Keith Grainge Calibration issuesAA-low Technical Progress meeting Calibration Issues Keith Grainge

Calibration issuesAA-low Technical Progress meeting Calibration – progress “Status Update on AA Calibration and Imaging Research” document (Wijnholds, Grainge, Nijboer) –Summary of recent work –References to 41 other document Cal Im meeting, July 2012: talks on web Context and some highlights…

Keith Grainge Calibration issuesAA-low Technical Progress meeting What limits the final map sensitivity? –Thermal noise (A/T; integration time) –Confusion noise (source counts; maximum baseline) –Calibration errors Related question of image fidelity –Issues of station and array configuration; uv-plane coverage Key questions: how accurate must calibration be such that it is not limiting performance? How do we achieve this? Limits to Imaging

Keith Grainge Calibration issuesAA-low Technical Progress meeting AA Calibration Aim is overall model for calibration of entire AA system –Investigate calibration requirements at all sub-system levels –Comment on specs; guide design decisions Eventually will be limited by non-linear effects of systematics –Need to consider interaction of various errors Scope: SKA Phase I is initial priority Also take forward Phase II considerations

Keith Grainge Calibration issuesAA-low Technical Progress meeting Approach Experience from existing mature telescopes Lessons from pathfinders – LOFAR, MWA, EMBRACE, AAVS … – Will become increasingly important – Especially for identifying “sub-optimal” design choices Simulations Theoretical studies

Keith Grainge Calibration issuesAA-low Technical Progress meeting L LOFAR calibration to reach the noise limit requires subtraction of several thousand sources along several hundred directions. OFAR calibration to reach the noise limit requires subtraction of several thousand sources along several hundred directions LOFAR calibration- SAGEcal Yatawatta and Kazemi Space Alternating Generalized Expectation Maximization Direction dependent calibration within facets Current LOFAR limits ≈ 1 million in dynamic range Sources outside the FOV almost as important as inside o LOFAR calibration requires subtraction of 1000s sources

Keith Grainge Calibration issuesAA-low Technical Progress meeting Before SAGEcal Cas A, 30 deg. away Center

Keith Grainge Calibration issuesAA-low Technical Progress meeting After SAGEcal Cas A, 30 deg. away Center

Keith Grainge Calibration issuesAA-low Technical Progress meeting (left) original, (middle) normal (right) hybrid SAGECal SAGEcal: ionosphere

Keith Grainge Calibration issuesAA-low Technical Progress meeting “Estimation noise” Self-calibration will be required to reach desired dynamic range Estimation of calibration parameters extracts information from the data –Effective noise in the image increases with number of calibration parameters –Imposes a stability requirement

Keith Grainge Calibration issuesAA-low Technical Progress meeting Imaging issues What is required accuracy in knowledge of station beam? – Translate this into tolerance requirements on physical parameters – Also determines the hierarchical level to which calibration is required – e.g. Individual beam former weights for each element of an AA-mid? How do we measure/assess image fidelity?

Keith Grainge Calibration issuesAA-low Technical Progress meeting Station configuration Bright sources in sidelobes are a major issue How best to assess station beam shapes? –Some useful metrics exist (Zarb-Adami) Need to consider interferometric case Seemingly sensible criteria may not lead to optimum design in practice

Keith Grainge Calibration issuesAA-low Technical Progress meeting BeamSims Smirnov Far Sidelobe Calibration Noise – Currently for dishes – Extend analysis to AAs Assess contribution from sources as a fn of radius; determine cost fn Sources attenuated by –PSF –PB –Time / bandwidth smearing

Keith Grainge Calibration issuesAA-low Technical Progress meeting Calibration with astronomical sources Do astronomical sources provide sufficient SNR for accurate estimation of calibration parameters? Particularly an issue for time varying DDEs A sensitivity issue Timescale Collector area; bandwidth etc May require specific calibration hardware to be built E.g. Ionosphere and implications for filling factor as a function of frequency (Wijnholds)...

Keith Grainge Calibration issuesAA-low Technical Progress meeting Ionospheric calibration implications (Wijnholds, 2012)

Keith Grainge Calibration issuesAA-low Technical Progress meeting Calibration “noise” Smirnov Residual calibration noise DDE “elimination” – Source peeling – SAGEcal – Differential gains Residual calibration noise Artifact assessment simulation –LSM + “contaminators” –G-Jones + E-Jones errors –Run self-cal for G-Jones 4.2 μJy rms residuals after subtraction of two 1 Jy contaminator sources; G-Jones solution

Keith Grainge Calibration issuesAA-low Technical Progress meeting Calibration “noise” 2.6 μJy rms residuals after subtraction of two 1 Jy contaminator sources; dE-Jones solution Improve with direction dependent solution, dE-Jones Far less spatial correlation But for long integrations... – Structure shows up – Artifacts do not decrease as √n

Keith Grainge Calibration issuesAA-low Technical Progress meeting Effects of Coupling AA elements very strongly coupled Use macro basis fns (Craeye) Issues: sparceness; configuration; channel bandwidth

Keith Grainge Calibration issuesAA-low Technical Progress meeting Beam Forming - Array Factor Station beam from Array factor and mean embedded element pattern (Razavi-Ghods, Lera Acedo) Array factor: configuration; convolutional gridding; FFT Interpolation for different directions and frequencies

Keith Grainge Calibration issuesAA-low Technical Progress meeting Xarray

Keith Grainge Calibration issuesAA-low Technical Progress meeting Simulation: OSKAR2 Oxford OeRC: Salvini, Mort and Dulwich Simulates AA visibility measurement sets Takes advantage of compute power of GPUs Very powerful tool for investigating calibration – Extended / polarised sources in sidelobes – System design aspects e.g. element geometry / alignment errors – Element complex gain errors How best to analyse the results?

Keith Grainge Calibration issuesAA-low Technical Progress meeting OSKAR2 - Future Next steps: ionosphere; noise; individual element patterns; dish simulations Increasing maturity of simulations allows: – Simulation / comparison with existing instruments – Currently working on LOFAR simulation o Validate OSKAR2 against BBS snapshot o Investigate possible deficiencies in LOFAR reduction pipeline on runs with changing beam

Keith Grainge Calibration issuesAA-low Technical Progress meeting Other Calibration Issues What is allowable gain/phase error in analogue chain –Encompasses all electronic and temperature effects Absolute amplitude calibration Can calibration be applied in real time? Beam formation –Bandpass changes; filter performance (aliasing) –Sample bit count through processing chain –What levels of frequency / time smoothing can be applied

Keith Grainge Calibration issuesAA-low Technical Progress meeting Summary Calibration critical to realise potential sensitivity Guide AA design decisions at station and system level Will specific calibration hardware be required? Address issues with pathfinders and simulation tested against measurements A great deal of progress in many areas Simulations should be extremely powerful ways of investigating several issues