Roshene McCool SKADS Workshop 2007 DS3 – T1 DS3 – T1 Network Infrastructure and Data Network Infrastructure and Data Transmission Transmission Roshene McCool Simon Garrington University of Manchester
Jan 08, 2018
Roshene McCoolSKADS Workshop 2007
DS3 – T1 DS3 – T1 Network Infrastructure and Data TransmissionNetwork Infrastructure and Data Transmission
Roshene McCoolSimon Garrington
University of Manchester
Roshene McCoolSKADS Workshop 2007
DS3 – T1 Network Infrastructure DS3 – T1 Network Infrastructure and Data Transmissionand Data Transmission
U.o.M Objectives1. Development of a prototype phase transfer system over
installed fibre links.2. Cost model analysis for all aspects of the data network
including installation and equipment costs.
Roshene McCoolSKADS Workshop 2007
Transfer of Local Oscillator Transfer of Local Oscillator Signals Over FibreSignals Over Fibre
• Motivation– Accurate timing signals to all antenna & data processing stations – Like to have transmission over fibre to avoid RFI
• Specifications– Not specified directly, yet…– Working to ±1 ps over 1 minute (driven by astronomy at high
frequencies) and ±10 ps over 10 minutes
• Other LO over fibre systems– EVLA, 22 km– ATCA, 4.5 km
• LO over fibre for the SKA & e-MERLIN– Same specifications, similar distances– e-MERLIN, 400km (120km longest unrepeatered hop)
Roshene McCoolSKADS Workshop 2007
Adapting a working design for the SKA Adapting a working design for the SKA MERLIN LBL MERLIN LBL
• The MERLIN telescope, based at Jodrell Bank, achieves coherent operation using a frequency standard transmitted at 1486.3MHz. This system is locally known as the ‘L-band Link’ (LBL).
• The LBL measures the round trip delay to a distant antenna and corrects for half this delay in the correlator.
• The system assumes reciprocity between the go and return path delays.
• The current MERLIN LBL works to 1 ps over periods of hours
Roshene McCoolSKADS Workshop 2007
Adapting a working design for the SKA Adapting a working design for the SKA LO Over FibreLO Over Fibre
• Plan to use our existing equipment, simply transmit the signal optically, rather over microwave links.
• Observe if the fibre link is reciprocal over distances up to 120 km
• Show the current system is no worse than the existing system
• Show that the system is accurate to 1 ps over 1 minute and 10 ps over 10 minutes
Roshene McCoolSKADS Workshop 2007
Experimental set upExperimental set upMaster L-Band Link
HP 8508A Vector Voltmeter
Aerial switchSlave L-band Link
499.9MHz over 3m of RG214
SMF28 Fibre Link (variable length)
1486.3 MHz signal
1486.3 MHz signal over 30m of RG214
Master Laser
Master Receiver
Slave Receiver
Slave Laser
-23dB
-40dB
Lab path
Link path
499.9 MHz over 30m of RG214
Roshene McCoolSKADS Workshop 2007
Results so far …..Results so far …..
Roshene McCoolSKADS Workshop 2007
Phase transfer - conclusionsPhase transfer - conclusions
• The stability of the link within a 1 minute and 10 minute timescale is within ±5 ps.
• This means the LO over fibre is in spec on the 10 minute timescale.
• The experimental error does not allow measurements down to 1 ps over 1 minute and therefore we need further work to obtain measurements at that accuracy.
• The 110 km fibre link works as well as a link with no fibre at all in both cases.
Roshene McCoolSKADS Workshop 2007
SKADs Costing ExerciseSKADs Costing Exercise
• Asked to look at costing for the communications sections of the SKADs design.– Assumed 8.6 Tbits/s per station.– Assumed the use of lasers mounted on custom built DSP boards
and transmission over a private fibre network. – Not restricted to the use of commercial protocols and
equipment.– Used detailed cable installation & test cost analysis
commissioned for the ALMA project (€ 45/km)– Civils costs included separately
Roshene McCoolSKADS Workshop 2007
SKADs Costing Exercise - SKADs Costing Exercise - ConclusionsConclusions
• Costing model shows communication costs are estimated at €155 Million, 8% of the total SKADs design. – (Much greater if links > 480km are included)
• For the costing model data links were restricted to 480km. – Further work is required to assess the best way to build links at
longer distances.– Illustrated the need for compromises on bandwidths at long
distances.
• Next steps– More optimisation of design solution required– Maintain a watching brief on technology solutions.
Roshene McCoolSKADS Workshop 2007
QuestionsQuestions