The WVR at Effelsberg. Thomas Krichbaum

Transcription

1 The WVR at Effelsberg Alan Roy Ute Teuber Helge Rottmann Thomas Krichbaum Reinhard Keller Dave Graham Walter Alef

2 The Scanning GHz WVR for Effelsberg ν = 18.5 GHz to 26.0 GHz Δν = 900 MHz Channels = 24 T receiver = 200 K sweep period = 6 s rms = 61 mk per channel Features Uncooled (reduce cost) Scanning (fewer parts, better stability) Robust implementation (weather-proof, temperature stabilized) Noise injection for gain stabilization Beam matched to Effelsberg near-field beam TCP/IP communication Web-based data access Improved version of prototype by Alan Rogers

3 The Scanning GHz WVR for Effelsberg

4 The Scanning GHz WVR for Effelsberg Front-end opened March 16th, 2004 Ethernet data acquisition system Temperature regulation modules Control unit

5 WVR Performance Requirements Phase Correction Aim: coherence = 0.9 requires λ/ 20 (0.18 mm rms at λ = 3.4 mm) after correction Need: thermal noise 14 mk in 3 s Measured: 12 mk = 0.05 mm Need: gain stability 3.9 x 10-4 in 300 s Measured: 2.7 x 10-4 Opacity Measurement Aim: correct visibility amplitude to 1 % (1 σ) Need: thermal noise 2.7 K Measured: 12 mk Need: absolute calibration 14 % (1 σ) Measured: 5 %

6 WVR View of Atmospheric Turbulence Absorber Zenith sky (clear blue, dry, cold) 280 Tantenna vs Time 57 Tantenna vs Time 10 2 Allan Variance of Tantenna Antenna temperature / K Absorber, on roof in Effelsberg 03apr to apr UT time / s Antenna temperature / K Zenith, clear sky, on roof in Bonn 02apr to 1205 UT time / s Allan Variance zenith sky Absorber, on roof in Effeslberg 03apr to apr UT absorber time / s 12 h 1 h gain stability: 2.7x10-4 over 400 s sensitivity: 61 mk for τ int = s (0.038 mm rms path length noise for τ int = 3 s)

7 Typical Water Line Spectrum

8 WVR Panorama of Bonn

9 Move to Effelsberg March 20th, 2003

10 WVR Panorama of Effelsberg

11 Spillover Cal: Skydip with Absorber on Dish detector output 0 V to 0.3 V el = 90 to 0 19 to 26 GHz

12 Gain Calibration Measure: hot load sky dip at two elevations noise diode on/off detector output / V Calibration Data from 04feb10 absorber + noise diode absorber 23.6d elevation 41.8d elevation frequency / GHz Derive: Tsky Treceiver gain Tsky at zenith, derived from 04feb10 cal data 300 Treceiver, derived from 04feb10 cal data Gain, derived from 04feb10 cal data Tsky / K 20 Treceiver / K gain / V/K frequency / GHz frequency / GHz frequency / GHz

13 WVR Beamwidth: Drift-Scan on Sun Drift scan through sun Detector voltage UT WVR in Bonn YIG freq = 26.0 GHz FWHM = 1.26d 26.0 GHz beamwidth = Detector voltage UT WVR in Bonn YIG freq = 18.0 GHz 6 FWHM = 1.18d 18.0 GHz beamwidth = Sun position / degrees

14 WVR Beam Overlap Optimization 2000 m Altitude 4.2 Water Vapour Density / g/m3 Water vapour density / kg m^ Altitude / km WVR 100 m RT Beam Overlap for three WV profiles Beam overlap Atmospheric WV Profiles at Essen from Radiosonde launches every 12 h (courtesy Dr. S. Crewell, Uni Cologne) (1) (2) (3) 0 m = low altitude water vapour 2 = mean water vapour profile 3 = high altitude water vapour Half power beamwidth / degrees

15 Scattered Cumulus, 2003 Jul 28, 1300 UT

16 Storm, 2003 Jul 24, 1500 UT

17 First Attempt to Validate Phase Correction 300 WVR Path Correction vs Time, 03apr UT 250 path correction / mm time / s (start = s) 180 WVR Predicted Phase and VLBI Measured Phase, 03apr UT phase at 86 GHz / degrees time / s (start = s)

