Eye Measurement

Calculate metrics from eye diagram

Since R2024a

• 

Libraries:
Mixed-Signal Blockset / Utilities
SerDes Toolbox / Utilities

Description

The Eye Measurement block generates eye pattern from time-domain waveform data and calculates metrics from the generated pattern.

Examples

Measure Eye Openings in Simulink

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 Systems and Export IBIS-AMI Models.

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 on

• Set Modulation to Modulation

• Set Enable clock input port to on

• Set Amplitude range to [-0.2, 0.2]

• Set Hold off time to IgnoreBits and its unit to UI

• Set Symbol time (s) to SymbolTime

• Set 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.

In the Eye Mask tab:

• Set Eye mask type to Rectangular

• Set W1 (UI) to 0.2

• Set H to 0.1

In the Metric Setup tab:

• Set Store results in base workspace to on

• Set 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"        {1x1 struct}
    "Best Eye Height"    <missing>    "NaN"        1e-05       "None"        {1x1 struct}
    "Eye Mask Margin"    "NaN"        "NaN"        1e-05       "None"        {1x1 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 =

    0.0062    0.0373

Measure Eye Openings from Intermediate Signals in SerDes System Model

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:

1. The input to the receiver

2. The output of the CTLE

3. 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 Systems and Export IBIS-AMI Models.

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 off

• Set Modulation to Modulation

• Set Symbol thresholds to 0

• Set Enable clock input port to off

• Set Phase offset to phaseOffset and its unit to UI

• Set Amplitude range to [-0.2, 0.2]

• Set Hold off time to IgnoreBits and its unit to UI

• Set Symbol time (s) to SymbolTime

• Set 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);

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.

s = warning("off", "shared_msblks_measurement:eye:BlockPathWarningWrapper");
sim(model);
warning(s);

Limitations

• Eye Measurement block does not support rapid accelerator mode.

Ports

Input

data — Input data

scalar

Input data used to construct the eye pattern, specified as a scalar.

Data Types: double

clock — Sampling clock for input data

scalar

Sampling clock signal for input data, specified as a scalar. Windows for the eye pattern are centered on the clock edges.

If you are using clock times at the clock port, new clock times must be within 1 UI of the current simulation time. Clock times should occur in the center of the corresponding eye opening.

Dependencies

To enable this port, select the Enable clock input port in the Configuration tab.

Data Types: double

thresholds — Symbol thresholds used to separate symbols

vector | matrix

Symbol threshold(s), used to separate symbols when building the eye diagram. If you provide this information, it increases the robustness of the PAMn metrics.

If the data at the thresholds port is a vector, the same vector is used across all time points during measurements. If the data at the thresholds port is a matrix, the block uses one vector for each time point.

Dependencies

To enable this port, select the Enable threshold input port in the Configuration tab.

Data Types: double

Parameters

To edit block parameters interactively, use the Property Inspector. From the Simulink^® Toolstrip, on the Simulation tab, in the Prepare gallery, select Property Inspector.

Configuration

Enable threshold input port — Enable threshold input port

on (default) | off

Select to enable the threshold input port. By default, this port is enabled.

Modulation — Number of symbol levels in eye diagram

2 (default) | nonnegative integer scalar

Number of symbol levels expected in the eye diagram, specified as a nonnegative integer scalar.

Programmatic Use

Block parameter: Modulation

Type: character vector

Values: nonnegative integer scalar

Default: 2

Enable clock input port — Enable clock input port

on (default) | off

Select to enable the clock input port. By default, this port is enabled.

Clock type — Determine how to sample waveform clock signal

Rising (default) | Falling | Both | Times

Determine how to sample waveform clock signals:

• Rising — Sample the data waveform at the rising edge of the clock waveform.

• Falling — Sample the data waveform at the falling edge of the clock waveform.

• Both — Sample the data waveform at both rising and falling edges of the clock waveform.

• Times — Sample the data waveform at the time values specified by the clock signal.

Programmatic Use

Block parameter: ClockType

Type: character vector

Values: Rising | Falling | Both | Times

Default: Rising

Clock threshold — Edge detection threshold for waveform clocks

0 (default) | real scalar

Edge detection threshold for waveform clocks, specified as a real scalar.

Dependencies

To enable this parameter, set the Enable clock input port option as on and set the Clock type as Rising, Falling, or Both.

