dont have CLR and PRE functions, so their internal circuits are a lot simpler.

hmm, how do you get balanced signal from paralleled circuits ? Anyway, thats true , a 6 // 74LS02 buffer from a 44,1kHz clock reduces jitter ...

[Zoran]

Hi Koldby
what part You use for Your no-dac from page 214 this topic?
You just wrote
74LVC1G?
Is this 74LVC1G74
thanks

[koldby]

Originally Posted by Zoran

Hi Koldby
what part You use for Your no-dac from page 214 this topic?
You just wrote
74LVC1G?
Is this 74LVC1G74
thanks

Ups. Seems I missed the last 74 ;-). But it was a 74AUC1G74 as it is a FF...

[mcspack]

Originally Posted by lasercut

No, single LVC1G79s.
It is 4 layer PCB from JLC, cost around €5, it's crazy.

Isolator is IL715, I will add it today.

How did it sound post isolator? Can more of these boards be produced if it sounds as expected?

[Zoran]

Originally Posted by koldby

Ups. Seems I missed the last 74 ;-). But it was a 74AUC1G74 as it is a FF...

thanks

[lasercut]

Originally Posted by mcspack

How did it sound post isolator? Can more of these boards be produced if it sounds as expected?

Worse, the problem of jitter>isolation in this case.
In the end there was an underlying metallic-ness to the sound that could never be helped, probably related to poor high speed PCB design.
Also noticed some subtle artifacts in one channel , might have been a damaged/faulty flip-flop but couldnt be bothered building another or trying to find the fault so it was abandoned after that.
Managing volume was a problem, minimum supply of 1.65V (3.3V differential output) still requires heavy digital attenuation for my headphones and although there wasn't a huge difference I felt performance suffered going below 3.3V supply.
It might make more sense to scale it up for sensitive speakers instead of fighting the volume issue, you can get up to 11V differential output or 13V if you push to absolute max.

[Zoran]

H
You said that single chip 4 X isolator IL715 used for digital isolation.
Maybe to use a single isolator for each digital line.
Like IL710.
(enable option could be of use too...)
.
I had good sonic results with ADuM1100 single isolators. 4 times for I2S bus.
Yes the jitter performance is probably slight worse, but in the other hand amount of digital noise is probably reduced?
.