Lectures on radio astronomy 3 Receivers and noise

  • Slides: 55
Download presentation
Lectures on radio astronomy: 3 Receivers and noise Richard Strom ASTRON, University of Amsterdam

Lectures on radio astronomy: 3 Receivers and noise Richard Strom ASTRON, University of Amsterdam & Qiannan Normal College for Nationalities 1

Signal we want (from sky) has unfortunate properties: Its statistical characteristics are just those

Signal we want (from sky) has unfortunate properties: Its statistical characteristics are just those of noise, indistinguishable from other background noise (from the ground, atmosphere, receiver, etc. ) n It is usually much weaker than other sources (ground, receiver, etc. ) Detecting it requires top technology, and a lot of clever tricks also n 2

An example of noise 3

An example of noise 3

Noise from solar pulse (spectrum) 4

Noise from solar pulse (spectrum) 4

Telescope receiver system: essential elements – Sensitivity n Tunable frequency n Total (instantaneous) bandwidth

Telescope receiver system: essential elements – Sensitivity n Tunable frequency n Total (instantaneous) bandwidth n Frequency channels/resolution n Stability of output signal n Dynamic range Also useful are simplicity, ease of operation & maintenance, flexibility n 5

Receiver sensitivity issues Choose low-noise components – nowadays often FET/HEMTs; can be cooled (even

Receiver sensitivity issues Choose low-noise components – nowadays often FET/HEMTs; can be cooled (even to < 4 K) for less thermal noise n Minimize effect of lossy components (cables, connectors, atmosphere, filters): loss of -0. 1 d. B (= 2. 3%, or factor 0. 977) will add 2. 3% × 290 K = 7 K to system noise. So, short cables, cooled filters, etc. before amplifier n 6

Troposphere significant for λ < 10 cm; solutions? Get above troposphere (mm telescopes are

Troposphere significant for λ < 10 cm; solutions? Get above troposphere (mm telescopes are on mountains); go into space? n Observe at high elevation. Atmospheric effect depends on thickness. Looking at zenith angle z > 0 (z = 0° is straight up) increases thickness (and effect) by sec z n Observe from dry site – effect is mainly due to water vapor in troposphere n Observe short λ when weather’s good n 7

This shows Tsys as a function of elevation (atmosphere) 8

This shows Tsys as a function of elevation (atmosphere) 8

Telescopes like 15 m SEST: high (2500 m) and dry site 9

Telescopes like 15 m SEST: high (2500 m) and dry site 9

Frequency requirements of radio astronomy receivers n Should be able to (quickly) tune to

Frequency requirements of radio astronomy receivers n Should be able to (quickly) tune to any frequency band n Within that band, may want to have many channels over some range (line observations) n For continuum, maximum possible bandwidth gives greatest sensitivity (includes more signal) 10

Basic modern receiver (VLBA) n n Cryogenic systems are housed in a cryostat, often

Basic modern receiver (VLBA) n n Cryogenic systems are housed in a cryostat, often with an inner cold level (He) at 4 -20 K Even if not cooled, the electronics should be in a temperature-stable box (held constant to ~ ± 1 C)

Structure of a typical superheterodyne receiver

Structure of a typical superheterodyne receiver

Receiver – basic blocks n n n n Now let’s look at the main

Receiver – basic blocks n n n n Now let’s look at the main elements in more detail; they are: Horn/polarizer LNA Filter + amplifier Calibration system Mixer LO system (not shown in detail)

Mixer used for frequency conversion of signal: Frf �Fif 14

Mixer used for frequency conversion of signal: Frf �Fif 14

Mathematics of the superheterodyne system Basic idea is to multiply together signals of two

Mathematics of the superheterodyne system Basic idea is to multiply together signals of two different frequencies n From the trigonometric identity: sinf 1 sinf 2 = ½cos(f 1 – f 2) – ½cos(f 1 + f 2) n The result of multiplying two sinusoidal signals together is signals at the sum and difference frequencies; the latter gives us an intermediate frequency: f. IF = |fsky – f. LO| n 15

Features of the superheterodyne system A local oscillator (LO) signal is mixed with sky