Programmatic Use

Block parameter: ClockThreshold

Type: character vector

Values: real scalar

Default: 0

Clock mode — Represent how data is captured by recovered clock

Clocked (default) | Ideal | Convolved

Represent how data is captured by the recovered clock from the system perspective.

• Clocked — The data is captured at the input to the decision latch using the clock from the output of the clock recovery circuit.

• Ideal — The data and clock are captured using an ideal reference clock source.

• Convolved — The data eye and clock PDF captured in normal mode are convolved together to present an eye and clock that look as though the simulation had been run in clocked mode.

For more information, see Clock Modes (Signal Integrity Toolbox).

Note

If you set the Enable clock input port option as off, Clock mode is set to Ideal and you cannot change it.

Programmatic Use

Block parameter: ClockMode

Type: character vector

Values: Clocked | Ideal | Convolved

Default: Clocked

Phase offset — Delay between clock edge and timing origin

0 (default) | nonnegative real scalar

Delay between the clock edge and the timing origin, specified as a nonnegative real scalar in seconds, samples, or UI.

Note

Phase offset delay must be causal.

Programmatic Use

Block parameter: PhaseOffset

Type: character vector

Values: nonnegative real scalar

Default: 0

Symbols per diagram — Width of eye diagram in symbols

2 (default) | scalar

The width of the eye diagram in symbols, specified as a scalar.

Programmatic Use

Block parameter: SymbolsperDiagram

Type: character vector

Values: scalar

Default: 2

Amplitude range — Minimum and maximum eye amplitudes in eye diagram

[] (default) | two-element vector

Minimum and maximum amplitudes contained in the eye diagram, specified as a two-element vector.

Programmatic Use

Block parameter: AmplitudeRange

Type: character vector

Values: two-element vector

Default: []

Amplitude bins — Number of amplitude bins in eye diagram

257 (default) | positive integer scalar

Number of amplitude bins in the eye diagram, specified as a positive integer scalar.

Programmatic Use

Block parameter: AmplitudeBins

Type: character vector

Values: positive integer scalar

Default: 257

Time bins — Number of time bins in eye diagram

256 (default) | positive integer scalar

Number of time bins in the eye diagram, specified as a positive integer scalar.

Programmatic Use

Block parameter: TimeBins

Type: character vector

Values: positive integer scalar

Default: 256

Hold off time — Delay before measurement analysis

0 (default) | nonnegative real scalar

Delays measurement analysis to avoid corruption by transients, specified as a nonnegative real scalar in seconds, samples, or UI.

Programmatic Use

Block parameter: HoldOffTime

Type: character vector

Values: nonnegative real scalar

Default: 0

Symbol time — Time for one symbol

16e-9 (default) | positive real scalar

Time for one symbol, specified as a positive real scalar.

Note

Symbol time must be evenly divisible by Sample time.

Programmatic Use

Block parameter: SymbolTime

Type: character vector

Values: positive real scalar

Default: 16e-9

Sample time — Time for one sample

[1e-9, 0] (default) | two-element vector | scalar

Time for one sample, specified as a positive real scalar.

If Sample time is specified as a 2- element vector, it represents the fixed-step discrete sample time in Simulink.

If Sample time is specified as a scalar, it represents the sample interval used in the SerDes Toolbox™.

Note

Symbol time must be evenly divisible by Sample time.

Programmatic Use

Block parameter: SampleTime

Type: character vector

Values: two-element vector | scalar

Default: [1e-9, 0]

Metric Setup

Add Metric — Add new metric to calculate

button

Click to add new eye diagram metric to calculate. You can define which metric to calculate at what point, the symbol error rate, and the extrapolation method.

Remove Selected Metrics — Remove selected metrics

button

Click to remove selected eye diagram metrics to focus on the area of interest.

Store results in base workspace — Store metric results to base workspace

off (default) | on

Select to store the metric calculation result to the base workspace at the end of simulation.

Metrics variable name — Name of metrics variable that stores measurement results

ans (default) | string

Name of the metrics variable that stores measurement results in the base workspace.

Programmatic Use

Block parameter: MetricsVariableName

Type: character vector

Values: string

Default: ans

Export to file — Export simulation results to file

button

Click to export the simulation results to a file to save for later use.

Export to base workspace — Export simulation results to base workspace

button

Click to export the simulation results to the base workspace.

