CS4334 Noise

[nevelo]

Any ideas on how i can reduce noise/humm coming out of the CS4334 on a PCB?

I have the recommended bypass caps mounted close to it (10uF elec and 0.1uF non elec)...

Image shows one of the output channels coming from the DAC

An externally hosted image should be here but it was not working when we last tested it.

I dont however have the ferrite bead that it recommends in series with VCC. Will this make a massive difference?

I cant really change the layout of the PCB, but i can tack on caps and the like if anyone has had experience with this DAC and noticed a good way of keeping the output clean.

Thanks

Rob
http://www.robs-projects.com

 

[alfsch]

Hi,
the 4334 has about 95db snr, so : how big is the noise out, and is your v+ clean? what comes out, if dig data=zero in?
alf

 

[sonnya]

Any ideas on how i can reduce noise/humm coming out of the CS4334 on a PCB?

I have the recommended bypass caps mounted close to it (10uF elec and 0.1uF non elec)...

Image shows one of the output channels coming from the DAC

An externally hosted image should be here but it was not working when we last tested it.

I dont however have the ferrite bead that it recommends in series with VCC. Will this make a massive difference?

I cant really change the layout of the PCB, but i can tack on caps and the like if anyone has had experience with this DAC and noticed a good way of keeping the output clean.

Thanks

Rob
http://www.robs-projects.com

I have been looking on your schematic and the photo's of PCB.

Not to be hard on you. But i think you should work it over again. Both the PCB and the schematic.

To start with the easy part. Yes ferrit beads does change a lot. especialy with high speed micro's near by. as the AVR.

2. remove R4 and replace it with a short. Add a ferrit bead in series with R3 and VCC on U2. Add a NPO/COG 10nF after the ferrite bead to ground on U2 and U3.

3. Add a ground plane and move all (as many as possible) traces onto the top side. So that you have a unbroken ground plane.. Keep analog areas free from digital tracks.

4. route VCC/VSS in stars.

Maybe (i am nearly shure) this is the cure...

 

[nevelo]

Thanks sonnya for your detailed post on a possible solution to my problem.

I have a few questions though;

"route VCC/VSS in stars" could you explain what 'stars' means, im not familar with this terminology.

I think im going to try modifying the current PCB first with the ferrite bead and the caps, to see if this makes this particular board better.

The reason i did not include a ground plane was that this was my first 'proper' PCB, and i could not get my CAD tool to change the copper pour size (as requested by the PCB manufacturer), so i dropped it in favour of proving the design software and operation wise. I plan to get this sorted for my second revision.

Its quite curious that a 'home made' single sided PCB that i knocked up to test the STA013 and the CS4334 (different layout) worked very well and had hardly any noise. Even more surprising as there were wires soldered on and hanging off everywhere.

Thanks for your feedback guys.

Rob
http://www.robs-projects.com

 

[greyhorse]

I built Project 85 on the ESP website a while ago for fun, because I had a spare CS8414 lying around. At the time I had no PCB etching tools around, so I build the circuit P2P. The only ferrite beads on the board were those used on VD of the CS8414, and the TORX173 optical reciever. Power supply bypassing of the CS4334 consisted of a small value Panasonic FC electrolytic paralleled with a 0.1uF film cap. Both were DIP type caps, and the only SMD caps were the ceramics used for the CS8414. Regulation for the whole board came from a single LM317.

I wasn't expecting much, but the first thing I noticed was how quiet this DAC was. Of course quiet is a relative term, and unfortunately I have no numbers to give you, but my other DAC at the time was Jukka Tolonen's JT-DAC3, and subjectively this DAC was quieter. While both DACs exhibited no hum when the volume was turned fully clockwise, Project 85 seemed to have a much blacker background when playing music.

I would show you a picture if I could, but when I was grafting Elso's comparator input scheme to the CS8414, I mistakenly let one of the DC power supply leads gently brush an AC120V terminal. I was sorry to see it go, because for such a cheap project it sounded unusually coherent and musical. Much more so than my recent efforts with the TDA1543 for which I made a proper PCB for.

To summarize, I found the CS4334 unusually forgiving of layout and power supply quality. In your case, you have a lot of other digital ICs nearby which changes everything, so I'm not sure how much help my post will be. You wouldn't happen to have any wierd ground loops issues would you?