Introduction to common knowledge about eye diagrams – ChangZhou Sun-Rise Electronic Co.,Ltd.

Signal rise, fall time, cross point position, extinction ratio, Q factor, signal-to-noise ratio, jitter, etc. and how to measure the pros and cons of an eye diagram from various aspects, we mainly discuss with CSA8000.

Introduction to eye diagrams and commonly used indicators

The picture below shows the eye diagram of a 10G optical signal. The left side is the shape of the eye diagram and the template of the 10G eye diagram. The right column shows some measured values of this optical signal. From top to bottom, they are the extinction ratio (ExdB); the intersection point. Ratio (Crs); Q factor (QF); average optical power (AOP); rise time (Rise); fall time (Fall); peak jitter (PFJi); root mean square jitter (RMS)

The extinction ratio is defined as the value of 1 level to 0 level in the eye diagram; in the recommendations, there are different requirements based on different rates and transmission distances; when testing the eye diagram of the optical module, different laser types have different requirements; General requirements for FP/DFB direct-tuned lasers; the extinction ratio is not less than 8.2dB; the extinction ratio of EML electro-absorption laser is not less than 10d. There is no maximum value for the extinction ratio in BITU-T, but this does not mean that the extinction ratio can be infinitely large. , If the extinction ratio is too high, the chirp coefficient of the laser will be too large, resulting in excessive channel cost, which is not conducive to long-distance transmission. It is generally recommended that the actual extinction ratio is 0.5 larger than the actual optical interface type, and the minimum required extinction ratio related to the rate transmission distance ~1.5dB, this is not an absolute value. The reason why such a value is given is that the extinction ratio is too high, and the signal will deteriorate too much after transmission, resulting in bit errors or excessive channel cost. If an optical module transmits its After the nominal distance, no bit errors occur and the channel cost meets the index requirements. As long as the extinction ratio is greater than the minimum value recommended by ITU-T, it can be any large.

The cross point ratio reflects the duty cycle of the signal. During the transmission process, the pulse width of the optical signal will be widened, causing the cross point on the receiving side to move up relative to the sending side. In order to facilitate long-distance transmission, ensure the crossing point on the receiving side The point ratio is about 50, which makes the sensitivity of the receiving side the best. We generally recommend that the position of the cross point be moved down slightly on the sending side. Generally, the cross point ratio of the sending side is recommended to be controlled at 40%-45%;

The Q factor comprehensively reflects the quality of the eye diagram; the higher the Q factor, the better, indicating that the quality of the eye diagram is better, and the Q factor is generally affected by noise; optical power, whether the electrical signal is impedance matched from the beginning to the end and other factors, generally speaking , the thinner and smoother the line at level 1 in the eye diagram, the higher the Q factor. In the case of no optical attenuation, the Q factor of the optical eye diagram at the transmitting side should not be less than 12. The higher the better, the receiving measurement The Q factor should not be less than 6. The higher the better, the CSA8000 can measure the average optical power approximately. If you need to measure the optical power accurately, it is recommended to use an optical power meter.

The rising time and falling time of the signal reflect the rising and falling speed of the signal, which generally refers to the change time of 20%-80% of the entire signal amplitude; it is generally required to rise; the falling time cannot be greater than 40% of the signal cycle; 9.95G signal requires its rising and falling time to be no more than 40ps.

Peak jitter and RMS jitter; it can qualitatively reflect the jitter of the signal. As a reference for comparison, the smaller the two measured values, the better. Quantitative measurement of output jitter requires a special instrument for testing jitter, such as Agilint’s 37718, ACTERNA's ANT20-SE When measuring jitter, the instrument generally needs to be warmed up for more than 30 minutes to ensure relatively accurate measurement values.

The signal-to-noise ratio can also qualitatively reflect the quality of the signal. As a comparison reference, the larger the measured value, the better. Generally, the measured value on the sending side is greater than 30dB; quantitative measurement requires spectral analysis.

Typical Eye Diagram Introduction

Next, let's look at some typical better eye diagrams and some problematic eye diagrams; and analyze where the problems of these eye diagrams are;

The following is a better 622M eye diagram. We can see that the eye diagram is relatively symmetrical; the eyeliner is very thin, the extinction ratio is moderate, and the Q factor is very high, reaching 24.

The following is the eye diagram of 622M without STM-4 filter; it can be seen that the eye line of the eye diagram is thin; especially the rising and falling edge 1 level is a little rippled. This is because without the low-pass filter, the signal The high-frequency harmonics are not filtered out, and the harmonic components add up to form a rippled square wave. We see that even if the 1 level is not flat, its Q factor still reaches 21.7.

