PCB Manufacturer in China: Explain several misunderstandings in differential signals

Differential transmission is a kind of signal transmission technology. It is different from the traditional one signal line and one ground line. Differential transmission transmits signals on these two lines. The two signals have the same amplitude and opposite phase. The signals transmitted on these two lines are differential signals. The signal receiving end compares the difference between the two voltages to determine the logic state sent by the transmitting end. On the board, the differential traces must be two lines of equal length, equal width, close proximity, and on the same level.

Strictly speaking, all voltage signals are differential because one voltage can only be relative to the other. In some systems, "system ground" is used as a voltage reference point. When 'ground' is used as a voltage measurement reference, this signal planning is called single-ended. We use this term because the signal is represented by the voltage on a single conductor. On the other hand, a differential signal acts on the two conductors. The signal value is the voltage difference between the two conductors. Although not necessary, the average of these two voltages will always be consistent.

It is conceivable that an equal voltage that is simultaneously added to the two conductors, the so-called common mode signal, has no effect on a differential amplification system, that is, although the input effective signal amplitude of a differential amplifier is only It takes a few millivolts, but it can be indifferent to a common-mode signal up to a few volts. This indicator is called the common-mode rejection ratio (CMRR) of the differential amplifier. The general op amp can reach more than 90db, and the high-precision op amp even reaches 120db. Since the interfering signal is generally in the form of a common mode signal, the application of the differential signal greatly increases the signal to noise ratio of the amplifier system.

Differential signal (DifferenTIal Signal) is widely used in high-speed circuit design. The most critical signal in the circuit is often designed with differential structure. What makes it so popular? How can we guarantee good performance in PCB design? With these two questions, we will discuss the next part. What is a differential signal? In layman's terms, the driver sends two equal-valued, inverted signals. The receiver compares the difference between the two voltages to determine whether the logic state is "0" or "1". The pair of traces carrying the differential signals is called a differential trace.

Compared with ordinary single-ended signal traces, the most obvious advantages of differential signals are reflected in the following three aspects:

1. Strong anti-interference ability, because the coupling between the two differential traces is very good. When there is noise interference from the outside, it is almost simultaneously coupled to two lines, and the receiving end only cares about the difference between the two signals. Therefore, the common mode noise of the outside world can be completely offset.

2. It can effectively suppress EMI. For the same reason, because the polarities of the two signals are opposite, the electromagnetic fields radiated by them can cancel each other. The closer the coupling is, the less electromagnetic energy is discharged to the outside.

3. Timing positioning is accurate. Since the switching change of the differential signal is located at the intersection of the two signals, unlike the ordinary single-ended signal, which depends on the high and low threshold voltages, it is less affected by the process and temperature, which can reduce the timing error. It is also more suitable for circuits with low amplitude signals. The currently popular LVDS (low voltage differenTIal signaling) refers to this small amplitude differential signaling technique.

For PCB engineers, the most important thing is to ensure that these advantages of differential routing can be fully exploited in the actual routing. Perhaps anyone who has been exposed to Layout will understand the general requirements for differential routing, which is "equal length, equidistance." The equal length is to ensure that the two differential signals maintain the opposite polarity at all times and reduce the common mode component; the equidistance is mainly to ensure that the differential impedances of the two are consistent and reduce reflection. "As close as possible to the principle" is sometimes one of the requirements for differential routing. But all these rules are not used to make a hard copy, and many engineers do not seem to understand the nature of high-speed differential signaling.

Cabling strategy

The differential signal pairs that are in close proximity to each other are also tightly coupled to each other. This mutual coupling reduces EMI emissions. The main disadvantage of differential signal lines is the increased PCB area. This paper introduces the circuit board design process. Wiring strategy for differential signal line routing. It is well known that signals exist along signal lines or under PCB lines. Even though we may not be familiar with single-ended mode routing strategies, the term single-ended distinguishes this transmission characteristic from that of differential and common-mode signals. The latter two signal transmission methods are usually more complicated.

Differential and common mode

The differential mode signal is transmitted through a pair of signal lines. One signal line transmits the signal we normally understand; the other signal line transmits a signal that is equal and opposite in direction (at least in theory). Differential and single-ended modes initially differed little because all signals have loops. Signals in single-ended mode are typically returned via a zero voltage circuit (or ground). Each of the differential signals is returned through the ground circuit. Since each signal pair is actually equivalent and inverted, the return circuits simply cancel each other out, so that the components of the differential signal return do not occur on the zero voltage or ground circuit. Common mode means that the signal appears on two signal lines of a (differential) signal pair, or both on a single-ended signal line and on the ground. The understanding of this concept is not intuitive, because it is difficult to imagine how to generate such a signal. This is mainly because we usually do not generate common mode signals. Most of the common mode signals are noise signals generated in the circuit according to imaginary conditions or coupled by adjacent or external signal sources. Common-mode signals are almost always "harmful" and many design rules are designed to prevent the appearance of common-mode signals.