18 WVR Noise Budget for Phase Correction Thermal noise: 75 mk in the water line strength, April mk per channel on absorber, scaled to 25 channels difference on-line and off-line channels (34 mk in Feb 2004 due to EDAS hardware & software upgrade) Gain changes: 65 mk in 300 s 2.7x10-4 multiplies T sys of 255 K Elevation noise: 230 mk typical elevation pointing jitter is 0.1 sky brightness gradient = 2.8 K/ at el = 30 Beam mismatch: 145 mk measured by chopping with WVR between two sky positions with 4 throw, Aug = 120 m at 1.5 km and el = mk to 145 mk Sramek (1990), VLA structure functions 95 mk Sault (2001), ATCA 2001apr UT Other? Spillover model errors, cloud liquid water removal, RFI, wet dish, wet horn Total (quadrature): 290 mk = 1.3 mm rms

19 Move to Focus Cabin March 16th, 2004

20 WVR Beam Geometry 1500 m Beam overlap, April 2003 Beam overlap, April 2004

21 Optical Alignment using Moon 60 Azimuth Scans across Moon 2004mar30 60 Elevation Scans across Moon 2004mar30 Antenna Temperature / kelvin Antenna Temperature / kelvin K Azimuth Offset / degrees Elevation Offset / degrees T antenna = 23 K T moon = 220 K at 22 GHz Beam filling factor = Beam efficiency = 92 %

22 19 to 26 GHz 19 to 26 GHz Spillover Reduction el = 90 to 0 detector output 0 V to 0.3 V

23 WVR Path Data from 3 mm VLBI, April Path Length and Elevation vs Time, 2004apr Path length / mm 150 Path length / mm path length elevation Elevation time / UT Time / UT hours 0

24 VLBI Path Comparison, 3 mm VLBI, April 2004 VLBI Phase and EB WVR Path Length Comparison, 2004apr DOY: 108 UT: 00:40 to 00:47 Baseline: Pico Veleta Effeslberg Path length / mm 83.4 VLBI phase + constant EB WVR path time / UT

25 VLBI Phase Correction Demo No phase correction EB phase correction 180 Baseline 1-3 Scan 7 NRAO150.UV.CL s s s 180 Baseline 1-3 Scan 7 NRAO150.UV.CL12 VLBI phase WVR phase NRAO 150 Pico Veleta - Effelsberg 86 GHz VLBI 2004 April 17 path 3.4 mm s s s Coherence function before & after 1.0 Baseline 1-3 Scan 7 Coherence Functions EB+PV phase correction 180 Baseline 1-3 Scan 7 NRAO150.UV.CL s 120 s 240 s 360 s s s s 420 s Path rms reduced 1.0 mm to 0.34 mm Coherent SNR rose 2.1 x

26 VLBI Phase Correction Demo 180 Baseline 1-3 Scan 11../NRAO150.UV.CL13 No phase correction Baseline 1-3 Scan 12../NRAO150.UV.CL NRAO 150 Pico Veleta - Effelsberg 86 GHz VLBI 2004 April VLBI phase s s s s s s WVR phase 180 Baseline 1-3 Scan 11../NRAO150.UV.CL15 EB phase correction Baseline 1-3 Scan 12../NRAO150.UV.CL path 3.4 mm s s s s s s 1.0 Baseline 1-3 Scan 11 Coherence Functions Coherence function before../nrao150.uv.cl15 & after../nrao150.uv.cl Baseline 1-3 Scan 12 Coherence Functions 0.8../NRAO150.UV.CL15../NRAO150.UV.CL s 120 s 240 s 360 s s 120 s 240 s 360 s Path rms reduced 0.85 mm to 0.57 mm Coherent SNR rose 1.7 x 100 d Baseline 1-3 Scan 11 Structure Functions 100 d Baseline 1-3 Scan 12 Structure Functions 420 s

