Simple DAC

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mcs

Well-known member
Joined
Jun 4, 2004
Messages
1,017
Location
Denmark
I need a DAC for testing some of my other projects. So I looked for a DAC chip that was simple to use, and had the features I need, and found the CS4340A. The features are 24bits, stereo and samplerates up to 192kHz BTW.

Here's the schematic: http://stiftsbogtrykkeriet.dk/~mcs/DAC1_sch.gif

It's mostly from the datasheet. I put in a mute relay instead of the transistors though. If anybody notices any errors please tell me about them before I make the PCB :wink:

Best regards,

Mikkel C. Simonsen
 
Instead of R7, I would have placed a small inductance to avoid digital noise to pollute analog ground.

Just my 2 cents ;)
 
speaking of noise I guess pcb layout can come into play here but this is a massive subject in its own right from what i've been reading/chin rubbin over so far...

What's ur thoughts for a regular dbl sided layout? Split A and D gnds (on both sides of the brd)? Then joined under the dac? what if your brd had 2, 4, 8 etc dac chips...
 
[quote author="daArry"]What's ur thoughts for a regular dbl sided layout? Split A and D gnds (on both sides of the brd)? Then joined under the dac?[/quote]
That's what I would do yes.

what if your brd had 2, 4, 8 etc dac chips...
Perhaps you could place the chips so that half the board was analog and the other half digital?

Best regards,

Mikkel C. Simonsen
 
Perhaps you could place the chips so that half the board was analog and the other half digital?

ye do that, but then if u join under each dac u'd create loops right - or just join in one spot i guess...or how about keeping them sep all the way bak to the psu?
 
[quote author="daArry"]or how about keeping them sep all the way bak to the psu?[/quote]
Maybe that's the best method.

Best regards,

Mikkel C. Simonsen
 
Here's my PCB drawing:

DAC1_PCB.gif


Actual size is only 7*4.2cm. I didn't do any groundplane on the top (I was too lazy), but I think it will still work...

Best regards,

Mikkel C. Simonsen
 
I use the standard sizes in my software. The thin ones are 0.32mm and the thick ones are 0.95mm (you do the mill convertion :wink:). I use the thin traces most of the time, but I like to use the thicker ones for the audio path, power and ground - when there's space for it.

Best regards,

Mikkel C. Simonsen
 
[quote author="slash14"]Instead of R7, I would have placed a small inductance to avoid digital noise to pollute analog ground.[/quote]
What size/type would you use? I placed R7 on the top layer, so I could use an SMD inductor instead of a resistor...

Best regards,

Mikkel C. Simonsen
 
MCS, I would use a small value ferrite bead for your coupling between Agnd and Dgnd. I was troubleshooting a video device proto for my company and came across a grounding setup like this one only there was just a copper bridge, no ferrite. the ground bounce was so bad that the analog power for the video circuit was actually causing severe noise on the video. I cut the track and placed a ferrite there and it works like a dream now. I don't remember what the value was though.. what are you using for your clk source? as i remember, you can also run the clock through ferrite too but the values have to be spec'd properly.

hope this helps..?

:guinness:
 
[quote author="Svart"]what are you using for your clk source?[/quote]
Since this is a test board, it could be either a "canned" crystal oscillator, the recovered clock from a CS841x or my PLL1707 board.

as i remember, you can also run the clock through ferrite too but the values have to be spec'd properly.
With a clock frequency that's not fixed, I guess that would be difficult.

hope this helps..?
Yes, thanks.

Best regards,

Mikkel C. Simonsen
 
Svart has been faster than I to answer but his answer is complete ;)

I used an inductance between AGND and DGND on the same kind of project but I can't remember the value, sorry.
 
Back
Top