WP2‐030.050.010‐SSDD‐003 Revision : 1 PHASED ARRAY FEED SIGNAL TRANSPORT AND NETWORKS CONCEPT DESCRIPTION FOR THE SKA Document number .............................................................. WP2‐030.050.010‐SSDD‐003 Revision ........................................................................................................................... 1 Author ..................................................................................................................... CSIRO Date ................................................................................................................. 2011‐06‐13 Status ............................................................................................... Approved for release Name Designation Affiliation Date Signature Submitted by: S. Amy CSIRO 2011‐06‐13 Accepted by: Approved by: R.McCool SPDO 2011‐06‐13
12
Embed
PHASED ARRAY FEED SIGNAL TRANSPORT AND ......2011/06/28 · WP2‐030.050.010‐SSDD‐003 Revision : 1 PHASED ARRAY FEED SIGNAL TRANSPORT AND NETWORKS CONCEPT DESCRIPTION FOR THE
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
WP2‐030.050.010‐SSDD‐003 Revision : 1
PHASED ARRAY FEED SIGNAL TRANSPORT AND NETWORKS CONCEPT DESCRIPTION FOR THE SKA
Document number .............................................................. WP2‐030.050.010‐SSDD‐003Revision ........................................................................................................................... 1Author ..................................................................................................................... CSIRODate ................................................................................................................. 2011‐06‐13Status ............................................................................................... Approved for release
Name Designation Affiliation Date Signature
Submitted by:
S. Amy CSIRO 2011‐06‐13
Accepted by:
Approved by:
R.McCool SPDO 2011‐06‐13
Revision
Error! Reference source not found.
DOCUMENT HISTORY Revision Date Of Issue Engineering Change
Number
Comments
A 2011‐03‐17 ‐ First draft release
B 2011‐06‐13 Sent to SPDO for publication
C 2011‐06‐13 Signed, approved and number added
DOCUMENT SOFTWARE Package Version Filename
Wordprocessor MsWord Word 2003
Block diagrams
Other
ORGANISATION DETAILS Name
Physical/Postal Address
Fax. Website
Revision
TABLE OF CONTENTS
1 INTRODUCTION ............................................................................................. 5 1.1 Purpose of the document ...................................................................................................... 5 1.2 Scope of the document ......................................................................................................... 5
2 SYSTEM CONTEXT FOR PHASED ARRAY FEEDS ....................................................... 5 2.1 SKA Hierarchy ........................................................................................................................ 5 2.2 Role of PAFs in the SKA .......................................................................................................... 7 2.3 First Draft Interface Description ............................................................................................ 8
4 FIRST DRAFT DESCRIPTION OF THE PAF DESIGN .................................................... 8
5 TECHNICAL PROGRESS TO DATE ......................................................................... 9
6 FIRST DRAFT POWER REQUIREMENTS ................................................................ 9
7 FIRST DRAFT RELIABILITY CONSIDERATIONS ....................................................... 10
8 IMPACT OF EXTENSIBILITY TO AIP AND SKA2 .................................................... 10
9 FIRST DRAFT COST REPORT ........................................................................... 10 9.1 Description of the procurement method ............................................................................ 11 9.2 Sensitivity Analysis of Cost estimates .................................................................................. 11
10 FIRST DRAFT RISK ANALYSIS ....................................................................... 11
11 PLANNING FOR THE NEXT PHASE ................................................................... 11 11.1 Identification of gaps and areas of further work ................................................................. 11 11.2 Strategy to proceed to the next phase ................................................................................ 11
Figure 1: Wide‐Band Direct Sampling PAF Receiver Assembly in the SKA Sub‐System Hierarchy ......... 6 Figure 2: RFoF PAF Receiver Assembly in the SKA Sub‐System Hierarchy .............................................. 6 Figure 3: Integrated IQ Mixer PAF Receiver Assembly in the SKA Sub‐System Hierarchy ...................... 7 Figure 4: PAF Concept Design Options .................................................................................................... 7
LIST OF TABLES
Table 1: Approxmate Data Rates and FIbre Count ................................................................................. 8 Table 2: Common Transceiver Power Consumption and Operating Temperature Range ................... 10 Table 3: Cost Comparison of 10Gbit/s Ethernet Transceivers .............................................................. 10
SKA ............................... Square Kilometre Array
SPDO ............................ SKA Program Development Office
STaN............................ . Signal Transport and Networks
SKA1............................ . SKA Phase 1
SKA2............................ . SKA Phase 2
Error! Reference source not found.
Revision
1 Introduction
1.1 Purpose of the document
This document outlines the data transport and network design concepts for Phased Array Feeds (PAFs). Much of the information used in this document is covered in detail in the supporting documentation developed by the PAFSKA group and will be presented in detail at the Dish CoDR to be held in July 2011. Many aspects of these designs are under active development by the PAFSKA partners, including CSIRO. Some of the information presented here, has been extracted from draft documents under development by the PAFSKA group. This document takes a high‐level snapshot view of the data transport requirements as at June 2011.
