DOC

ClockJitter 1

By Eleanor Porter,2014-03-24 11:07
13 views 0
ClockJitter 1

    ClockJitter 1.3

    Readme and Release Notes

ClockJitter 1.3

    ;;ClockJitter is a tool that performs jitter analysis on sample clock data in one of the

    following formats:

    o Regularly sampled voltage data.

    o Interval data.

    o Crossover data (array of edge crossing times).

    Details on the specific formats that are supported for each type of input file are described

    in subsequent sections.

    ;;ClockJitter uses a frequency domain algorithm for calculating peak to peak to jitter of

    input clock signals. The following steps outline the algorithmic approach used:

    o Format input data into array of intervals.

    o Perform discrete fourier transform on interval data.

    o Multiply frequency domain data by specified frequency domain jitter transfer

    function.

    o Perform inverse discrete fourier transform on data.

    o Form accumulated data array from inverse FFT data.

    o Calculate peak to peak jitter as the difference between the maximum and

    minimum values in accumulated data array.

    ;;ClockJitter is a Windows application. All testing of ClockJitter is performed under the

    Windows XP operating system. Earlier versions of Windows are not tested or

    supported.

To complete the installation:

    1. Run clockjitter.msi and follow the prompts.

Using ClockJitter

    Getting Started:

    Run the ClockJitter application by selecting ClockJitter from the start menu under Start >>Program Files >> ClockJitter 1.3. When the ClockJitter application runs it shows the main ClockJitter GUI shown in Figure 1:

Figure 1- ClockJitter Main Application GUI

Loading a Data File:

    The fist step to using ClockJitter is to load a data file to be analyzed. The type of data file is selected with the File Type control. The selections are listed as follows:

    ;;Differential: The expected input is an ascii or Tektronix binary .wfm single voltage file

    with differential voltages. To select the file enter its full path or select the browse

    button for a standard file selection dialog box. The files can be with or without

    timestamps. The Time stamps in file checkbox must be correct for an ascii file to be

    parsed correctly.

    ;;Single Ended: The expected input is two ascii or Tektronix binary. Wfm single ended

    voltage files. The full file paths can be entered manually into the Data File and Data

    File2 controls or selected with file section dialogs by selecting the browse controls. The

    files can be with or without timestamps. The time stamps in file checkbox must be

    correct for ascii files to be parsed correctly.

    ;;Interval: The expected input is an ascii file with a list of interval times with one

    interval value per line.

    A sample of the expected format is show below for a PCI Express reference clock:

    9.988888888885945E-9

    9.975000000040507E-9

    9.986111111095174E-9

    9.98055555551363E-9

    1.0008333333421347E-8

    9.981250000015533E-9

    1.0008749999905649E-8

    9.993333333372864E-9

    1.00303030303164E-8

    1.0005113636342954E-8

    The time values are not required to be in scientific notation. The expected resolution of

    the time values is in units of seconds.

    Crossover: The expected input is an ascii file with a list of crossover times with one

    crossover time per line. The time values are not required to be in scientific notation.

    The expected resolution of the time values is in units of seconds.

Selecting a File and Supported File Formats:

    Ascii voltage files can be either tab delimited or comma delimited. These formats can be used with either the single ended or differential file types. ClockJitter assumes that files with the .csv extension are comma delimited ascii files and files with the .tsv extension are tab delimited ascii files. ClockJitter assumes that files with any other extension except .wfm are tab delimited ascii files. ClockJitter assumes that files with the .wfm extension use the Tektronix binary wfm format. The standard file selection dialog is shown in Figure 2:

Figure 2 - File Selection Dialog

    The filetype selection can be used to show all files or to filter on .tsv, .csv, .txt, .dat, or .wfm files. Figure 3 shows the ClockJitter GUI after a single interval file has been selected.

Figure 3 - Selection Of An Interval File

Template Selections

    ClockJitter contains a pull down menus for a template file containing settings to use during the

    test. The ClockJitter template list from the main application screen is shown in Figure 4 with

    the list of technology selections:

Figure 4 - Template Pull Down Menu Selections

Each of the template selections are listed below with a summary of the template settings that

    ClockJitter includes for that technology. Templates set the jitter filter to be used in analyzing

    the data, the expected clock interval, and the jitter budget used in determining pass fail status for

    analyzed data.