Wiring of differential signal lines

Usually (with some exceptions) differential signals are also high-speed signals, so high-speed design rules are often applied to the routing of differential signals, especially when designing signal lines such as transmission lines. This means that we must design the routing of the signal lines very carefully to ensure that the characteristic impedance of the signal lines are continuous throughout the signal line and remain constant. During the placement and routing of differential pairs, we want the two PCB lines in the differential pair to be exactly the same. This means that in practical applications, every effort should be made to ensure that the PCB lines in the differential pairs have exactly the same impedance and that the length of the wiring is exactly the same. Differential PCB lines are usually always wired in pairs, and the distance between them remains constant at any position along the direction of the pair. In general, the layout of differential pairs is always as close as possible.

The following focuses on several common misunderstandings in PCB differential signaling design.

Misunderstanding 1: It is considered that the differential signal does not require the ground plane as the return path, or that the differential traces provide each other with a return path. The cause of this misunderstanding is that it is confused by surface phenomena, or the understanding of the mechanism of high-speed signal transmission is not deep enough. Differential circuits are insensitive to similar ground bounce and other noise signals that may be present on the power and ground planes. The partial return cancellation of the ground plane does not mean that the differential circuit does not use the reference plane as the signal return path. In fact, in the signal reflow analysis, the mechanism of the differential trace and the common single-ended trace is the same, that is, the high frequency signal is always The reflow is performed along the loop with the smallest inductance. The biggest difference is that in addition to the coupling to the ground, the differential lines have mutual coupling, which one is strong and which one becomes the main return path.

In the PCB circuit design, the coupling between the differential traces is generally small, often only 10 to 20% of the coupling degree, and more is the coupling to the ground, so the main return path of the differential traces still exists in the ground plane. . When the local plane is discontinuous, there is no reference plane, and the coupling between the differential traces will provide the main return path, although the discontinuity of the reference plane does not affect the ordinary single-ended trace. It is serious, but it will reduce the quality of differential signals and increase EMI. Try to avoid it. Some designers believe that the reference plane below the differential trace can be removed to suppress some of the common-mode signals in the differential transmission, but in theory this approach is not advisable. How is the impedance controlled? Not providing the ground impedance loop to the common mode signal will inevitably cause EMI radiation, which is more harmful than good.

Myth 2: It is considered more important to keep the equal spacing longer than the matching line length. In actual PCB layout, the requirements of differential design cannot often be met at the same time. Due to factors such as pin distribution, vias, and routing space, the wire length matching must be achieved by proper winding, but the result is necessarily that the partial regions of the differential pair cannot be parallel. The most important rule in the design of PCB differential traces is the matching line length. Other rules can be flexibly processed according to design requirements and practical applications.

Misunderstanding 3: It is believed that the differential traces must be close. Bringing the differential traces closer is nothing more than enhancing their coupling, both to improve immunity to noise and to take advantage of the opposite polarity of the magnetic field to counteract electromagnetic interference to the outside world. Although this practice is very beneficial in most cases, it is not absolute. If we can guarantee that they are fully shielded from external interference, then we do not need to let the strong coupling through each other to achieve anti-interference. And the purpose of suppressing EMI. How can we ensure that the differential traces have good isolation and shielding? Increasing the distance from other signal traces is one of the most basic ways. The electromagnetic field energy decreases with the square of the distance. When the line spacing is more than 4 times the line width, the interference between them is extremely weak. Can be ignored. In addition, ground plane isolation can also be used for shielding. This structure is often used in high-frequency (10G or more) IC package PCB design. It is called CPW structure and can guarantee strict differential impedance. Control (2Z0).

Differential traces can also be used in different signal layers, but this method is generally not recommended because differences in impedance, vias generated by different layers can disrupt the effects of differential mode transmission and introduce common mode noise. In addition, if the adjacent two layers are not tightly coupled, the ability of the differential traces to resist noise is reduced, but crosstalk is not an issue if the proper spacing from the surrounding traces is maintained. At normal frequencies (below GHz), EMI is not a serious problem. Experiments show that the differential excitation of 500Mils is 60dB away from the radiated energy of 3m, which is enough to meet the FCC electromagnetic radiation standard. The designer does not have to worry too much about the electromagnetic incompatibility caused by insufficient coupling of the differential lines.

Misunderstanding four

Differential Manchester coding is not a type of differential signal. It refers to a level transition at the beginning of each bit to indicate a logic state of "0" and no transition to indicate a logic state of "1". But each intermediate transition is used to make a synchronous clock, which has no logical meaning.

Misunderstanding five

The twisted pair does not necessarily have a differential signal. The electromagnetic radiation of the single-ended signal on the twisted pair is also smaller than that of the parallel trace.