I see in many schematics for this part that the analog and digital grounds are seperated by a 10 ohm resistor.
I thought instead on my board that I would try to create (if possible) isolated traces on all analog and digital connections with every trace grounded at the filter cap ground.
Wouldn't that be a superior method....and could I eliminate the 10 ohm resistor?
THANKS.
--Mike
I thought instead on my board that I would try to create (if possible) isolated traces on all analog and digital connections with every trace grounded at the filter cap ground.
Wouldn't that be a superior method....and could I eliminate the 10 ohm resistor?
THANKS.
--Mike
I wouldn't recommend this due to the inductance of the individual traces. I would read the application notes and see if they provide any general guidelines.
Conventional practice with DACs has been to have a digital ground plane and an analog ground plane that meet at just one point under the dac, and this seems like pretty good practice for a 2902 as well.
In some cases one very solid ground plane has been used with good results as well.
Kevin
Conventional practice with DACs has been to have a digital ground plane and an analog ground plane that meet at just one point under the dac, and this seems like pretty good practice for a 2902 as well.
In some cases one very solid ground plane has been used with good results as well.
Kevin
So the resistor isn't needed?
I thought the star system of dedicated traces and a SINGLE grounding point was ideal in any system?
I need to learn more about this......a large (low inductance) ground plane vs the star system.
Perhaps I should do the analog grounds in a star and use a single ground plane for the digital with everything going to the chip ground?
I thought the star system of dedicated traces and a SINGLE grounding point was ideal in any system?
I need to learn more about this......a large (low inductance) ground plane vs the star system.
Perhaps I should do the analog grounds in a star and use a single ground plane for the digital with everything going to the chip ground?
jmar said:
the resistor is a potential danger
I put my thoughts on paper, long ago
http://www.tentlabs.com/Info/Articles/Supply_decoupling.pdf
It may help you in learning about the subject
Again, don't trust semiconductors manufacturers app notes, I once improved a demo board using my philiosophy: The jitter decreased by a fair amount and noise went down 30 dB........
cheers
Hello!
My experience when designing a circuit with the PCM2906, was that, i simply didnt connect analog in/out grounds with the digital grounds (USB gnd) at all! ... instead, the analog grounds are taken from the external equipment to which you are going to connect the interface to. The digital grounds however, are connected to the USB ground. This way, i found that it had 0 noise. I checked it on a pair of professional studio monitors, with absolutelly silent output when nothing was being played.
just my bit
My experience when designing a circuit with the PCM2906, was that, i simply didnt connect analog in/out grounds with the digital grounds (USB gnd) at all! ... instead, the analog grounds are taken from the external equipment to which you are going to connect the interface to. The digital grounds however, are connected to the USB ground. This way, i found that it had 0 noise. I checked it on a pair of professional studio monitors, with absolutelly silent output when nothing was being played.
just my bit
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.