;;FBD_133_MHZ(BETA):

    o Clock Interval: 7.5 Nanoseconds

    o Jitter Budget 34.15 Picoseconds

    o Expected Sample Size 1 Million unit intervals

    o Expected Measurement Point System Board FBD Connector

    o Worst Case System PLL 1 (H1)

    ;;33 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;11 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;;Delay = 5 Nanoseconds

    o Minimum CDR (H3)

    ;;.2 Mhz High Pass.

    ;;20 DB/dec rolloff.

    o SSC Separation YES

    ;;FBD_166_MHZ(BETA):

    o Clock Interval: 6.25 Nanoseconds

    o Jitter Budget 27.32 Picoseconds

    o Expected Sample Size 1 Million unit intervals

    o Expected Measurement Point System Board FBD Connector

    o Worst Case System PLL 1 (H1)

    ;;33 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;11 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;;Delay = 5 Nanoseconds

    o Minimum CDR (H3)

    ;;.2 Mhz High Pass.

    ;;20 DB/dec rolloff. o SSC Separation YES

    ;;PCIE_1_1:

    o Clock Interval: 10 Nanoseconds o Jitter Budget 86 Picoseconds o Expected Sample Size 1 Million unit intervals

    o Expected Measurement Point System Board PCIE Connector

    o Worst Case System PLL 1 (H1)

    ;;22 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;1.5 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;;Delay = 10 Nanoseconds

    o Minimum CDR (H3)

    ;;1.5 Mhz High Pass.

    ;;20 DB/dec rolloff.

    o SSC Separation NO

    ;;PCIE_2_0_5MHZ_1_5M_H3_FIRST:

    o Clock Interval: 10 Nanoseconds o Jitter Budget 27.9 Picoseconds o Expected Sample Size 1 Million unit intervals o Expected Measurement Point System Board PCIE Connector

    o Worst Case System PLL 1 (H1)

    ;;16 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;5 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = 1.16

    ;;Delay = 12 Nanoseconds

    o Minimum CDR (H3)

    ;;1.5 Mhz High Pass.

    ;;20 DB/dec rolloff.

    o SSC Separation YES

    ;;PCIE_2_0_5MHZ_1_5M_H3_STEP:

    o Clock Interval: 10 Nanoseconds o Jitter Budget 27.9 Picoseconds o Expected Sample Size 1 Million unit intervals o Expected Measurement Point System Board PCIE Connector

    o Worst Case System PLL 1 (H1)

    ;;16 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;5 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = 1.16

    ;;Delay = 12 Nanoseconds

    o Minimum CDR (H3)

    ;;1.5 Mhz Step. o SSC Separation YES

    ;;PCIE_2_0_8MHZ_1_5M_H3_FIRST:

    o Clock Interval: 10 Nanoseconds o Jitter Budget 27.9 Picoseconds o Expected Sample Size 1 Million unit intervals

    o Expected Measurement Point System Board PCIE Connector

    o Worst Case System PLL 1 (H1)

    ;;16 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;8 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;;Delay = 12 Nanoseconds

    o Minimum CDR (H3)

    ;;1.5 Mhz High Pass.

    ;;20 DB/dec rolloff.

    o SSC Separation YES

    ;;PCIE_2_0_8MHZ_1_5M_H3_STEP:

    o Clock Interval: 10 Nanoseconds o Jitter Budget 27.9 Picoseconds o Expected Sample Size 1 Million unit intervals

    o Expected Measurement Point System Board PCIE Connector

    o Worst Case System PLL 1 (H1)

    ;;16 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;; Delay = 0.

    o Worst Case System PLL 2 (H2)

    ;;8 Mhz Low Pass.

    ;;40 DB/dec Rolloff

    ;;Damp = .54.

    ;;Delay = 12 Nanoseconds

    o Minimum CDR (H3)

    ;;1.5 Mhz Step. o SSC Separation YES

    Note: ClockJitter is packaged for release by several different organizations. Templates

    for all documented technologies are not included with all distributions of ClockJitter.

FBD Templates are based on pre-release versions of the specification and are not final.

Application Settings

    ClockJitter provides a number of settings through the application GUI.

    Time Stamps In File: If this setting is checked and the filetype is a voltage file (single ended or differential) ClockJitter will assume there is a timestamp preceding each data entry in the file and will parse out and ignore these timestamps.

