Process and machine for signal waveform analysis

• US 6,137,283 A
• Filed: 11/08/1999
• Issued: 10/24/2000
• Est. Priority Date: 02/22/1995
• Status: Expired due to Term

• Current Assignee: Michael K. Williams
• Original Assignee: Michael K. Williams
• Inventors: Williams, Michael K., Coffey, Daniel J.
• Primary Examiner(s): Metjahic, Safet
• Assistant Examiner(s): Solis, Jose M.
• Application Number:
  US09/435,707
• Time in Patent Office: 351 Days
• Field of Search: 702/79, 702/80, 702/70, 702/69, 324/76.42, 324/76.28, 324/76.38, 324/76.24, 324/76.15, 324/76.12, 324/121 R
• US Class Current: 324/76.12
• CPC Class Codes: G01R 23/16 Spectrum analysis; Fourier analysis

Abstract

Highly accurate time interval measurement is achieved for an electrical waveform. The electrical waveform is sampled and converted to a series of voltages, and the series of voltages is interpolated in order to form a time tag list, using interpolations that are optimized for time interval measurement and analysis. The time tag list accurately represents the times at which particular events of interest occur, and is used to generate displays and results analysis such as adjacent cycle jitter and accurate differential triggering and analysis.

Claims

1. An apparatus for interpolating a discrete time series that represents voltages of a waveform, to determine a point at which the waveform crosses a threshold voltage, the apparatus comprising:

means for upsampling the discrete time series to create an upsampled discrete time series;

means, coupled to the means for upsampling, for determining a first digital word and a second digital word of the upsampled series of digital words that surround the threshold voltage; and

means, coupled to the means for determining, for performing linear interpolation between the first digital word and the second digital word to determine a time tag corresponding to the point at which the waveform crosses the threshold voltage;

wherein the means for low-pass filtering includes;

a low-pass filter that has a plurality of coefficients;

means for inserting a digital word with a zero value between each adjacent pair of digital words in the series of digital words, to create a zero-valued series of digital words; and

means for low-pass filtering the zero-valued series of digital words with the low-pass filter, to create the upsampled series of digital words.

2. The apparatus of claim 1, wherein the waveform is sampled at a sampling frequency to generate the series of digital words, and wherein the low-pass filter includes:

a 500 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.35 multiplied by the sampling frequency;

means for upsampling the 500 point low-pass filter with respect to itself to generate a 2500 point low-pass filter;

a 1000 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.4 multiplied by the sampling frequency; and

means for upsampling the 2500 point filter with respect to a 1000 point filter to create a 25000 point filter.

3. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform; and

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

wherein the step of selecting an event of interest includes the steps of;

selecting measurement parameters;

specifying threshold information; and

selecting an interpolation filter that is used in the step of interpolating; and

wherein the step of selecting measurement parameters includes the steps of;

selecting an input type to be one of differential measurement and single ended measurement;

selecting at least one channel; and

selecting at least one slope, corresponding to each at least one channel, to be one of a positive polarity and a negative polarity;

and wherein the step of specifying threshold information includes one of;

specifying a threshold voltage;

selecting a predetermined voltage to be one of a standard TTL voltage, a 3.3 Volt TTL voltage, a standard ECL voltage, a positive ECL voltage, a shifted ECL voltage, and a ground voltage; and

selecting multiple threshold voltages of 0.8 volts, 1.5 volts, and 2.0 volts, each of which represents an event of interest.

4. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform; and

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

wherein the step of selecting an event of interest includes the steps of;

selecting measurement parameters;

specifying threshold information; and

selecting an interpolation filter that is used in the step of interpolating; and

wherein the step of selecting measurement parameters includes the steps of;

selecting a parameter to be analyzed to be one of period, pulse width, frequency, and delay;

selecting a presentation format to be one of time series line graph, time series text, histogram, and jitter spectrum; and

selecting an analysis type to be one of basic and deviation.

5. An apparatus for interpolating a discrete time series that represents voltages of a waveform, to determine a point at which the waveform crosses a threshold voltage, the apparatus comprising:

means for upsampling the discrete time series to create an upsampled discrete time series;

means, coupled to the means for upsampling, for determining a first digital word and a second digital word of the upsampled series of digital words that surround the threshold voltage; and

means, coupled to the means for determining, for performing linear interpolation between the first digital word and the second digital word to determine a time tag corresponding to the point at which the waveform crosses the threshold voltage;

wherein the waveform includes a plurality of cycles, each of the plurality of cycles having a rising edge and a falling edge, and wherein the means for upsampling includes;

a filter having characteristics that are optimized to accurately recreate the waveform at each rising edge and each falling edge;

means for inserting a digital word with a zero value between each adjacent pair of digital words in the series of digital words, to create a zero-valued series of digital words; and

means for filtering the zero-valued series of digital words according to the filter to create the upsampled series of digital words.

6. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform; and

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

wherein the step of interpolating includes interpolating according to one of;

a linear interpolation filter; and

a time optimal interpolation filter that is a combination of a non-linear interpolation filter and the linear interpolation filter;

wherein the step of interpolating includes interpolating according to one of;

a linear interpolation filter; and

a time optimal interpolation filter that is a combination of a non-linear interpolation filter and the linear interpolation filter;

wherein the step of selecting an event of interest includes selecting a threshold voltage, wherein the series of digital words correspond to a sampling frequency, and wherein the step of interpolating according to a linear interpolation filter further includes the steps of;

determining a first digital word and a second digital word that surround the threshold voltage;

determining a first difference between the first digital word and the second digital word;

determining a second difference between the first digital word and the threshold;

determining a relative time to be the second difference divided by a multiplication product of the first difference and the sampling frequency; and

determining one of the series of time tags to be equal to a sum of the sample time of the first digital word and the relative time.

7. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform; and

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

wherein the step of interpolating includes interpolating according to one of;

a linear interpolation filter; and

a time optimal interpolation filter that is a combination of a non-linear interpolation filter and the linear interpolation filter;

wherein the step of selecting an event of interest includes selecting a differential measurement, wherein the step of converting includes steps of converting a first waveform into a first series of digital words and converting a second waveform into a second series of digital words, each of the first series of digital words corresponding to one of the second series of digital words, wherein the first and second series of digital words correspond to a sampling frequency, and wherein the step of interpolating according to a linear interpolation filter includes the steps of;

determining a difference between each digital word of the first series and the corresponding digital word of the second series, to generate a series of differences;

determining a first difference and a second difference that surround a value of zero;

determining a third difference between the first difference and the second difference;

determining a relative time to be the first difference divided by a multiplication product of the third difference and the sampling frequency; and

determining the time tag to be equal to a sum the sample time at which the digital words that correspond to the first difference were sampled and the relative time.

8. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform; and

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

wherein the step of interpolating includes interpolating according to one of;

a linear interpolation filter; and

a time optimal interpolation filter that is a combination of a non-linear interpolation filter and the linear interpolation filter;

wherein the step of interpolating according to a time optimal interpolation filter includes the steps of;

upsampling the series of digital words to create an upsampled series of digital words;

determining a first digital word and a second digital word of the upsampled series of digital words that surround the threshold voltage; and

performing linear interpolation between the first digital word and the second digital word to determine a time tag corresponding to the event of interest.

9. The method of claim 8, wherein the waveform includes a plurality of cycles, each of the plurality of cycles having a rising edge and a falling edge, and wherein the step of upsampling includes the steps of:

providing a filter having characteristics that are optimized to accurately recreate the waveform at each rising edge and each falling edge; and

inserting a digital word with a zero value between each adjacent pair of digital words in the series of digital words, to create a zero-valued series of digital words; and

filtering the zero-valued series of digital words according to the filter, to create the upsampled series of digital words.

10. The method of claim 8, wherein the step of upsampling includes the steps of:

providing a low-pass filter that has a plurality of coefficients;

inserting a digital word with a zero value between each adjacent pair of digital words in the series of digital words, to create a zero-valued series of digital words; and

low-pass filtering the zero-valued series of digital words with the low-pass filter, to create the upsampled series of digital words.

11. The method of claim 10, wherein the waveform is sampled at a sampling frequency to generate the series of digital words, and wherein the step of providing a low-pass filter includes the steps of:

providing a 500 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.35 multiplied by the sampling frequency;

upsampling the 500 point low-pass filter with respect to itself to generate a 2500 point low-pass filter;

providing a 1000 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.4 multiplied by the sampling frequency; and

upsampling the 2500 point filter with respect to a 1000 point filter to create a 25000 point filter that is the low-pass filter provided in the step of providing a low-pass filter.