1.2 Scope of the document
The data and signal transport requirements for three possible PAF receiver designs:
• wide‐band direct sampling, • RF over Fibre (RFoF), • integrated IQ mixers.
The context within the STaN domain is restricted to the provision of the physical network infrastructure (e.g. fibre cabling, termination) and the digital data transport. In the case of RFoF, it is envisaged that STaN would have an advisory role but as the electronics would need to be highly integrated into the PAF design, the PAF group would take a lead role in this aspect.
2 System Context for Phased Array Feeds
2.1 SKA Hierarchy
This concept is applicable for where PAF receivers are implemented as part of the SKA dish feed array. The SKA sub‐system hierarchy for three PAF receiver designs are presented in the figures below.
Error! Reference source not found.
Revision
Figure 1: Wide‐Band Direct Sampling PAF Receiver Assembly in the SKA Sub‐System Hierarchy
Figure 2: RFoF PAF Receiver Assembly in the SKA Sub‐System Hierarchy
Error! Reference source not found.
Revision
Figure 3: Integrated IQ Mixer PAF Receiver Assembly in the SKA Sub‐System Hierarchy
2.2 Role of PAFs in the SKA
The PAF receiver assembly includes the necessary analogue and digital signal processing to present digital data streams to the correlator that represent a number of PAF beam‐formed beams. As noted above, there are a number of PAF Receiver Assembly concepts which require different data transport requirements. The relationship between the various PAF elements are summarised in Figure 4.
Figure 4: PAF Concept Design Options
Error! Reference source not found.
Revision
2.3 First Draft Interface Description
The interface of various PAF Concept design with the STaN domain remains uncertain as development is underway on multiple different designs. In some cases, multi‐mode or single mode fibre can be used. It is strongly recommended that only standard single‐mode fibre be utilised as this is likely to provide the most flexible data transport physical layer regardless of the technology that is used. It is certainly the case that wider‐bandwidth and longer distance transmission can be achieved on single‐mode fibre if COTS networking hardware (e.g. transceivers) are used. There is an additional advantage in that a smaller range of spare components, such as patch leads, feed‐through connectors, fibre pigtails etc need to be stocked along with less confusion for technical staff as to whether a particular connection uses multi‐ or single‐mode fibre.
3 Target Requirements and Functionality
3.1 Functional Requirements
For the three concept design cases, the data rates are summarised in the table below (for a PAF with approximately 200 elements).
Table 1: Approxmate Data Rates and FIbre Count Data Rate Fibre Count @ 10Gbit/s Direct‐Sample 3GS/s @ 8‐12bit ~30Gbit/s 3 x 200 IQ Mixer 2 x 1.6GS/s @ 8bit <30Gbit/s 3 x 200 RFoF 200 elements
The digital transmission component assumes the use of 10Gbit/s transceivers which are currently (in 2011) the “sweet‐spot” in terms of cost, flexibility and performance. In addition, no passive multiplexing has been used. It is important to note that, particularly for the RFoF case, the fibre must be buried at a suitable depth to minimise temperature related phase‐shifts. Whilst this technology would preclude the use of other fibre installation methods (e.g. overhead suspension systems) buried fibre will be needed for the distribution of phase stable local oscillator reference signal(s).
4 First Draft Description of the PAF Design
For the two digital based systems, 10Gbit/s transmission has been used as a basis. As the growth of the Internet increases, the demands for high‐bandwidth, per‐channel systems will continue to drive the cost of 10Gbit/s downwards and also lead to cheaper, higher‐bandwidth per‐channel systems. Using fixed wavelengh DWDM transceivers in the optical C‐band (centered around 1550nm) would realise 32 channels (100GHz spacing) that could be passively MUXed, significantly reducing the fibre core count. A recent development has been the introduction of a fully‐tunable DWDM transceiver that can be tuned across the full C‐band at 50GHz spacing realising 80 channels on a single fibre pair.
Error! Reference source not found.
Revision
As of mid‐2011, 40Gbit/s and 100Gbit/s per wavelength long‐haul transmission systems are available from various vendors as commercially supported products. However, these are unsuitable for this application which distances are typically < 10km from receptor to processor. There is active development of 40Gbit/s and 100Gbit/s ”pluggable” transceiver modules, again driven by the networking and telecommunications sector. Significantly, these use built‐in multplexiing/de‐multiplexiing to realise the total bandwidth spread over a number (typically four) lower bit‐rate channels. An example of 100Gbit/s transceiver development is demonstrated by the commercial availability of a 100Gbit/s C Form‐factor Pluggable (CFP) module which is an industry standard. A CFP module capable of 100Gbit/s over 10km uses four wavelengths in the 1310nm band on single‐mode fibre.
