It can be much simpler to use only one ground plane. Much less thought involved and less chance to mess things up. For your use it might come down to how many bits the A/D is- how much ground noise you can tolerate getting into the analog circuitry. There will be "DC" noise, or offset, as the digital circuitry draws more or less current. This is fairly easy to estimate.
For example, say you have a single layer devoted to ground and that it is 1/2 oz copper. The DC resistance is 1 milliohm per square. Say the digital current is 100 mA and it is all going to the D/A and flowing through the same section of the ground plane that the analog signal is. There could be a 100 uV offset seen by the analog circuit due to the voltage drop in the last 1/2 inch approaching the IC. For high frequency digital current we need to use another method that is even more of a guesstimate. Remember that the analog circuitry may or may not be very sensitive to high frequency noise. The power plane will appear as an exponentially expanding transmission line- above about 100 MHz the decoupling capacitors are out of the picture. The digital current will flow through the "line". Say the P-P digital current is 100 mA and the line has an effective impedance of 1/2 ohm for the digital currents. There will be 50 mV of noise on the digital and analog power pins. The voltage drop along the ground plane is what I'm not 100% sure on. Perhaps other's can clarify this. If this were coaxial cable there would be no voltage drop due to the inductance of the internal surface of the line. With a PCB we don't have complete magnetic coupling be it is close. So, I think we can ignore the voltage drop due to surface inductance. We do have to consider the resistance, however. At 100 MH! z it will be about 3 milliohms per square. If so, then the 100 mA of digital current will produce a voltage drop of 300 uV that gets directly into the analog inputs. We also want to consider the inductance of vias. I usually model them as 50 ohm lines of the appropriate length. Or simplify to just an inductor. Figure 500 pH for a 0.062" PCB via. Less when going from, for example, layer 3 to layer 1 on your 8 layer board. Now it's about 200 pH. Dave Cuthbert Micron Technology This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. Visit our web site at: http://www.ieee-pses.org/ To cancel your subscription, send mail to: [email protected] with the single line: unsubscribe emc-pstc For help, send mail to the list administrators: Ron Pickard: [email protected] Dave Heald: [email protected] For policy questions, send mail to: Richard Nute: [email protected] Jim Bacher: [email protected] All emc-pstc postings are archived and searchable on the web at: http://www.ieeecommunities.org/emc-pstc