12. The method of claim 10, further including the steps of:

determining statistical properties of the time tag list; and

displaying the time tag list and the statistical properties of the time tag list in one of a basic mode and a deviation mode as one of a line graph, text, a histogram, and a jitter spectrum;

and wherein the step of selecting an event of interest includes steps of;

selecting a parameter to be analyzed to be one of period, pulse width, frequency, and delay; and

wherein the step of selecting an event of interest further includes one of;

specifying a threshold voltage;

selecting a predetermined voltage to be one of a standard TTL voltage, a 3.3 Volt TTL voltage, a standard ECL voltage, a positive ECL voltage, a shifted ECL voltage, and a ground voltage; and

selecting multiple threshold voltages of 0.8 volts, 1.5 volts, and 2.0 volts, each of which represents an event of interest.

13. A method of interpolating a discrete time series that represents voltages of a waveform, to determine a point at which the waveform crosses a threshold voltage, the method comprising the steps of:

upsampling the discrete time series to create an upsampled discrete time series;

determining a first digital word and a second digital word of the upsampled series of digital words that surround the threshold voltage;

performing linear interpolation between the first digital word and the second digital word to determine a time tag corresponding to the point at which the waveform crosses the threshold voltage;

wherein the waveform includes a plurality of cycles, each of the plurality of cycles having a rising edge and a falling edge, and wherein the step of upsampling includes the steps of;

providing a filter having characteristics that are optimized to accurately recreate the waveform at each rising edge and each falling edge; and

inserting a digital word with a zero value between each adjacent pair of digital words in the series of digital words, to create a zero-valued series of digital words; and

filtering the zero-valued series of digital words according to the filter, to create the upsampled series of digital words.

14. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform;

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

determining statistical properties of the time tag list; and

displaying the time tag list and the statistical properties of the time tag list;

wherein the step of displaying includes displaying one of;

a time series line graph that includes occurrences of the event of interest;

a time series line graph that includes an amount that a subsequent occurrence of the event of interest deviated from a previous occurrence of the event of interest;

text representing a time series that includes occurrences of the event of interest;

text representing a time series that includes an amount that a subsequent occurrence of the event of interest deviated from a previous occurrence of the event of interest;

a histogram that includes occurrences of the event of interest; and

a histogram that includes an amount that a subsequent occurrence of the event of interest deviated from a previous occurrence of the event of interest;

a jitter spectrum that includes the frequency components of jitter of the event of interest; and

a jitter spectrum that includes the frequency components of an amount that a jitter for an event of interest deviated from a previous amount of jitter for the event of interest.

15. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform;

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

determining statistical properties of the time tag list; and

displaying the time tag list and the statistical properties of the time tag list;

wherein the step of displaying includes displaying adjacent cycle data.

16. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform;

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

determining statistical properties of the time tag list; and

displaying the time tag list and the statistical properties of the time tag list;

wherein the step of determining statistical properties includes determining at least one of;

a mean;

a standard deviation;

a largest negative slope displacement;

a largest positive slope displacement;

a maximum; and

a minimum.

17. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform;

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

generating locally synthesized data as a second series of digital words, each of the second series of digital words representing a voltage of an ideal waveform and a synthesized sample time corresponding to each voltage; and

interpolating the second series of digital words to generate a second time tag list.

18. A method for analyzing time interval data of a waveform, comprising the steps of:

converting the waveform into a series of digital words, each word of the series of digital words representing a voltage of the waveform and a sample time at which the voltage occurred;

selecting an event of interest defined by a characteristic of the waveform; and

interpolating the series of digital words to generate a time tag list, the time tag list including a series of time tags, each of the series of time tags representing a time at which the event of interest occurred during the waveform;

wherein the step of converting includes steps of converting a first waveform into a first series of digital words and converting a second waveform into a second series of digital words;

the step of interpolating includes interpolating the first series of digital words to create a first time tag list and interpolating the second series of digital words to create a second time tag list;

the method further including the steps of;

determining a first set of statistical properties of the time tag list;

determining a second set of statistical properties of the second time tag list;

providing inter-acquisition analysis by comparing the first set and second set of statistical properties, to generate a third set of statistical properties.

19. A method of interpolating a discrete time series that represents voltages of a waveform, to determine a point at which the waveform crosses a threshold voltage, the method comprising the steps of:

upsampling the discrete time series to create an upsampled discrete time series;

determining a first digital word and a second digital word of the upsampled series of digital words that surround the threshold voltage;

performing linear interpolation between the first digital word and the second digital word to determine a time tag corresponding to the point at which the waveform crosses the threshold voltage;

wherein the step of upsampling includes the steps of;

providing a low-pass filter that has a plurality of coefficients;

inserting a digital word with a zero value between each adjacent pair of digital words in the series of digital words, to create a zero-valued series of digital words; and

low-pass filtering the zero-valued series of digital words with the low-pass filter, to create the upsampled series of digital words.