5 Technical progress to date
Development work on a suitable RFoF system is currently being undertaken at CASS as part of the ASKAP project. Whilst RFoF systems have been around for many years, the relatively in‐expensive systems used Vertical Cavity Surface Emitting Lasers operating which have been unsuitable for the required distance and signal to noise. Recent work has concentrated on using Direct Feedback Single Mode Lasers at 1310nm. Early results appear encouraging with the required SNR, phase‐stability and distance being achieved. However, this would require a single fibre for each analogue link along with associated termination and cable routing issues. Although, further development is required, leveraging industry standards such as Coarse Wavelength Division Multiplexing (CWDM) would result in a significant reduction in fibre count. The current COTS systems used for 1Gbit/s networking applications can support eight wavelengths, each separated by 20nm over a single fibre pair (although in this application only uni‐directional transmission is required). More recent passive multiplexing systems has seen the introduction of an eight wavelength passive system using standard DWDM transceivers at 10Gbit/s.
6 First Draft Power Requirements
This is a difficult parameter to quantify at this early stage of the concept design. Indicative power consumption and operating temperature range for a given transceiver type is detailed in the following table.
Error! Reference source not found.
Revision
Table 2: Common Transceiver Power Consumption and Operating Temperature Range Maximum Power
Consumption (W) Operating Temperature
Range (°C) 10Gbit/s LR transceiver (SFP+) 1 0 – 70 10Gbit/s DWDM transceiver (XFP) 3.5 0 – 70 100Gbit/s LR transceiver (CFP) 24 0 – 70
7 First Draft Reliability Considerations
Using industry standard fibre along with COTS components would provide similar reliability to that achieved in the telecommunications transmission systems. Typically these would provide well in excess of 100000 hours MTBF.
8 Impact of Extensibility to AIP and SKA2
The impact of using PAFs on STaN is across two broad areas:
• the quantity (and type) of fibre required, and • the selected transmission technology.
Regardless of the use of PAFs in either phase of the SKA or the AIP, these factors remain essentially constant with a multiplier required depending on the number of PAFs deployed (assuming that the distance from a PAF equipped antenna to data processor is no greater than 10km).
9 First Draft Cost Report
It is difficult if not impossible to undertake a full cost analysis at this early stage of the development of PAF‐based receiver technology for the SKA. This is particularly true given the number of different designs currently under consideration. However, it is worth looking at the cost of 10Gbit/s COTS Ethernet transceivers at mid‐2011 prices (in AUD). In the table below, the cost is given for vendor‐specific transceivers in small quantities (i.e. well under 100 per order). A comparison of the cost when purchased “branded” transceivers directly from a networking vendor (at a discount level that is realistic) and third‐party 100% compatible devices are shown.
Given the potential quantities involved, a direct approach to manufacturers rather than vendors would be prudent. Clearly, having the design finalised and quantities known will provide leverage to negotiate with suppliers to obtain a price‐point the project can sustain. Taking into account exchange rates, import duties and the like will be necessary.
9.2 Sensitivity Analysis of Cost estimates
The discounts used in the above table are based on current levels that are provided to either CSIRO or AARNet for 1‐off/small number quantities. It is likely the cost of 10Gbit/s transceivers will fall in the coming years as this technology begins to replace existing 1Gbit/s campus and enterprise backbones.
10 First Draft Risk Analysis A formal risk analysis is difficult to complete at this concept design stage, particularly given the range of technologies being deployed. Some areas that contain significant risk are:
• RFoF: stability of the fibre over long distances even for the direct‐buried case taking into account both short and long‐term temperature variations. This is less of a concern for phase‐stable LO distribution given it is likely a round‐trip system is likely to be employed as a mitigation strategy.
• The quantity of fibre required along with terminations and management of the cable plant is likely to be a significant cost.
• Using COTS hardware appears to be the correct approach but the quantity of transceivers, related power and cooling needs to be carefully considered.
11 Planning for the next phase
11.1 Identification of gaps and areas of further work
Given there are at least three PAF concept designs which differ in their data transfer requirements it is too uncertain to attempt to complete the costing model or risk analysis. These will become clearer once further development work is undertaken with the PAFSKA group. Further development work, which is underway, is needed before the RFoF system design is finalised.
11.2 Strategy to proceed to the next phase
Significant further design work along with laboratory and field studies are required. The path forward in terms of the receiver technology may be clearer after the upcoming Dish CoDR to be held in July. Field studies of RFoF over different distances with the fibre buried to different depths would also be helpful to determine the stability issues associated with this technology.
Error! Reference source not found.
Revision
Error! Reference source not found.
12 References The following draft documents, in preparation for the Dish CoDR, were used in the preparation of this report.
• Concept Description PAF receiver with wide‐band direct sampling (19 May 2011), • Concept Description PAF receiver with integrated IQ mixers (06 June 2011), • Concept Description PAF receiver with RFoF (06 June 2011).