Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Electrical signal parameter measurement system
Reexamination Certificate
2000-08-31
2003-05-27
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Electrical signal parameter measurement system
C702S066000, C702S079000, C702S187000
Reexamination Certificate
active
06571186
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to time stamping of threshold crossing points in an acquired waveform input signal where the time stamps are separated by one or more nominal time intervals and more particularly to a method of dithering a threshold level for making time stamp measurements of the crossing points. Such a method is usable in time interval distribution measurements for minimizing digitization artifacts.
Oscilloscopes are tradition measurement instruments for time interval measurements of an electrical or optical signals, such as pulse width, time period and the like. A threshold level is established and the signal crossing points through the threshold are time stamped, such as by placing cursors at the crossing points. The time interval between the time stamped locations is calculated and displayed as the time interval measurement.
Digital oscilloscopes have essentially replaced traditional analog oscilloscopes for measuring electrical signals. A digital oscilloscope receives an input analog signal and samples the signal at a defined sample rate established by the instrument setting parameters. The sampled signal is quantized by an analog-to-digital converter to discrete levels and the digitized signal samples are stored in memory. A threshold level is established and time marks of the signal crossing points through the threshold are determined by interpolating between signal samples above and below the threshold crossing. The interpolated time marks are used to calculate the time interval between the time marks.
The computer and telecommunications industry uses digital oscilloscopes to generate statistical time interval distribution measurements for verifying the telecommunications signal meets set standards for jitter. A waveform record of the telecommunications signal is acquired and stored. A threshold level is set and time stamps are determined for the threshold crossing points using interpolation of data point or points above and below the crossing points. Time interval values between time stamps are determined and the statistical time interval distribution measurements are performed. A histogram may be made of the time interval values versus the number of occurrences of the each value and displayed as well as a display of the statistical time interval distribution measurements.
A drawback to using digital oscilloscopes for calculating the time interval of a repetitive electrical signal is artifacts caused by the digitization of the electrical signal. Because quantization levels are fixed, sampled data values above and below the nominal threshold take on a limited set of combinations. This results in time marks that fall at a limited set of interpolated locations between sample points. Time interval measurements made with these time marks will result in a time interval variances. A distribution histogram of the time intervals with digitization artifacts appears as discrete spikes in the plot. This may appear to a user as jitter in the incoming electrical signal where as in fact it is caused by the digitization of the signal.
One way of reducing the digitization artifact is to add a random amount of noise to the acquired signal samples by dithering the samples. A random number generator provides random numbers that are applied to each signal sample. The result of offsetting the signal samples with random numbers is statistically spreading the threshold crossing points of the signal. Even though the time interval distribution is spread out in relation to the digital artifact distribution, the statistical time interval distribution measurements provide a more accurate information on time interval and jitter.
The above method for reducing digitization artifacts has the drawback of being computationally expensive and/or slow. Random numbers are added to each signal sample. For a digital oscilloscope sampling systems with high sample rates, this add substantial digital processing overhead that is performed by the processor. A separate processor or digital signal processor (DSP) may be used to perform the dithering operation but such an addition adds cost to the oscilloscope.
What is needed is a method for time stamping threshold crossing points of a input signal having one or more nominal time intervals that has improved processing throughput. The method should not require adding additional hardware to the oscilloscope. The time stamping method is usable in generating statistical time interval distribution measurements and histogram displays having reduced the digitization artifacts.
SUMMARY OF THE INVENTION
Accordingly, the present invention is a method for time stamping an input signal having one or more nominal time intervals that is usable for minimizing digitization artifacts in time interval distribution measurements. The time stamping method randomly dithers the threshold level to vary the threshold crossing points of the input signal. The steps for time stamping includes the steps of establishing a threshold level for measuring a time mark of a threshold crossing point and acquiring digital data samples representative of the input signal prior and subsequent to the threshold crossing. The threshold crossing point is determined by interpolating at least a first data sample above and a first data sample below the threshold crossing point. The time mark for the interpolated threshold crossing point is stored as the time stamp. The threshold level is varied, using threshold offset values, for subsequent threshold crossings of the input signal with digital data samples acquired prior and subsequent to the crossing points interpolated to determine the threshold crossing points. The time marks for the crossing points are stored as the time stamps.
The threshold level establishing includes the further step of generating a distribution threshold values. Preferably, the distribution is a gaussian distribution. The interpolating step includes the further step of linearly interpolating between the first data sample above the threshold crossing point and the first data sample below the threshold crossing point. A higher order interpolator using multiple data samples above and below the threshold crossing point may be implemented, such as by generating a windowed sin(x)/x function.
The time stamping method includes the additional steps of generating time interval values defined as the difference between two time stamps and accumulating the number of occurrences of the time interval values in time bins. Alternately, the time interval values may be generated between time stamps in two separate waveforms or between time stamps in a waveform and a set of reference times. Further steps include generating and displaying a distribution histogram of the number of occurrences of the time interval values in the accumulated time bins. Still further steps include generating and displaying statistical distribution data of the time interval values from the accumulated time bins for mean, standard deviation, and minimum and maximum time intervals.
The time stamping method is incorporated into a method of minimizing digitization artifacts in time interval distribution measurements on an input signal having one or more nominal time intervals. The previously described time stamping method uses a dithered threshold level to generate the time stamps of the input signal threshold crossing points. Linear or higher order interpolation, such as a windowed sin(x)/x function, are two alternative steps for determining the threshold crossing points. The threshold dithering is accomplished by the step of generating a distribution of threshold values. Time interval are generated and the number of occurrences of the time interval values are accumulated in time bins. Statistical distribution data is generated from the time bin accumulated time interval values for mean, standard deviation, and minimum and maximum time interval values. An additional step includes the display of the statistical distribution data for the time interval values. Further steps include gener
Bucher William K.
Hoff Marc S.
Textronix, Inc.
West Jeffrey R.
LandOfFree
Method of waveform time stamping for minimizing digitization... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of waveform time stamping for minimizing digitization..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of waveform time stamping for minimizing digitization... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3015486