3.12 Cross Coupling between Parallel Transmission Lines of Equal Length

Parallel transmission lines can be series or parallel terminated with the same or opposite signal directions. Culprit and victim logic transitions can be in the same direction or in opposite direction. The culprit and victim can be unterminated at either end. There can be a culprit on both sides of one victim. A complete analysis of signal integrity requires an analysis of each case. In the following discussion, we will assume that the culprit wave is positive and goes to the right.

Consider the case where the culprit line is series terminated and the culprit makes a round trip during the rise time. The cross talk is composed of reverse coupling for half the round trip time and forward coupling during the return time. The coupled signal is maximum when the culprit signal has made a round trip. This line length is called the critical length.

The distance a wave travels in a rise time is called a TEL that stands for transitional electrical length. Lines longer than one-fourth TEL that are not terminated are apt to fail. In the example above, the critical length is half of TEL.

The following cases are a sample of the cross talk problem. A treatment of every case would be of little value, as there are just too many variables.

Case 1. Victim line is series terminated at the left.
At the point of first coupling, the reverse wave is absorbed in the victims matching impedance. There is no reflection. The forward pulse has a width equal to the rise time of the culprit and is usually of small amplitude. This pulse is reflected at the open end of the victim line, and it is not present at the next clock time. If the victim is at logic one and the coupling exceeds the clamping voltage then a reflection will result.
Case 2. Victim line is open circuited at the left.
The amplitude of the forward coupling on the victim line is proportional to time. When this pulse reaches the victim's series termination, the wave is absorbed and poses no problem. The reverse coupled wave is doubled and is reflected forward. If the doubled voltage exceeds the clamping voltage then a reflection or damage can result. This reverse wave lasts twice as long as the culprit wave, so it could add to the logic signal at the next clock time. Under some conditions this could cause a logic error.
Case 3. Forward crosstalk when the victim line is a zero impedance at the left.
In this case, the victim line is parallel terminated at the right end of the line. The victim voltage, when the coupling starts, can be a logic 1 or a logic 0. Logic errors are apt to occur when the coupled wave drives the logic state toward the transition region. In this region, the logic state is undefined. This case is shown in Figure 3.11.

Figure 3.11 Coupling to a transmission line with a zero impedance source.

3.11
N.B.
In any logic design, the error budget must allow for all modes of cross coupling.

The coupling between traces is often likened to transformer action where the transformer coils are single turns. This approach does not consider the character of the forward and reverse coupled waves.

Guard traces are sometimes used in an attempt to reduce cross coupling. The approach assumes that added grounds act as shields. The approach is flawed because these traces might function as resonant couplers. The wide spectrum associated with logic transmission means that logic signals can cause ringing in any nearby unterminated transmission lines. The result is added coupling rather than shielding.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
18.191.235.210