    Number of Header Lines To Skip: ClockJitter will skip this number of lines at the beginning of the file. If the file to be analyzed contains header information before the data, this setting must be used to inform ClockJitter to skip to the data portion of the file. ClockJitter will not automatically attempt to detect and skip header information in the file.

    Sample Interval: The main application GUI contains a control that allows the user to enter the sample interval for the file being tested. This field is required input for ClockJitter to operate accurately when voltage files are analyzed (it is not used for interval or crossover files). ClockJitter does not attempt to parse the sample interval from data file headers.

    Num Intervals: When a file is selected for testing this control is updated with the number of clock intervals that are contained in the selected file.

    Ave Interval: When a file is selected for testing this control is updated with the average interval for the clock intervals contained in the selected file. When the file selected is a voltage time record only the appropriate edges are used in analysis. For example, a voltage time file tested with the PCIE_1_1 template will only use the rising edges of the waveform for clock interval analysis.

Running a Full Test and Viewing Results

    Once data files have been loaded, the correct template selected, and application settings configured as desired, it is time to run a test. A full test run is started by pressing the Test button on the main application GUI. When a full test completes the results screen is automatically displayed as shown in Figure 5:

Figure 5 - Test Results Screen

    Each of the results on the full test results screen are summarized as follows:

    ;;Clock Jitter Test Result: Indicates an overall pass or fail result. The result is fail if the

    peak to peak jitter exceeds the limit specified in the selected template.

    ;;High frequency and low frequency jitter values for each of the following. The low

    frequency values are greyed out for technologies that only specify high frequency jitter

    limits.

    o Peak to Peak Jitter (ps): The difference in the maximum and minimum values

    in the accumulated phase jitter array.

    o RMS Jitter (ps): The root mean square value of all the values in the

    accumulated phase jitter array.

    o Maximum Allowed Jitter (ps): The maximum allowed jitter (per spec) for the

    specified template.

    Command Line Execution And Batch Processing Mode

    Note: Command line mode has not been fully validated for the ClockJitter 1.1 release.

    ClockJitter can be executed from the command line using a variety of command line flags to set

    application options. The command line options support a batch mode that will process an entire

    directory of files without the need for user intervention. Each of the command line options for

    ClockJitter are described as follows:

    ;;/si <sample interval>: The /si switch is used to specify the sample interval in

    picoseconds.

    ;;/ts: The /ts flag indicates that the file(s) to be processed contain time stamps.

    ;;/nots: The /nots flag indicates that the file(s) to be processed do not contain time stamps

    ;;/d <directory>: The /d option specifies a directory to be used for batch processing

    multiple files. The directory can be a relative path from the ClockJitter.exe executable

    or a full path. The /d option is not used if a single file (or pair of single files) is being

    analyzed from the command line.

    ;;/p <positive file>: The /p option specifies a positive single ended file to be analyzed.

    The file can be specified with a relative path from the ClockJitter.exe directory or with an

    absolute path. The /n option must also be used to specify the corresponding negative

    single ended file.

    ;;/n <negative file>: The /n option specifies a negative single ended file to be analyzed.

    The file can be specified with a relative path from the ClockJitter.exe directory or with an

    absolute path The /p option must also be used to specify the corresponding positive

    single ended file.

    ;;/s <differential file>: The /s option specifies a differential file to be analyzed. The

    differential file can be specified with a relative path from the ClockJitter.exe directory or

    with an absolute path. If the /d option is used the differential file can contain wildcards

    to process multiple differential files in the indicated directory.

    ;;/i <interval file>: The /i option specifies an interval file to be analyzed. The interval

    file can be specified with a relative path from the ClockJitter.exe directory or with an

    absolute path. If the /d option is used the interval file can contain wildcards to process

    multiple interval files in the indicated directory.

    ;;/c <crossover file>: The /c option specifies a crossover file to be analyzed. The

    crossover file can be specified with a relative path from the ClockJitter.exe directory or

    with an absolute path. If the /d option is used the crossover file can contain wildcards to

    process multiple crossover files in the indicated directory.

    ;;/l <log file>: When ClockJitter is run from the command line it produces a text output

    file containing a summary of each of the results from the full test run along with a data

    and timestamp for when the test results finished. The output file is named

    clock_jitter_test.tsv by default if the /l flag is not used. When multiple files are

    processed a new line is appended to the output file containing the results for each file

    processed. The default location of the output file if the /l flag is not used is in the

    directory of the data file(s) being processed.