27 VLBI Phase Correction Demo Before phase correction at EB 180 Baseline 1-3 Scan 21../NRAO150.UV.CL13 CLOUDED 90 0 NRAO 150 Pico Veleta - Effelsberg 86 GHz VLBI 2004 April Baseline 1-3 Scan 22../NRAO150.UV.CL VLBI phase s s s -180 WVR phase s s s After phase correction at EB 180 Baseline 1-3 Scan 21../NRAO150.UV.CL15 CLOUDED 180 Baseline 1-3 Scan 22../NRAO150.UV.CL path 3.4 mm s s s s s s Coherence function before & after 1.0 Baseline 1-3 Scan 21 Coherence Functions 0.8../NRAO150.UV.CL15../NRAO150.UV.CL Baseline 1-3 Scan 22 Coherence Functions 0.8../NRAO150.UV.CL15../NRAO150.UV.CL s 120 s 240 s 360 s Baseline 1-3 Scan 21 Structure Functions Path rms saturated at 0.95 mm Coherent SNR decrease 7.5 x s 120 s 240 s 360 s 420 s Baseline 1-3 Scan 22 Structure Functions

28 VLBI Phase Correction Demo Coherence function after phase correction at EB divided by CF before phase correction Improvement factor Baseline 1-3 Corrected/Uncorrected Coherence Functions NRAO 150 Pico Veleta - Effelsberg 86 GHz VLBI 2004 April s s 120 s 240 s 360 s 120 s 240 s 360 s Coherent integration time Baseline 1-3 Corrected/Uncorrected Structure Functions Coherence improves for most scans

29 Cloud Removal EB WVR path time series Keep VLBI scan times only Subtract linear rate 200 WVR path length vs time, EB, 2004 Apr WVR path length vs time, EB, 2004 Apr 17 WVR path length vs time, EB, 2004 Apr Kept data during VLBI scans Kept data during VLBI scans Fit and subtracted linear slope from each scan Path / mm Path / mm Path / mm Time / seconds Time / seconds Time / seconds NRAO GHz VLBI 2004 April 17 Cloud contamination shows up as large scatter in the path lengths

30 VLBI Phase Correction Demo

31 VLBI Phase Correction Demo

32 Validation of Opacity Measurement Comparison of Opacity Measured with 100 m RT and WVR, 2004feb m RT RCP 100 m RT LCP opacity WVR UT / hours

33 Path Length & Opacity Statistics at Effelsberg

34 Path Length Stability at Effelsberg RMS path fluctuation over 120 s vs hour of day - July RMS path fluctuation over 120 s vs hour of day - December 2 mm 1 mm 0 mm 0 h 24 h sunrise UT sunset 0 h sunrise UT sunset 24 h

35 Absolute Calibration for Astrometry & Geodesy 120 km

36 Opacity Effects and the Mapping Function

37 Issues: TCP/IP time overhead

38 Issues: Temperature stability Physical temperature near LNA vs time 20 mk 3 min T sys vs time 250 mk

39 Issues: Temperature stability Solution: weaken thermal coupling between Peltier and RF plate Effects: No more 3 min temperature oscillation Worse long-term temperature stability Weak thermal coupling Strong thermal coupling Temperature vs time Temperature vs time 0.7 C 5.5 C 0.75 days 2.5 days

40 Issues: Noise Diode Stability Tsys vs time on absorber Calibrate using temp. Calibrate using noise diode 2.0 K 22 h Tsys rms / K Structure function of Tsys on absorber 1 K Original data Calibrated with noise diode Calibrated with temperature 0.1 K Time / s

41 Issues: Beam Mismatch at Low Elevation?

42 Future Developments Software development: (Helge Rottmann, RadioNet) data paths into JIVE correlator, AIPS and CLASS improve calibration accuracy (allow for opacity effects) Hardware development: temperature stabilization: reduce Tsys? spillover: integration time efficiency: beam overlap: better insulation, regulation Cooling? reduce with new feed? Data acquisition system upgrade move to prime focus receiver boxes?

43 Conclusions WVR running continuously Phase correction of 3 mm VLBI has been demonstrated (but in four experiments WVR made things worse.) Opacities agree with those from 100 m RT Zenith wet delays agree with GPS & radiosonde within 10 mm Web-based display & archive access available Radiometer stability is 2.7 x 10-4 in 400 s Radiometer sensitivity is 61 mk in s integration time