Eye Diagram and Digital Signal Testing

Updated Jun 17, 2021

Reported In


  • PXI Oscilloscope


  • LabVIEW
  • LabVIEW Jitter Analysis Toolkit

Issue Details

  1. What is an eye diagram?
  2. What is it used for?
  3. What information is reflected in the waveform?
  4. What is the corresponding solution for NI?


The Eye Diagram can show the transmission quality of digital signals. It is often used in applications where electronic devices, serial digital signals or high-speed digital signals in chips are tested and verified. In the final  analysis, the quality of digital signals is fast by intuitive means of observation. In consumer electronics, high-speed signal transmission is often used inside the chip, between the chip and the chip. If the corresponding signal quality is not good,, the device will be unstable, function execution error, or even malfunction. The eye diagram reflects that the digital signal is affected by the physical device and the channel. Engineer can quickly obtain the measured parameters of the signal in the product to be tested through the eye diagram, and can predict the problems that may occur in the field.

1 Formation of the Eye Diagram
For digital signals, the high and low transitions can be combined in multiple sequences. Taking 3 bits as an example, there may be a combination of 000-111 and 8 in total. In the time domain, enough sequences are aligned according to a certain reference point, and then the waveform are superimposed to form an eye diagram. As in Figure 1. For the test instrument, the clock signal of the signal is first recovered from the signal to be tested, and then the eye diagram is superimposed according to the clock reference, and finally displayed.
Figure 1: Formation of the Eye Diagram

2. Information contained in the Eye Diagram

For a real eye diagram, as show in Figure 2, we can first see the average erise time (Rise Time), fall time (Fall Time), Overshoot, Undershoot, and other basic level shift parameters such as Threshold/ Crossing Percent of the digital waveform.

Figure 2: Level shift parameter
It is impossible for the signal to maintain the same voltage value every high and low level, and it is not guaranteed that the rising and falling edges of each high and low level are at the same time. As shown in Figure 3, due to superposition of multiple signals, the signal line of the eye diagram becomes thicker, and a blur (Blur) phenomenon occurs. Therefore, the eye diagram also reflects the noise and jitter of the signal: on the vertical axis voltage axis, it is reflected as Voltage Noise; on the horizontal axis time axis, it is reflected in the time domain(Jitter).

Figure 3: Voltage Noise and Jitter

The blank area on the eye diagram becomes smaller due to noise and jitter. As shown in Figure 4, on the basis of removing jitter and noise, the distance of the blank area on the eye digram on the horizontal axis is called Eye Width. When the data superimposed on the eye diagram is sufficient, the eye width is well reflected. The stability time of the signal on the transmission line; similarly, the distance on the vertical axis of the blank area on the eye diagram is called Eye Height. When the data superimposed on the eye diagram is sufficient, the eye height reflects the transmission line very well. The noise margin of the upper signal. At the same time, the highest eye height in the eye diagram is the best judgement moment.

Figure 4: Eye Height and Eye Width

Before and after sampling, the digital signal needs to have a certain setup time and hold time. The digital signal should be stable during this period to ensure correct sampling, as shown in the blue part of Figure 5.1. For the input level decision, the high level voltage value is required to be higher than the input high level VIH, and the low level voltage value is input with the low level VIL, as shown in the green portion of Figure 5.1. Therefore, we can know that the earliest sampling time and the latest sampling time are shown in Figures 5.1 and 5.2.

Figure 5.1: Sampling and judgement a

Figure 5.2: Sampling and judgement b

At the optimal sampling time, the bit error rate of the sample is the lowest, and the bit error rate increases as the sampling time moves to both sides of the time axis, as shown in Figure 6. Therefore, the engineering often also draws a curve of the bit error rate during the signal sampling period, called the Bathtub Curve.

Figure 6: Bathtub Curve

In the actual test, in order to improve the efficiency of the testing, the method often used is Mask Testing. That is, according to the requirements of signal transmission, an area (such as the diamond-shaped area in Figure 7) is specified on the eye diagram, and all the left and right signals are required  to appear outside the area. Once a signal appears in the diamond area, the test fails.

Figure 7: Mask Testing

3 Tools from NI

For transmission performance test analysis of digital signals, acquisition and analysis can be separated.

At the acquisition end, it is necessary to select an oscilloscope with the appropriate sampling rate, appropriate bandwidth and effective quantized bits according ​​​to the actual signal frequency, bandwidth, level characteristics, etc. NI provides a variety of oscilloscope cards for measurement, among them, PXIe-5186 oscilloscope can achieve 12.5 GS/s sampling, 5GHz bandwidth.

On the analysis side, you cna use the Signal Processing Toolkit (Function Palette - Signal Processing) in LabVIEW to perform simple processing on digital signals, including waveform measurement, analysis, and transformation. Testing for high speed digital signals;

LabVIEW also offers the Jitter Analysis Toolkit, which provides VIs for clock recovery, eye diagram measurements, jitter measurements, level detection, jitter signal generation, and timing measurements.

The correlation display analysis function of the eye diagram is shown in Figure 8, including eye diagram display, Mark Testing, eye diagram parameter calculation. The detection of other parameters mentioned in the paper can find the corresponding function in the board of clock recovery, jitter measurement, level detection, jitter signal generation, timing measurement, see Table 1.

Figure 8: Jitter Analysis Toolkit

Clock Recovery VIRecovery of embedded clocks from data streams. VIs for :
Mean (constant) clock recovery
First and second order PLL clock recovery
Eye Diagram
Measurement VIs
Eye visualization (color, scaling, data segmenting)
Eye mask definition
Eye width, height, and quality factor measurements
Jitter VIs
Cycle-to-cycle jitter, N-cycle jitter, Period jitter, N-period jitter
Periodic jitter (PJ)
Time interval error (TIE)
Total jitter (TJ)
Duty cycle distortion (DCDI)
Data dependent jitter (DDJ)
Intersymbol inteference (ISI)
Phase noise, Cycle-to-cycle measurement, N-cycle measurement
Bathtub plot support
Level VIsArea, Find transitions, Max and min voltages
Overshoot and undershoot, Peak-to-peak voltage
RMS voltage, Reference levels, State levels
Signal Generation VIsGenerates a signal with jitter that can be used to validate and/ or
debug user-defined VIs
Timing VIsBurst width, Channel-to-channel skew, Duty cycle, Fall time, Level crossing
Negative width, Period, Phase, Positive width, Rise time, Setup and hold time
Standard LabVIEW FunctionsStatistics, Spectral analysis
Table 1: Jitter Analysis Toolkit VIs Function