The following is a better 2.5G, the eye pattern is relatively symmetrical, the eyeliner is relatively thin, the 0; 1 levels are relatively smooth, the extinction ratio is moderate, and the Q factor is relatively high.

The following is a better 10G eye diagram, the eye diagram is symmetrical, the eye diagram is relatively thin, especially the 0, 1 level rises, and the falling edge is a little thicker. It can be seen that the jitter of the signal is larger, the extinction ratio is moderate, and the Q factor is higher. The crossover point is slightly higher. In actual debugging, the crossover point can be lowered a little.

In general, the higher the rate, the worse the quality of the eye diagram, which is mainly caused by two aspects, the first is jitter, the higher the rate, the more difficult it is to control the jitter, and the second is noise, because the test process generally needs to add corresponding The low-pass filter of the 10G signal has a bandwidth of about 8GHz, and the bandwidth of the low-pass filter of the 622M signal is about 500MHz. The noise in the frequency range from 500MHz to 8GHz is filtered out by the filter of the 622M signal , but it was not filtered out by the filter of the 10G signal, so the noise of the 10G signal is larger from the eye diagram.

Problematic Eye Diagram Introduction

The following is a problematic 622M eye diagram; this eye diagram has many problems; let’s analyze them one by one. First; the eye diagram has two very obvious rising and falling edges (commonly known as double eyelids); 0 level 1 level is not flat ; The signal has overshoot and undershoot; the extinction ratio is low, only 4.1dB; the reason for these phenomena is suspected to be the impedance mismatch of the signal; resulting in overshoot, undershoot and multipath of the signal; this eye diagram also illustrates another Question: It is only the most basic requirement of the eye diagram that the eye diagram can fit the template; it is not the only requirement; let's take a look and see that there is still a certain margin between the edges of the eye diagram and the template.

Let's take a look at the following 622M eye diagram. The problem is that the noise is very large. It is estimated that the signal filtering is not processed well.

The following is a 2.5G eye diagram. The overall quality is not bad. The problem is that the eye diagram is a bit skewed and asymmetrical. This has a certain relationship with the modulation characteristics of the laser.

The following 2.5G eye diagram, the problem is that the jitter is relatively large. Note that compared with the previous eye diagram, its rising and falling edges are thicker. Pay special attention to comparing its peak-to-peak jitter and root mean square jitter. big.

The following 2.5G eye diagram is relatively bad, the eye diagram twists and turns, the rise and fall are very slow, the signal quality is not good, the Q factor is only 6.4, and the extinction ratio is only 6.6dB. The reason may be the driver laser The problem itself, or the impedance is very mismatched.

The following is a 2.5G eye diagram. It can be clearly seen that there is ringing on the rising edge of the eye diagram. This may be caused by two reasons. The first is the impedance mismatch on the signal line, and the second is the oscillation caused by the relaxation oscillation of the direct-controlled laser. bell.

The following is a 10G eye diagram. There are two problems with the eye diagram. First, the extinction ratio is too low, only 10dB. The level of eye diagram 1 is very thick and uneven. The possible reasons are: signal mismatch.

The following 10G eye diagram does not have its measurement data, but it can be seen from the thicker rising and falling edges of the eye diagram that the signal jitter is relatively large.

The following is the 10G eye diagram. The problem with this eye diagram is that the noise is relatively large. It can be seen from there. Please note that the rise and fall of the eye diagram are relatively thick, and the entire eye diagram has many scattered points, which is very dirty.

For the above three eye diagrams, we have analyzed the three situations that lead to poor eye diagrams, impedance mismatch, jitter, and noise. How can these three situations be seen from the eye diagrams?

If the 1-level line of the eye diagram is thick and uneven, it is caused by impedance mismatch

If the rising and falling edges of the eye diagram are relatively thick, it is caused by jitter. To solve the problem, we must start by reducing signal jitter, such as improving the quality of the input clock, and reasonably designing the phase-locked loop, especially the low-pass filter part.

If the eye diagrams are relatively thick, then it is caused by noise. Generally speaking, it is caused by power supply noise, poor ground loop or large interference sources around the signal. To solve the problem, we must start from these aspects. For eye diagrams, you cannot Measured by a ruler, the higher the rate, the more difficult it is to guarantee the quality of the eye diagram. The longer the target transmission distance, the better the quality of the eye diagram. At the same time, the optical module with data and clock input is better than the optical module with only data input. The picture quality will be better, especially in terms of jitter. The eye diagram of the EA modulation method will perform better than the eye diagram of the direct modulation method.

Sun-Rise Electronic / Calukii US Store

Sep 4th, 2023