Eye Measurement
Libraries:
Mixed-Signal Blockset /
Utilities
SerDes Toolbox /
Utilities
Signal Integrity Toolbox /
Utilities
Description
The Eye Measurement block generates eye pattern from time-domain waveform data and calculates metrics from the generated pattern.
Examples
This example shows how to produce eye diagrams and related metrics for sampled data systems in Simulink.
Set Up Eye Measurement Block
Open the model eyeMeasurementInDFECDRExampleModel attached with this example. The model was originally created in the SerDes Designer app and then exported to Simulink®. For more information, see Design SerDes System and Export IBIS-AMI Model (SerDes Toolbox).
model = "eyeMeasurementInDFECDRExampleModel";
open_system(model);
In the eyeMeasurementInDFECDRExampleModel model, double-click on the Rx subsystem. Then right-click the DFECDR and select Look under mask to enter the DFECDR subsystem. Add an Eye Measurement block.
Note The Eye Measurement block is not supported for IBIS-AMI model generation. Remove the Eye Measurement block before exporting the model to IBIS-AMI.
Double-click on the Eye Measurement block to open its dialog.
Set the Configuration tab parameters to match the signal:
Set Enable threshold input port to
onSet Modulation to
ModulationSet Enable clock input port to
onSet Amplitude range to
[-0.2, 0.2]Set Hold off time to
IgnoreBitsand its unit toUISet Symbol time (s) to
SymbolTimeSet Sample time (s) to
SampleInterval
phaseOffset = 0.7;
Modulation, IgnoreBits, SymbolTime, and SampleInterval are model properties controlled by the Configuration block in the top level of the model.
In the Eye Mask tab:
Set Eye mask type to
RectangularSet W1 (UI) to
0.2Set H to
0.1
In the Metric Setup tab:
Set Store results in base workspace to
onSet Result variable name to
eyeResults
This saves the metrics to the base workspace in the variable named eyeResults once the simulation ends.
Apply the changes.
Connect Eye Measurement Block to DFECDR
Connect the data input of the Eye Measurement block to the output signal from the DFECDR. Add a Bus Selector block. Connect the input of the bus selector to the ClkAMI output port on the DFECDR. Use the Bus Selector to select the signal clockTime from the bus.
Connect the output port from the bus selector to the clock input port on the Eye Measurement block. Connect the PAM_Thresholds signal from the DFECDR to the thresholds input to the Eye Measurement block.
open_system(model + "/Rx/DFECDR", "force");
Run the simulation and confirm that the eye opening is centered, the correct number of unit intervals is visible, and that the dynamic range of the signal is fully within the diagram.
sim(model);
Export Metric Results
Once the simulation is complete, find the metric results in the base workspace.
disp(eyeResults);
Metric Position Unit SER Extrapolation Results
_________________ ________ _________ _____ _____________ ____________
"Eye Height" {[ 0]} "Seconds" 1e-05 "None" {1×1 struct}
"Best Eye Height" {[NaN]} "NaN" 1e-05 "None" {1×1 struct}
"Eye Mask Margin" {[NaN]} "NaN" 1e-05 "none" {1×1 struct}
To retrieve the width and height margins, select the last element in the results column of the table, and from that structure retrieve the field named Margin.
eyeResults.Results{end}.Margin
ans = -0.0035 -0.0255
This example shows how to produce eye diagrams and related metrics for specific signals in a SerDes system using the Eye Measurement block.
Set Up Constants
This example compares eye openings at three different signals in the model:
The input to the receiver
The output of the CTLE
The output of the DFE
Open the model eyeMeasurementExampleModel_1Eye attached with this example. The model was originally created in the SerDes Designer app and then exported to Simulink®. For more information, see Design SerDes System and Export IBIS-AMI Model (SerDes Toolbox).
model = "eyeMeasurementExampleModel_1Eye";
open_system(model);
The Eye Measurement block does not have a phase detector or CDR, but the eye opening must be centered in the diagram for metric calculations to succeed. Use a manual phase offset to center the eye opening.
phaseOffset = 0.7; % UI
Measure Eye at DFE Output
In the eyeMeasurementExampleModel_1Eye model, double-click on the Rx subsystem. Add an Eye Measurement block.
Note Eye Measurement block is not supported for IBIS-AMI model generation. Remove the Eye Measurement block(s) before exporting the model to IBIS-AMI.
Double-click on the Eye Measurement block to open the Block Parameters dialog box.
Set the Configuration tab parameters to match the signal:
Set Enable threshold input port to
offSet Modulation to
ModulationSet Symbol thresholds to
0Set Enable clock input port to
offSet Phase offset to
phaseOffsetand its unit toUISet Amplitude range to
[-0.2, 0.2]Set Hold off time to
IgnoreBitsand its unit toUISet Symbol time (s) to
SymbolTimeSet Sample time (s) to
SampleInterval
Modulation, IgnoreBits, SymbolTime, and SampleInterval are model properties controlled by the Configuration block in the top level of the model.
Apply the changes.
Connect the data input of the Eye Measurement block to the output signal from the DFECDR. Name the Eye Measurement block Eye 3: DFE.
open_system(model + "/Rx");
Run the simulation and confirm that the eye opening is centered, the correct number of unit intervals is visible, and that the dynamic range of the signal is fully within the diagram.
sim(model);
Warning: Warning from 'eyeMeasurementExampleModel_1Eye/Rx/Eye 3: DFE': Number of eye openings at center (3) was not the expected number of eye openings (1), metrics may be incorrect. Ensure 'Modulation' is correct, open the eye further, or manually center the eye using 'PhaseOffset'.