Features of the superheterodyne system A local oscillator (LO) signal is mixed with sky signal: converts sky to intermediate frequency (f. IF) n In general, 2 sky frequencies (upper and lower sidebands) are produced; may not be desirable, so filter one out if not wanted n The same IF system can be used for different antenna signals n 16

Scanning filter receiver (ancient) n n n Use bandwidth of desired resolution Tune over

Scanning filter receiver (ancient) n n n Use bandwidth of desired resolution Tune over expected frequency range OK, but inefficient 17

Filter bank backend (old) 18

Filter bank backend (old) 18

Digital autocorrelation spectrometer 19

Digital autocorrelation spectrometer 19

FT 20

FT 20

Two related issues of t �f Giant pulses from PSR 0531+21 n n n

Two related issues of t �f Giant pulses from PSR 0531+21 n n n The Crab pulsar emits giant pulses of S > 1000 Jy lasting 1 ns Since (1 ns)-1 = 1 GHz such pulses can only be resolved with a receiver BW 1 GHz And the emission itself has to be at least 1 GHz wide 21

A mystery of first years in Westerbork (WSRT) n n n WSRT had twelve

A mystery of first years in Westerbork (WSRT) n n n WSRT had twelve 25 m dishes, all the same All cables were of equal length, carefully measured But there were differences of several meters in delay BW = 4 MHz; (4 MHz)-1 = 25 μsec; c = 75 m FE BWs likely differ a bit 22

Available noise power R at T R’ The signal from resistor R, at temperature

Available noise power R at T R’ The signal from resistor R, at temperature T, is filtered and limited to a bandwidth, B (= f 2 – f 1) n The resulting signal in resistor R’ has a power, Pn = Bk. T n 24

In a similar way, we can determine equivalent noise R’ In a similar setup,

In a similar way, we can determine equivalent noise R’ In a similar setup, we filter the signal from an unknown source n From the power in resistor R’, we can calculate: Ts = Pn/k. B n 25

What is the equivalent noise of the amplifier? 26

What is the equivalent noise of the amplifier? 26

HFET noise temperature 27

HFET noise temperature 27

HFET (HEMT) Low Noise Amplifier (LNA) 28

HFET (HEMT) Low Noise Amplifier (LNA) 28

What is the noise contribution of amplifiers in series? 29

What is the noise contribution of amplifiers in series? 29

Example of receiver temperature calculation Receiver system, 3 amplifiers: 1, 2 & 3 n

Example of receiver temperature calculation Receiver system, 3 amplifiers: 1, 2 & 3 n T 1 = 50 K, G 1 = 20 d. B (= 100×) n T 2 = 300 K, G 2 = 10 d. B (= 10×) n T 3 = 500 K n TN = 50 K + 300 K/100 + 500 K/(10× 100) = 50 K + 3 K + 0. 5 K = 53. 5 K (divide each Tn by product of all gains before it) So we see, 1 st amplifier has main effect. 30

Noise contribution of input loss 31

Noise contribution of input loss 31

The lesson, when designing a receiver, is. . . Put a low noise amplifier

The lesson, when designing a receiver, is. . . Put a low noise amplifier (LNA) with high gain (G ≥ 20 d. B) at front n Avoid any losses before the LNA (so, keep cables short, or use waveguide; avoid filters if possible, or cool them; keep atmospheric loss low – choose a good site) n Noise from lossy element after LNA gets divided by LNA gain (G 1) n 32

Final noise determined by TN, bandwidth, duration 33

Final noise determined by TN, bandwidth, duration 33

Here is an example of time integration in practice n n Observation of pulse

Here is an example of time integration in practice n n Observation of pulse from ms pulsar PSR 1937+21 Integration time increases by 2×each step from bottom See pulse better, lose some detail [frequency smoothing gives similar result] 34

Having sensitivity is useless if stability is poor n Amplifiers with high gain tend

Having sensitivity is useless if stability is poor n Amplifiers with high gain tend to be less stable n To keep output stable, often add feedback loop: automatic gain control (AGC) n Physicist Robert Dicke invented technique: switch to reference noise source, to monitor receiver. 35

