SKA(DS) System Design Aspects: building a system

1
SKA(DS) System Design Aspectsbuilding a system

• Laurens Bakker

2
System model
Distributed beamforming
N/16
N/16
N
N
N
Central beamforming
N
N
N
N
N
N
N
Assumes 2 indep. Beams, 2 pol, FOV 250sq.degree
at 1GHz (both beams)
3
System model
Distributed beamforming
N/16
N/16
N
N
N
Central beamforming
N
N
N
N
N
N
N
Central beamforming closely resembles FPA
architecture
4
Central beamforming

• rather different RF architecture for distributed
  and central
• We cant use the same front-end (cost) in both
  cases
• Central resembles FPA to a certain extend
• Some numbers of FPA at ASTRON (APERTIF)

5
Central beamforming

• rather different RF architecture for distributed
  and central
• We cant use the same front-end (cost) in both
  cases
• Central resembles FPA to a certain extend
• Some numbers of FPA at ASTRON (APERTIF)

6
8 x 7 x 2 elements Vivaldi array Dual
polarisation 112 antenna elements 112
amplifiers 60 cables 60 receivers Frequency range
1.0 1.7 GHz Element separation 10 cm (l/2 _at_
1.5 GHz) 30 MHz bandwidth (backend) Data
recording backend (6.7 s)
APERTIF prototype
APERTIF prototype
7
Some APERTIF numbers

• It took about 2 days to assemble the antenna
• It took about 2 days to connect all amplifiers
  and power
• It took about 2 hours to connect the cables on
  both sides
• The front-end is expensive
• Voltage regulators needed for performance
  (noise)

8
Some APERTIF numbers

• It took about 2 days to assemble the antenna
• It took about 2 days to connect all amplifiers
  and power
• It took about 2 hours to connect the cables on
  both sides
• The front-end is expensive
• For AA
• Ease of deployment important
• Manufacturability important
• In general getting a system operational takes a
  lot of time

9
System temperature

• LNA noise temperature vs. Tsys
• Current APERTIF LNA is 50K (with 15dB gain, 50
  ohm)
• Current installed APERTIF front-end is 65K (40dB
  gain, 50 ohm)
• Current measured Tsys 115K

10
System temperature

• LNA noise temperature vs. Tsys
• Current APERTIF LNA is 50K (with 15dB gain, 50
  ohm)
• Current installed APERTIF front-end is 65K (40dB
  gain, 50 ohm)
• Current measured Tsys 115K
• So Tsys about 65 K higher than LNA
• 15K second stage front-end
• Feed loss and loss connectors 20K (expensive
  RF material used)
• Low cost high performance connectors needed (or
  no connectors at all)
• Noise coupling/mismatch about 10K (LNA has low Rn
  value)
• Sky noise 3K
• (spillover about 15K, not relevant for AA?)

11
System temperature

• LNA noise temperature vs. Tsys
• Current APERTIF LNA is 50K (with 15dB gain, 50
  ohm)
• Current installed APERTIF front-end is 67K (40dB
  gain, 50 ohm)
• Current measured Tsys 115K
• So Tsys about 65 K higher than LNA
• 15K second stage front-end
• Feed loss and loss connectors 20K (expensive
  RF material used)
• Noise coupling/mismatch about 10K (LNA has low Rn
  value)
• Sky noise 3K
• (spillover about 15K, not relevant for AA)
• Quite some challenges ahead achieving Tsys
  numbers of (or even below) 50K as specified at
  low cost

12
Sky noise and survey speed

• Sky noise rather dominant below 500 MHz

Tinst40K Efficiency75 A/T10000
13
Sky noise and survey speed

• Survey speed increases when scaled with 1/?2
• A rather constant survey speed from 300-1GHz can
  be achieved with aperture array

Tinst40K Efficiency75 A/T10000
14
Some SKA system optimization points

• Should optimize SKA system (and cost) as a whole
• What should be the switchover frequency of AA
  -gtdishes
• How many different antenna technologies are
  required to cover the whole band?
• 100-500MHz requires probably 2 different antenna
  types
• 500-800Mhz can easily be met with one antenna
  type
• 300MHz-800MHz (or even 1000MHz) is also
  achievable

15
Some SKA system optimization points

• Should optimize SKA system (and cost) as a whole
• What should be the switchover frequency of AA
  -gtdishes
• How many different antenna technologies are
  required to cover the whole band?
• 100-500MHz requires probably 2 different antenna
  types
• 500-800Mhz can easily be met with one antenna
  type
• 300MHz-800MHz (or even 1000MHz) is also
  achievable
• We should try to minimize the required number of
  different antenna types
• Running cost (esp. power consumption) should be
  taken in account early on in the design process