Successful PCB grounding with mixed-signal chips - Part 1: Principles of current flow

Board-level designers often have concerns about the proper way to handle grounding for integrated circuits (ICs) which have separate analog and digital grounds. Should the two be completely separate and never touch? Should they connect at a single point with cuts in the ground plane to enforce this single point or "Mecca" ground? How can a Mecca ground be implemented when there are several ICs that call for analog and digital grounds?

This series of three articles comprises a basic tutorial on proper printed circuit board (PCB) grounding for mixed-signal designs. For most applications a simple method without cuts in the ground plane allows for successful PCB layouts with this kind of IC. Below we will learn about this method and that the basic principles used here can be extended to handle more complex and difficult applications.

We begin this Part 1 of the article with the basics: where the current flows. In Part 2 we will learn how to place components and route signal traces to minimize problems with crosstalk. In Part 3 we move on to consider power supply-currents and end by discussing how to extend what we have learned to circuits with multiple mixed-signal ICs.

And, yes, all digital circuits are analog circuits; they are a subset for which we assign meaning to only two states. The transistors and other components, as well as the currents and voltages within the circuit, still operate by the same physical principles as other analog circuits. They will induce return currents in the same way as any other circuit. 

Figure 1 illustrates the simplest of connections in a design: a direct connection from one chip to another. Taken as an ideal circuit in an ideal world,¹ the output impedance of IC1 would be zero and the input impedance of IC2 would be infinite. Therefore, there would be no current flowing. In the real world, however, current will flow from IC1 and into IC2, or the reverse. What happens to this current? Does it just fill up IC2 or IC1? That is a facetious rhetorical question.

Actually, there must be another connection between IC1 and IC2 to allow the current flowing into IC2 from IC1 to return to IC1 and vice-versa. This connection is usually ground and is often not indicated in a digital section of a schematic (Figure 1). It is at most implied by use of ground symbols as shown in Figure 2a. Figure 2B shows the full circuit for current flow.

Of course, the ICs themselves are not the sources of current. The power supply for the circuit is. To keep things simple, we assume a single power rail and think of the supply as a battery. To be complete, we bypass the supplies to ICs with capacitors. 

All DC currents ultimately start and end at the power source. Figure 3 shows the complete circuit with DC current flow when IC1 is sourcing the current indicated.

For high-frequency signals ("high" largely determined by the bypass capacitance and power-source impedance) the current starts and ends with the bypass capacitor. Figure 4 shows the high-frequency signal current flow.

It is important to remember that an output is not always the source of currents. For example, consider the case where an output from IC1 is connected to an input of IC2 which has a pullup resistor to V_DD. Figure 5 shows transient (high frequency) current flow for this situation with the current coming from C2 through the pullup in IC2 over to the low-side FET in IC1, which is on, and then through the ground lead of IC1 to the ground lead of C2. While IC1 is the "driving" device, sinking current at its output pin by shorting it to ground with a FET, the current source is from C2 through IC2.

If the output pin of IC1 in Figure 5 stays low for a long time, then the static current that will be drawn will come directly from the power source (Figure 6).