Example of a simple Dicke switch radio telescope Generate switching frequency, faster than system

Example of a simple Dicke switch radio telescope Generate switching frequency, faster than system drift n Demodulate at same frequency after detection n Disadvantage is not all time spent on source: lose some observing time n 36

Avoid loss of observing time with two receivers Always observing sky and reference n

Avoid loss of observing time with two receivers Always observing sky and reference n At end, average two difference signals n Always need stable reference n This system costs more (2 channels) n 37

Dicke’s technique widely used, in different ways For example, with two receivers, we can

Dicke’s technique widely used, in different ways For example, with two receivers, we can make two beams n We can point one beam at source, other on empty sky. n Using Dicke’s switch, one beam becomes reference – can “switch out” effect of atmosphere. n 38

Effelsberg λ 2. 8 cm system (Emerson et al. , 1979) 39

Effelsberg λ 2. 8 cm system (Emerson et al. , 1979) 39

What dual-beam measures & example of data (in fog) 40

What dual-beam measures & example of data (in fog) 40

Observation of strong source 3 C 84: data & result NEGATIVE 41

Observation of strong source 3 C 84: data & result NEGATIVE 41

Technique can also be used for mapping extended sources n n n For Effelsberg

Technique can also be used for mapping extended sources n n n For Effelsberg dish (100 m diameter) observing at λ = 2. 8 cm Rayleigh distance: DR ≈ D 2/λ = 1002/0. 028 = 360 km Troposphere (where water is) is at 2 -3 km altitude, so should be same in both beams 42

Single-beam map of 3 C 10, showing effects of atmosphere 43

Single-beam map of 3 C 10, showing effects of atmosphere 43

Cas A, beam separation = 8. 2’ arc: 2 images well separated NEGATIVE 44

Cas A, beam separation = 8. 2’ arc: 2 images well separated NEGATIVE 44

Images not always separated: 3 C 10, 5. 5’ arc beam distance 45

Images not always separated: 3 C 10, 5. 5’ arc beam distance 45

3 C 10, final map separates and averages two images 46

3 C 10, final map separates and averages two images 46

Triple-horn system: 3 beams are even better 47

Triple-horn system: 3 beams are even better 47

Two final sensitivity issues: polarization and confusion n If we only observe with one

Two final sensitivity issues: polarization and confusion n If we only observe with one antenna and receiver, we miss half the signal (for unpolarized sources). Must also have a second antenna with opposite sense of polarization Second antenna must have its own receiver system (so it costs more, but gives better sensitivity by √ 2) �Why is sensitivity better by √ 2? 48

Antenna with one polarization misses half the signal 49

Antenna with one polarization misses half the signal 49

Horn/polarizer n n n Horn determines the dish illumination (hence dish efficiency), spillover and

Horn/polarizer n n n Horn determines the dish illumination (hence dish efficiency), spillover and polarization properties Shown here, hybrid mode prime focus feed designed for high efficiency and low spillover Horn may be mounted on OMT to separate the two polarization channels

Orthomode transducer (OMT) n n φ n n Used to separate two orthogonal polarizations

Orthomode transducer (OMT) n n φ n n Used to separate two orthogonal polarizations Can be 2 linear or (after phase shift φ) 2 circular polarizations Pure linear is easier to realize, but circular is desirable for use in VLBI Examples of OMTs, which can be large, heavy, and difficult to design & build

Confusion is where sources (almost) overlap each other 52

Confusion is where sources (almost) overlap each other 52

To overcome confusion requires a smaller beam A common definition of “confusion level” is

To overcome confusion requires a smaller beam A common definition of “confusion level” is 1 source/20 beams n Source density increases as we go to weaker sources n To estimate, need source number vs. strength curves (from observations); see next panel n A more sensitive telescope needs greater angular resolution n 53

Curves showing source density vs. S (flux density) 54

Curves showing source density vs. S (flux density) 54

Hubble Deep Field (HDF), observed by WSRT & VLA 55

Hubble Deep Field (HDF), observed by WSRT & VLA 55

Next lecture we will look at interferometers 56

Next lecture we will look at interferometers 56