I must be crazy to build my DAC this way

Status
Not open for further replies.
Jump to Latest
Peter Daniel

Peter Daniel

Member
Joined 2002
audiosector.com
Paid Member
#1
I finally decided to build TDA1541S2 based DAC. As usually I'm looking for best ways to implement the mechanical aspects of the design, so this is what I think could be a proper method to dress up S2 chip😉
 

Attachments

  • d1.jpg
    d1.jpg
    57.7 KB · Views: 3,102
#2
Initially I was considering PCB, but somehow didn't like the restrictions dictated by 2 dimentional plane. So last evening I took a piece of brass and played with it a bit. I mounted the chip on top of 1/2 square rod and attached to it on top a piece of PCB forming ground plane. The rest is pretty self explanatory. I wonder myself if this is a good move or not?😉

Any comments on 560p Wima? I didn't know if this a good choice, but that's what seemed to fit there.
 

Attachments

  • d2.jpg
    d2.jpg
    56.5 KB · Views: 2,915
#6
I was thinking about splitting grounds, but checked the PCB board on my Marantz CD 94 and both grounds were connected at the chip (same ground plane). Doing it this way definitely simplifies things, but I'm not sure how this affects sound. If they are separated at the chip, were should they meet then?
 
#7
Pedja said:
Hi Peter,

Yes, it is OK I'd say (excellent actually), but try to divide digital and analog grounds before you joint them. Decoupling caps (pins 26 and 28) belong to the digital, those 14 caps to the analog part.

Pedja
Click to expand...

So 5v supplies are digital and -15 is analog? I could use a center ground strip to route grounds from pins 26 and 28 to the common point on the other side.
 
#11
Nice work, Peter!

I've found that these DACs benefit enormously from good PSU decoupling and screening and it isn't possible to optimise this using a PCB alone.

As regards connecting analogue and digital grounds: ignore the method used by Marantz, on my CD80 they're daisy-chained together by a thin track:xeye:. They should meet at a low impedance point like a big decoupling capacitor.
 
#12
Peter Daniel said:
Pedja,
I just checked your site and the info is good as usually (on TDA1541 Dac)
Click to expand...
Thanks.

Richard C said:
I've found that these DACs benefit enormously from good PSU decoupling and screening and it isn't possible to optimise this using a PCB alone.
Click to expand...
No doubt, it benefits. I screened the chips by the copper sheets later but still have to upload newer pics. And yes, the way the Peter is doing this might be better than usage of PCB, regardless of how careful it is designed. Though it is certainly not the easier way.

As regards connecting analogue and digital grounds: ignore the method used by Marantz, on my CD80 they're daisy-chained together by a thin track:xeye:. They should meet at a low impedance point like a big decoupling capacitor.
Click to expand...
Two pages of extremely useful text on the topic:
http://www.sigcon.com/Pubs/edn/adcgrounding.htm
http://www.sigcon.com/Pubs/edn/multipleadc.htm

Pedja
 
#15
Pretty as always. Have you also consider to mount the receiver chip in the same manner, but face it inverted and mount it underneath the DAC chip. This way, the 3 I2S pins (FSYNC, SCK, SDATA) will be P2P also. This will solve your ground plane problem too, by sandwiching a ground plane (say digital ground under the chips) between the two chips. Analogue ground will be on the top of both chips.
 
#18
Peter,
You are crazy 😀

I would suggest not using any cap between pin 16 & 17, but instead implement the DEM reclocking circuit.
You'll be suprised how much it does.
Thread about DEM is somewhere in the archives.

BTW TDA1543 board is finished, just need to install it in a cd player. Probably start a thread tonight.
 
Status
Not open for further replies.
Top Bottom