Crosstalk is one of those more complex signal integrity problems, and it is sometimes misunderstood. All crosstalk arises due to unintended coupling through the electromagnetic field, but there are different types of crosstalk to consider in any PCB design. Crosstalk can never be eliminated, but it can be reduced to the point where it does not affect signal quality at the receiver.
For digital boards operating at high speeds and high data rates, crosstalk can be hugely problematic and difficult to pinpoint, simply because there can be so many nets involved in producing crosstalk simultaneously. Analog boards have fewer signals and crosstalk can be identified easily, but it can be difficult to eliminate down to an acceptable level as analog components and traces react so strongly to external noise. Let’s look at the different types of crosstalk in your PCB and some steps you can take to eliminate them.
All crosstalk occurs between a victim line (the net or transmission line where a crosstalk signal is induced) and one or more aggressor lines (the net or transmission line that induces the crosstalk signal in the victim). If you read through different tutorials and websites, you’ll see that there are several terms used to describe types of crosstalk:
In reality, the only types of crosstalk are forward and backward crosstalk. The other terms in the above list refer to either where the crosstalk is measured (NEXT and FEXT), how crosstalk is quantified (power sum measurements), or the crosstalk source (AXT). All crosstalk is induced through two mechanisms: inductive and capacitive coupling. Let’s look at how these two mechanisms give rise to crosstalk in a real PCB.
The circuit diagram below shows how circuit theory is used to model different types of crosstalk. In this diagram, there is some parasitic capacitance between the two traces, which exists due to broadside coupling between the traces. Because each trace is a loop of conductor, each trace acts like an inductor and has some parasitic inductance. The two parallel inductors have some mutual inductance, which defines the strength of inductive coupling between the two traces.
This circuit diagram shows the classical circuit model describing different types of crosstalk in a PCB.
The capacitive coupling is not just meant to show the equivalent capacitor created by the edges of the traces, although this does contribute to the mutual capacitance. The native capacitance of each trace and the broadside capacitance combine to give the total mutual capacitance; they are all in series and coupled back to the ground plane.
When crosstalk occurs and is observed as a snapshot in the time domain, it is difficult to differentiate capacitive and inductive contributions to crosstalk. This is why we have FEXT and NEXT measurements; they basically allow you to quantify forward and backward crosstalk. Let’s look at how these coupling mechanisms produce the two types of crosstalk in a PCB.
When an analog signal oscillates, or as a digital signal rises along the length of a trace, the changing current along the signal generates an oscillating magnetic field. The oscillating magnetic field then generates a changing magnetic flux within the cross section of the victim trace. Thanks to Faraday’s law, this induces a back EMF in the victim trace. In other words, the voltage signal created in the victim points opposite to the signal in the aggressor. Therefore, the Inductive coupling generates backwards crosstalk on the rising signal edge, and it generates forward crosstalk on the falling signal edge.
What about capacitive coupling? This is a bit more complicated in that you need to consider the distributed nature of the line. When the voltage on the aggressor line changes, a displacement current is induced in the victim trace. In this situation, the displacement current is only induced in the victim within a certain region (called the coupling length, more on that below). The displacement current then moves away from the point where the signal is induced along the line, i.e., towards the transmit and received ends.
Note that, for digital signals, the two lines act like a capacitor that needs to charge when there is a potential difference between the aggressor and victim lines. However, This charging effect occurs on the order of microseconds, so any stream of digital aggressor signals faster than ~1 MHz will induce glitches that appear on the victim with similar timescale.
Equations for describing the voltage and current of a crosstalk signal due to inductive and capacitive coupling. These values are defined in terms of the mutual inductance and capacitance, respectively.
Note that both types of crosstalk only occur within a certain region that exists between the coupled aggressor and victim. Remember that a wave travelling on a transmission line is extended out over some distance, which depends on the dielectric constant of the substrate. Because the aggressor signal only changes within a certain portion of the aggressor trace, the crosstalk signal will only be induced in a small region in the victim trace. This is called the coupling region, and it is the region over which the signal is changing (see the above equations). The image below shows the coupling region for the rising edge of a digital signal on a pair of coupled transmission lines.
In this example with a digital signal, the coupling region only exists in the portion where the signal edge is rising. The coupling region on the two lines is outlined in blue.
Note that forward and backward crosstalk can exist on a victim line simultaneously, which is why we care more about measuring NEXT and FEXT than about tracking forward or backward crosstalk. Once a summed crosstalk signal reaches the transmit or receive ends of the line, it can be measured as either NEXT or FEXT, respectively. With the right high speed layout practices and stackup design techniques, you can reduce crosstalk between different portions of your design. Here, we’ve only addressed single-ended crosstalk, but the same effects occur in differential pairs. I’ll get into more depth on crosstalk in analog systems and differential crosstalk in the future.
At NWES, we regularly work with high speed PCB designs and sensitive analog designs that are sensitive to different types of crosstalk. We know how to create a high quality, fully manufacturable PCB layout for your system. We're here to help electronics companies design modern PCBs and create cutting-edge technology. We've also partnered directly with EDA companies and advanced PCB manufacturers, and we'll make sure your next layout is fully manufacturable at scale. Contact NWES for a consultation.