Eye Mask

Eye mask type — Shape of eye mask

Off (default) | Rectangle | Diamond | Hexagon | Custom

Shape of the eye mask for the mask margin measurement. For more information, see eyeMask (Mixed-Signal Blockset).

W1 (UI) — Width of rectangle, diamond, and center of hexagon eye masks

0.2 (default) | positive real scalar

The width of the rectangle or diamond eye masks or the width at the center of the hexagon eye mask, specified as a positive real scalar.

Dependencies

To enable this parameter, set Eye mask type to Rectangle, Diamond, or Hexagon.

Programmatic Use

Block parameter: W1

Type: character vector

Values: positive real scalar

Default: 0.2

W2 (UI) — Width of top and bottom of hexagon eye mask

0.1 (default) | positive real scalar

The width at the top and bottom of the hexagon eye mask, specified as a positive real scalar.

Dependencies

To enable this parameter, set Eye mask type to Hexagon.

Programmatic Use

Block parameter: W2

Type: character vector

Values: positive real scalar

Default: 0.1

H — Height of rectangle, diamond, or hexagon eye masks

0.1 (default) | positive real scalar

The height of the rectangle, diamond or hexagon eye masks, specified as a positive real scalar.

Dependencies

To enable this parameter, set Eye mask type to Rectangle, Diamond, or Hexagon.

Programmatic Use

Block parameter: H

Type: character vector

Values: positive real scalar

Default: 0.1

Position mask in eye — Position of origin of shape of mask in eye diagram

At clock (default) | For best height margin | For best width margin

Define the position of origin of the shape of the mask in the eye diagram. You can position it at the clock, or for the best eye height or width.

Dependencies

To enable this parameter, set Eye mask type to Rectangle, Diamond, or Hexagon.

Programmatic Use

Block parameter: PositionMaskInEye

Type: character vector

Values: At clock | For best height margin | For best width margin

Default: At clock

Custom eye mask — Define custom eye mask from eye mask object

[] (default) | object

Define a custom eye mask from an eye mask object.

Dependencies

To enable this parameter, set Eye mask type to Custom.

Programmatic Use

Block parameter: CustomEyeMask

Type: character vector

Values: object

Default: []

Margin SER — Symbol error rate to measure eye mask margin

1e-5 | nonnegative real scalar in the range [0,1]

Symbol error rate at which the block measures the eye mask margin, specified as a nonnegative real scalar in the range [0,1].

Dependencies

To enable this parameter, set Eye mask type to Rectangle, Diamond, Hexagon, or Custom.

Programmatic Use

Block parameter: MarginSER

Type: character vector

Values: nonnegative real scalar in the range [0,1]

Default: 1e-5

Margin extrapolation — Extrapolation method to generate eye contours

None (default) | DualDirac

Extrapolation method to generate eye contours used for measuring eye margins.

Dependencies

To enable this parameter, set Eye mask type to Rectangle, Diamond, Hexagon, or Custom.

Programmatic Use

Block parameter: MarginExtrapolation

Type: character vector

Values: None | DualDirac

Default:

Mask variable name — Name of mask variable to export mask

ans (default) | string

Name of the mask variable to export the mask to the base workspace.

Programmatic Use

Block parameter: MaskVariableName

Type: character vector

Values: string

Default: ans

Export mask to file — Export mask to file

button

Click to export the mask to a file to save for later use.

Export mask to base workspace — Export mask to base workspace

button

Click to export the mask to the base workspace.

Plot Setup

Show plots at end of simulation — Automatically show plots at end of simulation

on (default) | off

Select to automatically show the simulation results as plots at the end of simulation. You can define which type of plot to generate, including the eye, clock, and bathtub.

Plot grouping — define how plots are grouped together

Subplot on same figure (default) | One figure per plot

Define how the plots are grouped together.

Add Plot — Add new plot to show

button

Click to add new plot to show.

Remove Selected Plots — Remove selected plots

button

Click to remove selected plots to focus on the area of interest.

Show Plots — Show plots

button

Click to show the plots.

More About

Extrapolation

Extrapolation method is used to constitute the relation between specific SER values and the shape of the 2-D histogram. It is the core technology behind the generation of bathtub plots and eye contours.

During extrapolation method, the data is pre-processed, usually one symbol at a time on only a 1D slice of the said symbol.

The None extrapolation method is 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.

The Dual-Dirac 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.

Version History

Introduced in R2024a