An eye diagram is a visual tool used to assess the quality of a digital signal. It’s created by overlaying many consecutive bit sequences on top of each other on an oscilloscope display, producing a pattern that resembles a human eye. The wider and taller that “eye” appears, the cleaner the signal. A closed or narrow eye means the signal is degraded and more likely to produce errors.
How an Eye Diagram Is Created
A digital signal is just a stream of ones and zeros, represented as high and low voltage levels over time. To build an eye diagram, an oscilloscope captures this stream and chops it into segments, each one or two bit periods long. It then stacks all those segments on top of each other in a single display. Every possible transition between bits (high to low, low to high, high to high, low to low) gets drawn repeatedly in the same frame.
The result is a composite image. The traces pile up and form a shape with an open area in the middle, the “eye.” Clean signals produce crisp, well-defined traces with a wide-open eye. Noisy or distorted signals produce fuzzy, scattered traces that shrink the opening. You can think of it like laying hundreds of photographs on top of each other: if the signal is consistent, the images line up neatly. If something is off, you get blur.
What the Eye Shape Tells You
Every feature of the eye diagram maps to a specific aspect of signal quality.
The height of the eye opening represents the voltage margin available for the receiver to correctly distinguish a one from a zero. A taller eye means more room for error, so the receiver can reliably decode the signal even if some noise is present. As the eye closes vertically, the receiver has less margin and starts misreading bits.
The width of the eye opening represents the timing margin. A wider eye means the receiver has a larger window in which to sample the signal and get a correct reading. When the eye narrows horizontally, the timing window shrinks, and the receiver must sample at precisely the right moment or risk errors.
The crossing points, where traces intersect on the left and right sides of the eye, reveal how cleanly the signal transitions between voltage levels. Ideally these crossings are tight and well-defined. When they spread out vertically or horizontally, it indicates noise or timing problems.
When the center of the eye pattern is completely closed, the signal is too degraded for reliable communication. The receiver can no longer tell ones from zeros at any sampling point.
Connection to Bit Error Rate
The size of the eye opening directly correlates with the bit error rate (BER), which is the fraction of bits that a receiver decodes incorrectly. BER is one of the most important metrics in any communication system because it defines the channel’s reliability. A wide, tall eye corresponds to a low BER. As the eye closes due to noise, jitter, or distortion, the BER climbs because the receiver’s voltage and timing margins shrink simultaneously. Engineers use measurements from the eye diagram, combined with jitter data, to estimate BER without needing to count billions of individual bit errors.
How Jitter Appears in the Eye
Jitter is the variation in the timing of signal transitions. Instead of arriving at perfectly regular intervals, the edges of the signal shift slightly earlier or later than expected. This shows up in the eye diagram as a horizontal thickening of the crossing points, which narrows the eye’s width.
There are two main types. Random jitter comes from unpredictable sources like thermal noise in electronic components. If your system had only random jitter, every signal edge would have the same probability of timing error, and the crossings would blur symmetrically. Deterministic jitter, on the other hand, is pattern-dependent. Some edges consistently shift in one direction while others shift the opposite way, creating distinct, repeating trace paths within the diagram.
You can often distinguish between the two using an oscilloscope’s color-graded or variable-intensity display. Well-defined bright paths within the eye indicate deterministic jitter, because certain transitions are consistently landing in the same offset positions. The random component then causes those shifted edges to bounce around their offset locations, adding fuzz on top of the structured pattern. Identifying which type dominates helps engineers target the right fix: deterministic jitter often points to a specific design flaw, while random jitter is limited by fundamental physics.
Common Causes of a Degraded Eye
Several physical problems in a transmission channel cause the eye to close. Signal loss over long cables or circuit board traces reduces the voltage swing, shrinking the eye vertically. This is especially problematic at high frequencies, where losses increase. Impedance mismatches, such as an open circuit, short circuit, or abrupt change in trace width on a circuit board, cause reflections that send energy bouncing back through the channel. These reflections arrive at the receiver at unpredictable times and distort the waveform.
Crosstalk from adjacent signal lines injects unwanted energy into the channel. And intersymbol interference (ISI), where the residual energy from one bit bleeds into the next, is one of the most common culprits at high data rates. ISI happens because real-world channels can’t perfectly switch between voltage levels instantaneously. The leftover energy from previous bits smears into the current bit, closing the eye from both the voltage and timing sides simultaneously.
NRZ vs. PAM4 Eye Diagrams
The simplest and most common type of eye diagram comes from NRZ (non-return-to-zero) signaling, where the signal has two voltage levels representing a zero or a one. This produces a single eye opening. There are only two distinct transitions: low-to-high and high-to-low.
As data rates have increased, many modern systems have adopted PAM4 (pulse amplitude modulation with four levels) signaling. PAM4 uses four voltage levels instead of two, with each level representing a two-bit combination: 00, 01, 10, or 11. Because it encodes two bits per symbol, PAM4 requires only half the bandwidth of NRZ for the same data throughput.
The tradeoff is complexity. A PAM4 eye diagram has three stacked eye openings instead of one, corresponding to the spaces between the four voltage levels. There are 12 distinct transitions instead of two. The individual eyes are smaller than a single NRZ eye because the voltage range is divided among more levels, which means tighter margins and greater sensitivity to noise. The lower and upper eyes are also typically asymmetric, with their widest points slightly off-center relative to the voltage levels, making analysis more involved. Engineers use specialized algorithms to measure the height of each of the three eyes independently.
How Engineers Use Eye Diagrams in Practice
The most common practical application is mask testing. A “mask” is a predefined template that sits inside the eye opening on the oscilloscope display. If any signal traces cross into the mask area, the signal fails the test. Industry standards for interfaces like USB, Ethernet, HDMI, and PCIe each define their own mask shapes and sizes, setting the minimum acceptable eye opening for that standard. This gives engineers a quick pass/fail result without needing to calculate BER directly.
Eye diagrams are also used during the design and debugging of high-speed circuit boards, cables, backplanes, and optical links. When a new design produces a closed or marginal eye, engineers can work backward from the shape of the distortion to identify whether the problem is jitter, noise, loss, reflections, or crosstalk. The visual nature of the diagram makes it faster to diagnose issues than purely numerical measurements, which is why it remains one of the most widely used tools in signal integrity work despite being a decades-old technique.