The main goal of this project is to analyze the usefulness of different clock models for NTP, contributing for an enhanced performance. The goal is to use filtering techniques to analyze two issues. The first is a parameter that provides an explicit connection between the polling interval and the accuracy of the synchronization process. The second issue is a more sophisticated model for the performance of the clock on the local system and for the network delay. Different concepts (e.g. Allan Variance and it's application to determining the stability) related to clock statistics will also be studied.
Brief description of topics completed till now (After Community Bonding Period till week 2):
Reviewed the articles that have been completed in community bonding period
The concepts of the frequency Domain and time Domain
Detailed study and work on these topics:
Analysis of Time Domain data
Study of method for calculated the range of “true” Allan Variance from the calculated estimate for a specific tau values
Study of Spectrum Analysis and translation between spectral Density of frequency and Allan Variance
Study of modified Allan Variance estimate for enhancing the knowledge of Allan Variance.
Calculation of range of true Allan Variance from the estimated Allan Variance on the data series provided by Mentor and plotting the relation between Τ (tau) values and degree of freedom for each data series.
Plotting the relation between Τ (tau) values on x-axis and estimated Allan variance, lower range of true Allan Variance, upper range of true Allan Variance on the Y-axis as calculated by the method studied in 2 (completed in starting of week 2).
Preparing to work on the problem “Estimating the crystal oscillator drift in motherboard using temperature sensors” (started discussion)
worked on collecting the temprature data of the CPU and individual cores and written script for getting the information using lm_sensors and written parser program for extracting the exact temprature value.
Program for calculating the standard fourier transform of the a given data series. This will be used in analysis in cross spectrum method.
Preparing for collect the real time data of time difference of local clock and ntp server using ntpdate. The data will be collected for at least two week continuously (the basic time interval of 64 seconds in difference data as well as in temprature data ). Analysis (Allan Variance estimaet analysis and Cross Spectrum Analysis) will be performed on this data when the collection period (two week) will be completed.
Starting to study the cross spectrum (cepstrum) Method which will be usiful in finding the correlation between the temperature and time difference data described in the above points.
Real Time Data collection have been started which will be used in analysis in later stage.
studying topics : Ensemble correlation functions, Properties of Correlation Functions, Tim Average, Time Correlation functions, Power Spectral Density , Cross Spectral Densities (will complete in 1-2 Days).
Still collection of data is in progress (till 4 July) and studying the topics specified Above.
Reference Articles for the Above study:
Properties of Signal Sources and Measurement Methods by D.A. Howe, D.W. Allan and J.A. Barnes
Frequency and Time- Their Measurement and Characterization By Samuel R. Stein
Introduction of time and frequency Metrology
Characterization of Frequency Stability: Analysis of the Modified Allan Variance and Properties of its Estimate
Hi Sir, I have updated the week 2 report in task section.
-- DurgeshPandey - 2015-06-14
Week 3 Updated in task section (point 12 and 13).
-- DurgeshPandey - 2015-06-20
Week 4 Updated in Task section (Just point 14) as point 13 is still in progress.
-- DurgeshPandey - 2015-06-26
I | Attachment | Action | Size | Date | Who | Comment |
---|---|---|---|---|---|---|
docx | Week_1.docx | manage | 5.5 K | 2015-06-06 - 13:18 | DurgeshPandey |