20. The method of claim 19, wherein the waveform is sampled at a sampling frequency to generate the series of digital words, and wherein the step of providing a low-pass filter includes the steps of:

providing a 500 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.35 multiplied by the sampling frequency;

upsampling the 500 point low-pass filter with respect to itself to generate a 2500 point low-pass filter;

providing a 1000 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.4 multiplied by the sampling frequency; and

upsampling the 2500 point filter with respect to a 1000 point filter to create a 25000 point filter that is the low-pass filter to be provided.

21. An apparatus for analyzing time interval data of a waveform with respect to an event of interest, comprising:

a measurement control unit, that has an input for receiving the waveform and an output providing an input file, the input file including a series of digital words, each of the series of digital words representing a voltage of the waveform and an input time at which the voltage occurred;

a time tag interpolator, having an input coupled to the measurement control unit for receiving the input file, and an output providing a time tag list according to an interpolation, the time tag list including a series of time tags, each of the series of time tags representing an output time at which the event of interest occurred during the waveform; and

a results analyzer, having an input coupled to the time tag interpolator for receiving the time tag list, and an output that provides the series of time tags and statistics of the series of time tags;

wherein the time tag interpolator includes;

a low-pass filter having an input that receives the input file, for generating an upsampled interpolated data stream as an output; and

a linear interpolation filter having an input that receives the input file, for interpolating based on linear relationships between time tags;

a time tag generator, having an input that receives the upsampled interpolated data stream, for comparing the upsampled interpolated data stream with the event of interest to generate the time tag list.

22. The apparatus of claim 21, wherein the waveform has a sampling frequency, and wherein the low-pass filter includes:

a time optimal filter, including;

a 500 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.35 multiplied by the sampling frequency;

means for upsampling the 500 point low-pass filter with respect to itself to generate a 2500 point low-pass filter;

a 1000 point low-pass filter having a pass band to 0.25 multiplied by the sampling frequency and a stop band beginning at 0.4 multiplied by the sampling frequency; and

means for upsampling the 2500 point filter with respect to a 1000 point filter to create a 25000 point filter; and

at least one hybrid filter that is a combination of the linear interpolation filter and the time optimal filter;

wherein the time tag interpolator further includes a filter selector to select one of the time optimal filter and the at least one hybrid filter.

23. The apparatus of claim 21, wherein the waveform includes a plurality of cycles, each of the plurality of cycles having a rising edge and a falling edge, and the low-pass filter is structured and arranged to accurately recreate each rising edge and each falling edge.

24. An apparatus for analyzing time interval data of a waveform with respect to an event of interest, comprising:

a measurement control unit, that has an input for receiving the waveform and an output providing an input file, the input file including a series of digital words, each of the series of digital words representing a voltage of the waveform and an input time at which the voltage occurred;

a time tag interpolator, having an input coupled to the measurement control unit for receiving the input file, and an output providing a time tag list according to an interpolation, the time tag list including a series of time tags, each of the series of time tags representing an output time at which the event of interest occurred during the waveform; and

a results analyzer, having an input coupled to the time tag interpolator for receiving the time tag list, and an output that provides the series of time tags and statistics of the series of time tags, wherein the results analyzer includes;

an event processor having an input that receives the time tag list, for processing the time tag list with respect to the selected event of interest to generate data indicative of the event of interest as an output;

a statistical processor, having an input that receives the data indicative of the event of interest, for performing statistical analysis on the data indicative of the event of interest to generate statistical data as an output; and

a display/output controller, coupled to the statistical processor and the event processor, having an input that receives the data indicative of the event of interest and the statistical data, and having an output that provides display data corresponding to the data indicative of the event of interest and the statistical data.

25. An apparatus for analyzing time interval data of a waveform with respect to an event of interest, comprising:

a measurement control unit, that has an input for receiving the waveform and an output providing an input file, the input file including a series of digital words, each of the series of digital words representing a voltage of the waveform and an input time at which the voltage occurred;

a time tag interpolator, having an input coupled to the measurement control unit for receiving the input file, and an output providing a time tag list according to an interpolation, the time tag list including a series of time tags, each of the series of time tags representing an output time at which the event of interest occurred during the waveform; and

a results analyzer, having an input coupled to the time tag interpolator for receiving the time tag list, and an output that provides the series of time tags and statistics of the series of time tags;

wherein the time results analyzer includes means for displaying adjacent cycle analysis parameters.