12 GHz Radio-Holographic surface measurement of the RRI 10.4 m telescope. Ramesh Balasubramanyam , Suresh Venkatesh and Sharath B. Raju † Department of Astronomy & Astrophysics, Raman Research Institute, INDIA. †
†
email-id: {ramesh,sureshv}@rri.res.in
1. Abstract
The millimeter-wave telescope at the Raman Research Institute has a 10.4m paraboloidal primary. The telescope is being fitted with a modern Q-band LNA frontend to support observations in the 40-50 GHz frequency range. In order to assess the goodness of the surface for this purpose, we measured the deviations of the primary surface from an ideal paraboloid by the method of radio holography. For this purpose, we used the 11.6996 GHz beacon from the GSAT3 satellite, a 1.2 m reference antenna, commercial Ku-band LNBCs as the receiver frontends and a Stanford Research Systems (SRS) lock-in amplifier as the backend. Though the LNBCs had independent free-running first local oscillators, we recovered the correlation by using a radiatively injected common tone that served as the second local oscillator. At the end, we mapped the surface deviations on a 64×64 grid and measured an rms surface deviation of ~350µm with a measurement accuracy of ~50µm. In this poster, we report the details of this experiment, analyse the data, discuss the results and present our conclusions.
3. Experimental Setup •
•
5. Results
Basic Principle of Radio Holography: Measure the complex beam pattern; its Fourier Transform yields the aperture plane field (APF) distribution; the surface deviations can be derived from the phases of APF. Signal to Noise Ratio v/s Measurement Accuracy: For Vectorial Holography, to measure the surface deviation to an accuracy of δ µm over N×N pixels, the required SNR is proportional to N/δ (Scott & Ryle, 1977). => δ ≤ 80µm demands SNR ≥ 1600 at 25 mm (Ku-band) • Receiver Built: Both test (10.4m) and reference (1.2m) antennae were fitted with satellite TV lownoise block converters (LNBCs) as frontends. A 11.315 GHz reference tone was radiatively fed to both the antennae for use as second LO to remove the effects of the free-running first LOs in the LNBCs. A tracking PLL improves the SNR by locking to the signal from the reference antenna down-converted and filtered to 57.5 MHz. On both chains, signals down-converted to below 100 kHz and fed to the test (10.4m) and reference (1.2m) inputs of the SRS lockin amplifier (analog correlator). The correlation amplitudes and phases were digitized with a 12 bit ADC and recorded in the control computer every 100ms.
2. Need for Surface Measurements
Fig 5: Beam Power Pattern in logarithmic scale.
Fig 4: Aperture Plane Amplitude Distribution.
• Software developed to automatically: Raster-scan region of interest. Acquire scan data (2 Position Encoders, ADC, DAC) in coordination. Correct for satellite drift by pointing at regular intervals. Acquire calibration data at regular intervals. Fig2: Dual Channel Holography Receiver Layout. Fig 7: Difference Map showing measurement accuracy.
Fig 6: Measured Surface Deviations.
4. Observations
●
Source: 11.6996 GHz beacon from EDUSAT (GSAT3 geostationary satellite, Orbital slot: 74°E longitude, translates to El: 74.40°& Az: 195.75°)
●
Fig1: 10.4m Radio Telescope inside RRI campus.
RRI mm-wave Leighton telescope, of 10.4 m diameter with 81 hexagonal panels, is being rejuvenated to undertake Q-band observations. The immediate key science project: A sensitive unbiased survey for 43 GHz SiO masers towards the Southern Galactic bulge. 2 Ruze's relation: η = η exp(4 π δ / λ ) aperture 0 => δ ≤ 370µm (λ/16) for η7mm ≥ 50%
To confirm this rms error and to identify the panels that may need correction, radio holography was performed during Aug - Sep in 2007.
On the fly mapping; Map extent of ± 4.3° in Az & El about the satellite position.
Effect of free running LO in LNBC solved by radiatively injecting a tone.
Effect of satellite drifting removed, pointing every 30 min during observation.
●
Critical Sampling @ 11.7 GHz = 8.8 arc min.
●
10% Oversampling (8 arcmin spacing); N = 64
●
Measured Beam-width = 9.6 arcmin.
●
●
Fig3: Scheme for Raster Scan
6. Conclusions
Spatial Resolution: D/N = 16 cm (~1/6th the panel size) One Observation: 7- 8 hours which includes 65 Az scans, 11 satellite pointings & 8 to 9 repeats of central block of 5 scans.
●
4 independent observations lead to same results.
●
All observations were done during night time.
Fig 8: Histogram of the Measured Surface Deviations.
SNR > 5000 achieved by making the receiver chain robust, repeating and co-adding the central block scans & using satellite pointing data for amplitude and phase calibration. Fig6, shows that some of the panels require correction. Fig7, is the residual obtained by subtracting two independent surface deviation measurements and rms of this residual is ~70µm. Measurement accuracy of ~50µm achieved using commercial low cost LNBC frontends. Surface error rms is ~350µm and also within λ/16, Q-Band observations possible.
Acknowledgements Thanks to RAL for the components used in the receiver chain. Also, thanks to Prof. V. Lakshminarayanan (LC Lab) for lending the SRS Lockin Amplifier.