Holdover in Synchronization Applications
Introduction
“Synchronization is as important as power at the cell site.” [1]
The quote above suggests that we can think of Holdover as analogous to running on backup power.
Modern wireless communication systems require at least knowledge of frequency and often knowledge of phase as well in order to work correctly. Base stations need to know what time it is, and they usually get this knowledge from the outside world somehow (from a GPS Time and Frequency receiver, or from a synchronization source somewhere in the network they are connected to).
But if the connection to the reference is lost then the base station will be on its own to establish what time it is. The base station needs a way to establish accurate frequency and phase (to know what time it is) using internal (or local) resources, and that’s where the function of Holdover becomes important.
The Importance of GPS Derived Timing
GPS as a source of timing is a key component in not just Synchronization in telecommunications but to critical infrastructure in general [2]. Of the 18 CIKR sectors, 15 use GPS derived timing to function correctly [3]
A key application for GPS derived timing is to provide synchronization in wireless basestations and in these applications Holdover is used in base stations to ensure continued operation while GPS is unavailable and to reduce the costs associated with repairs [4]
Base stations depend on timing to operate correctly, particularly for the handoff that occurs when a user moves from one cell to another.[5]
1us for base-stations [6]
Search for reliable timing based on something other than $3k Rubidium spurred on by the need for CDMA to compete. [7]
Holdover is important for E911 for example. Most critical telecom applications require precise time and frequency and synchronization to operate properly such as VoIP, video streaming, TDM (time division multiplex) services, voice switching, mobile services, and any LBS (___location based services, with E911 being the most important).[8]
1us for synchrophasors [9]
How GPS Derived Timing Can Fail
GPS is sensitive to jamming etc because the signal levels are so low [10]
GPS outage not initially an issue because clocks can go into Holdover [11]
GPS suffers from interference which can be alleviated by a GPSDO, up to the stability of the GPSDO[12]
Defining Holdover
Clock Accuracy in MIL-PRF-55310[13]
Time Error Model in ITU G.810[14]
Definition of Holdover[15]
Definition of a Disciplined Oscillator[16]
Holdover Performance Aspects
Two independent clocks once synchronized will walk away from one another without limit. How fast this happens depends on the quality of the oscillator [17]
Holdover relies on the OCXO, the PLL design, and correction mechanisms[18]
Aging and temp stability are taken to be the dominant factors. During Holdover the maximum error due to the OCXO is limited by control mechanisms. [19]
Algorithms and quartz get good results[20]
Allan variance can measure instabilities
Time feviation can measure instabilities
Implementing Holdover Solutions
GPS Clocks are used
It is the TFS, not the GPS Receiver alone, that should be considered as the building block for timing, frequency, and time-of-day services.
GPS TFS built to provide time derived from GPS
Usually built around a GPSDO
How a GPSDO works
GPS clock block diagram
The holdover capability is provided by either by a free running local oscillator, or a local oscillator that is steered with software that retains knowledge of its past performance.
An addition of a Microprocessor can improve temperature stability and aging
Aging can be effectively compensated for
Basic aim of a control mechanism is to improve the stability of a clock or oscillator while minimizing the number of times it needs calibration
In Holdover the learned behaviour of the OCXO is used to anticipate and correct for future behavior
Holdover problem solved by predicting current errors from past history
Prediction allows the system to remain stable in holdover
All sort of choices for algorithms and techniques to do this correction extrapolation, interpolation, predictive filters, including the Kalman filters
Kalman filters are used to generate correction signals
A high quality quartz oscillator matched with a Kalman filter algorithm seems to provide the best compromise between quality and reliability versus cost.
The only limitation to extended hold-over performance in such a GPDO is irregularity in drift rate.
Stability definitions have been around for a long time
Random Walk in an oscillator mostly from outside
Complexity in trying to implement has resulted in tailor made Holdover solutions in the market
References
- ^ http://www.juniper.net/us/en/local/pdf/whitepapers/2000400-en.pdf
- ^ http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA503921
- ^ http://www.swpc.noaa.gov/sww/SWW_2011_Presentations/Wed_830/GPS-PNTTimingStudy-SpaceWeather4-27.pptx
- ^ http://www.eetimes.com/design/communications-design/4213947/Understanding-the-concepts-of-synchronization-and-holdover
- ^ http://www.gmat.unsw.edu.au/snap/publications/khan&dempster2007b.pdf
- ^ http://www.telecom-sync.com/pdf/2008/Day1/WCDMA_and_LTE_Synchronisation_Aspects_(Stefano_Rufini,_Ericsson).pdf
- ^ http://www.4timing.com/SyncGPS.pdf
- ^ http://www.eetimes.com/design/communications-design/4213947/Understanding-the-concepts-of-synchronization-and-holdover
- ^ http://tf.nist.gov/general/pdf/2193.pdf
- ^ tf.nist.gov/sim/2010_Seminar/SIM_2010_GPS_Lombardi.ppt
- ^ http://www.syncuniversity.org/drsync/q45.php
- ^ http://www.gmat.unsw.edu.au/snap/publications/khan&dempster2007b.pdf
- ^ http://standards.gsfc.nasa.gov/reviews/mil/mil-prf-55310d/mil-prf-55310d.pdf
- ^ http://www.itu.int/rec/T-REC-G.810-199608-I
- ^ http://www.etsi.org/deliver/etsi_i_ets/300400_300499/30046201/01_60/ets_30046201e01p.pdf
- ^ http://tf.nist.gov/general/enc-d.htm
- ^ http://tf.nist.gov/general/pdf/988.pdf
- ^ http://www.analog.com/static/imported-files/application_notes/AN-1002.pdf
- ^ http://kunz-pc.sce.carleton.ca/thesis/CrystalOscillators.pdf
- ^ http://tf.nist.gov/general/pdf/2297.pdf