on the call: Maurizio, Brendan, Karl, Shaul, Kris

• Maurizio's post: calibration_simulations
  • another toast bug fixed: galactic/ecliptic mixup
  • note about plot that shows 1/f noise before and after destriper. Blue curve has the dipole removed, but orange curve still has the dipole remaining.
  • some questions about why orange peak of dipole in Fourier domain is low. Maurizio will try to make the effects of destriping clearer.
  • The gain variation plot shows the punchline.
    • first point is that there is very small bias around 5e-4. This might be a discrepancy between DaCapo and toast dipole computation.
    • 0.04% RMS error in gain with a 10-hour window. more than factor 2 better than Planck, though of course no Galaxy in this sim.
    • noise level correlates with seasonal dipole changes as expected.
  • Next step: the gain rms error will go back into toast sim to produce I,Q,U frequency maps. These gain estimates will then show the EE and BB due to leakage from temperature to polarization.
  • Shaul: eventually final parameters (noise level, scan parameters, etc.) will be in place.
    • how to best use parameters: stop Maurizio's work until final parameters are in place, or re-run later? Maurizio would like to keep pushing ahead. Once bugs are fixed, compute time is pretty small. Can be re-run when final parameters are settled.
    • What is the time scale? Sensitivities are settled. Scan angle might change just a little. Biggest remaining uncertainty: different between Current Best Estimate noise and Requirement noise. aiming for late may or possibly into June.
  • Shaul: how much can this infrastructure be adapted to other simulation purposes?
    • Easily! toast infrastructure is set up
• Karl: GRASP sims status.
  • main beam masked beam created with center 1 x 1 degree pixel zeroed. This is a Far-Sidelobe-only beam.
    • Plots: (all .png)
    • center pixel = 0, no other effects
  • Andrea also asked for simple cases to test code: Karl made a 10 degree circular Gaussian and a 1 x 10 elliptical Gaussian
    • Plots of these test beams, just for sanity.
    • (.png)
    • (.png)
  • Karl to make a quick check to verify the main lobe actually is removed. Karl will be gone after next week for an internship will be in less contact
• Kris: precession
  • More to come on Wednesday: a preview is as follows:
  • 10 hour precession period: (.png)
  • 48 hour precession period: (.png)
  • 48 hour precession leads to coherent structures, lines of high coverage.
  • Kris suggests a fiducial toast simulation with 4-detector set to show full I,Q,U reconstruction in the presence, with CMB+dipole first. Extract power spectra.
    • this demonstrates PICO's basic ability to extract polarization with PICO scan, 1/f, etc.
    • Andrea would be the one to run this. Andrea reports that the easiest thing to do is to set a 4-detector single horn. We agree that we should quickly run this, then revisit later if we want to do 2 horns, scaled noise, etc.