Measure Eye at CTLE Output
Copy the Eye Measurement block and paste the copy into the model. Rename the copy to Eye 2: CTLE. Connect its data input to the output of the CTLE.
Measure Eye at Receiver Input
Paste another copy of the Eye Measurement block and rename it to Eye 1: Channel. Connect its data input to the input to the CTLE.
bdclose(model); model = "eyeMeasurementExampleModel_3Eye"; load_system(model); open_system(model + "/Rx");
Running the simulation as-is produces errors when the Eye Measurement block connected to the CTLE input attempts to calculate its metrics because its eye is closed. To avoid this, double click on the block to open its dialog and change the following parameters:
In the Metric Setup tab, click on each metric and then click Remove Selected Metrics. You can shift-click to select multiple metrics.
Additionally, to fit the data in the eye diagram, in the Configuration tab set Amplitude range to [-0.5, 0.5].
Apply the changes.
Run the simulation. There will be a warning because Eye 1: Channel is measuring a closed eye. This is expected, because that eye is measured at the output of the channel, before the receiver. The subsequent eyes (Eye 2: CTLE and Eye 3: DFE) are open due to RX equalization performed by the CLTE and DFE.
sim(model);
Warning: Warning from 'eyeMeasurementExampleModel_3Eye/Rx/Eye 1: Channel': Number of eye openings at center (0) was not the expected number of eye openings (1), metrics may be incorrect. Ensure 'Modulation' is correct, open the eye further, or manually center the eye using 'PhaseOffset'. Warning: Warning from 'eyeMeasurementExampleModel_3Eye/Rx/Eye 2: CTLE': Number of eye openings at center (3) was not the expected number of eye openings (1), metrics may be incorrect. Ensure 'Modulation' is correct, open the eye further, or manually center the eye using 'PhaseOffset'. Warning: Warning from 'eyeMeasurementExampleModel_3Eye/Rx/Eye 3: DFE': Number of eye openings at center (3) was not the expected number of eye openings (1), metrics may be incorrect. Ensure 'Modulation' is correct, open the eye further, or manually center the eye using 'PhaseOffset'.
Limitations
Eye Measurement block does not support rapid accelerator mode.
Ports
Input
Parameters
More About
The Eye Measurement block extrapolates its 2-D histogram to a specified symbol error rate whenever it generates bathtub curves or eye contours.
During extrapolation method, the block pre-processes the data, one symbol at a time on only a 1-D slice of the said symbol. The block extrapolates the eye diagram at the specified SER value using the interpolation between the adjacent bins.
The block provides five interpolation methods (none,
linear, spline,
pchip, and makima) and two
extrapolation methods (gaussian and
dualdirac) to extrapolate data.
This image shows how the block uses interpolation methods to extrapolate data.
The blue dots represent the cumulative sum of the eye diagram going outward from the center of the eye. The blue dot at the x-axis value zero represents the eye opening. The dotted red line shows the interpolated samples.
The log scale image is used to create the bathtub curves. The discontinuity at the logarithmic scale is to show the zero x-axis value. The linear scale image shows how the cumulative sum of the 1-D slice of the eye looks on the same scale as the slice itself..
For example, the none extrapolation method uses a previous
neighbor interpolation of the cumulative sum an eye slice. For horizontal eye slices, the
extrapolation uses the timing origins. For vertical eye slices, the extrapolation uses the
symbol thresholds.
When moving outward from the center of the eye, it is a previous neighbor interpolation. When moving across the eye from one side to the other, it appears as a next neighbor interpolation that switches to a previous neighbor interpolation as you pass the center of the eye. This way, the result for a symbol error rate is a conservative estimate from the perspective of the eye opening, based on the data.
On the other hand, the dualdirac extrapolation algorithm
first fits the Dual-Dirac PDF to a column of the split histogram. Then it uses those
coefficients to calculate the inverse Dual-Dirac CDF for the specified SER(s). It is only
applicable to systems with an exponential impulse response whose time constant is on the
order of the time for one symbol, or less. The algorithm is also significantly slower that
the None extrapolation method.
This table summarizes the different extrapolation methods the block supports.
| Method | Description | Comments |
|---|---|---|
none | Uses previous neighbor interpolation of each eye slice to find the SER values. The interpolated value at any query point is the value at the previous sample grid point. |
|
dualdirac | Fits a Dual-Dirac model to each slice of the eye, then uses the fitted model to find the SER values. |
|
gaussian | Uses the Gaussian defined by the sample mean and sample standard deviation of each eye slice to find the SER values. |
|
linear | Uses linear interpolation of each eye slice to find the SER values. The interpolated value at any query point is based on linear interpolation of the values at neighboring grid points in each respective dimension. |
|
spline | Uses natural spline interpolation of the cumulative sum out from the center of each eye slice to find SER values. |
|
pchip | Uses shape-preserving piecewise cubic interpolation of each eye slice to find the SER values. The interpolated value at any query point is based on a shape-preserving piecewise cubic interpolation of the values at neighboring grid points. |
|
makima | Uses modified Akima cubic Hermite interpolation of each eye slice to find the SER values. The interpolated value at a query point is based on any piecewise function of polynomials with degree at most three. The Akima formula is modified to avoid overshoots. |
|
Version History
Introduced in R2024aSee Also
eyeDiagramSI (Mixed-Signal Blockset) | eyeContour (Mixed-Signal Blockset) | eyeMask (Mixed-Signal Blockset)
Topics
- Choose Extrapolation Method Based on Application (Mixed-Signal Blockset)