    ;;/t <template file name>: Specifies the template file to be used in command line

    processing. The template file name can be specified as a relative path from the

    ClockJitter.exe directory or as an absolute path. The template name specified does not

    need to be one of the templates that install with the ClockJitter application. The

    templates that install with ClockJitter are located in the templates directory under the

    main ClockJitter program directory.

    ;;/q Causes the application to close once the last file from command line execution is

    processed. By default the application will stay open until it is closed manually.

Command Line Examples

    This section contains a set of examples for processing a single file or multiple files with ClockJitter command line processing mode.

ClockJitter /s c:\Diffout.txt /ts /si 50 /t templates\PCIE_1_1.dat

    Analyzes a differential tab delimited file c:\Diffout.txt with time stamps and a sample interval of 50 picoseconds. The PCIE_1_1.dat PCI Express 1.1 template is used. When the file to be analyzed uses a non-standard extension (such as .txt) ClockJitter assumes it is a tab delimited ascii file.

ClockJitter /p c:\Singlep.csv /nots /n c:\singlen.csv /si 25 /t templates\PCIE_1_1.dat

    Analyzes two single ended files, c:\singlep.csv and c:\singlen.csv, that do not contain time stamps and have a 25 picosecond sample interval. The PCIE_1_1.dat PCI Express 1.1 template is used.

ClockJitter /d c:\sample_data /i *.csv /ts /t /t templates\PCIE_1_1.dat

    Analyzes all *.csv files in the c:\sample_data directory as interval files. The template file PCIE_1_1.dat is used.

Summary of Additions to ClockJitter 1.3 Release

    ;;PCI Express 2.0 templates require a minimum of 90,000 clock periods for official testing.

    Smaller records will still be tested - but a warning will be presented to the user instead of

    a pass/fail result.

    ;;PCI Express 2.0 templates require a sample interval of 50 picoseconds or less for

    waveform data. Data with larger sample intervals will still be tested - but a warning will

    be presented to the user instead of a pass/fail result.

    ;;The beta distinction has been removed from the PCIE 2.0 templates. The templates

    match the final PCI Express 2.0 specification.

    ;;Pre-filtering of voltage records above 5 Ghz is now applied with all templates.

    ;;Tektronix .wfm files support now supports versions 1, 2, and 3 of the .wfm format.

Summary of Additions to ClockJitter 1.2 Release

    ;;PCI Express 2.0 templates test for high frequency (HF) and low frequency (LF) jitter.

    ;;PCI Express 20 templates support testing with a 1.5 Mhz step transfer function.

Summary of Additions to ClockJitter 1.1 Release

    ;;SSC removal is supported for PCI Express 2.0 and FBD analysis.

    ;;When voltage records are analyzed only rising edges are used in calculations for

    appropriate technologies.

    ;;Tektronix binary .wfm voltage files are supported.

    ;;The phase jitter frequency content is plotted before and after filtering and/or SSC

    removal.

    ;;The GUI has been updated to provide the average interval and number of intervals in the

    file under test.

    ;;PCI Express 2.0 templates are provided for performing analysis with or without a

    minimum CDR (“H3”) assumption.

    ;;Memory management has been improved for large voltage records. 64M point records

    have been tested.

    ;;Voltage records are pre-filtered to remove noise above 5 Ghz. This noise is expected to

    be from instrumentation and is unlikely to affect clock receivers.

Known Issues With The ClockJitter 1.0 Release

    ;;SSC removal is not performed for FBD and PCI Express 2.0 analysis. A minimum

    CDR bandwidth assumption is used instead as described in the templates section.

    ;;Creating custom templates is not supported.

    ;;The command line based batch processing mode has not been validated in this release.

    ;;There is no ability to view the frequency content of the input file for debug purposes.

Copyright ? 2005-2006, Intel Corporation. All rights reserved.

Intel assumes no responsibility for any errors in this document. Intel makes no commitment to update nor to keep current the information contained in this document. * Other product and corporate names may be trademarks of other companies and are used only for explanation and to the owners' benefit, without intent to infringe.

Report this document

For any questions or suggestions please email
cust-service@docsford.com