Hi all,
A reminder of our first BICEP2 team meeting of the new year, back at our
usual time today:
Tuesday 8 Jan 2013
9:00 am Pacific, 12:00 eastern, 6:00am Pole
Phone: 1-866-890-3820 (toll: 1-334-323-7229) Passcode:59702175
As the agenda below reflects, I expect we will have a very limited
pipeline discussion today. Walt, can you please take notes?
1) General business:
- online logbook updated for Nov/Dec? [Jon, Colin, …]
- status of getting B2 summer data onto odyssey [Walt, Colin]
- votes on telecon time for next 5 weeks: 9am or 11am Pacific?
2) Organizing efforts toward results:
BICEP2 operations are complete. We'll be reorienting these meetings
toward organizing our efforts to produce and publish results. As we
discussed in our last (28 Dec) meeting, we'll be starting to match
individuals with assignments for "final reports" on B2 sub-threads on
reduction and instrument characterization. These are an evolution of
the threads we started outlining in November.
On the B2 analysis call today Walt volunteered to start a separate index
page for these reports, which we see as roughly mapping into sections of
papers. Proposed guidelines:
1. unlike logbook postings, these documents get edited and updated
until we publish
2. each report connects the topic to what is needed for the B2
science results--what final numbers or tests are needed, and why.
3. review past experiments experience, esp. B1 approach
4. provide an overview of tasks remaining
5. summarize (with links) work done in postings, and where
appropriate summarize the existing B2 data.
6. Format can be either html or google doc. Authors should seek and
incorporate all group feedback to converge on a consensus doc, ready to
extract conclusions for publication.
Example report topics:
- thermal systematics
- magnetic systematics
- spectral response
- temporal transfer function
- RPS-derived polarization chi's and epsilons
- pol cal consistency checks from Yukical / pixel polrasters
- far sidelobes
...
To drive this forward we need your buy-in both on format and
assignments. Thread coordinators, please think about topics within your
threads that need reports. Those who can't join tomorrow, send feedback
and/or offers to take assignments.
Today we'll agree on format, organization of topics under threads,
and initial assignments. Walt will continue this process on next
Tuesday's meeting (I'll be in transit). For future meetings I hope we
can schedule reviews of specific reports.
3) B-spectrum Pipeline
limited attendance expected today
brief update on data reduction [Walt, Angiola]
brief update from Jamie T on few degree sidelobes T->P sims
(Jamie=monopole, Chris=quadrupole)
other updates from ongoing work
4) AOB
--
___________________________________________________________________
John Kovac jmkovac(a)cfa.harvard.edu
Assistant Professor, Astronomy and Physics, Harvard University
160 Concord Ave rm 310, Cambridge MA 02138, 617-496-0611
Hi Jamie,
Good posting although result not very encouraging.
- When you fit for epsilon what area of the F-plane do you use? presumably you
should restrict to the central area where TB "lobes" are?
- You might try an extension which allows a radial variation of the strength
of the coupling - our false B seems even more concentrated to low ell than T -
an exponential or power law rolloff.
Clem
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
We are having porblems with phone connection...
You can't hear us!
Clem
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
Hi Chris,
I can't see your posting at the moment but sounds good.
It sounds like your work indicates that a uniform differential monopole
sidelobe with <10% power in the difference beam may be a viable candidate?
It's visibility (or not) in (non-diff) beam maps would be similar as would
it's ability to cause failure in the WMAP cross comparison. It also would
produce the proper TB correlation pattern (and evade jackknifes).
[I am aware that Jamie T has made a new posting on monopole sidelobes but I
can't see it either at the moment...]
Clem
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
Hi,
I've updated my posting from last week:
http://bicep0.caltech.edu/~spuder/analysis_logbook/analysis/20130121_ellipt…
...and made a new posting:
http://bicep0.caltech.edu/~spuder/analysis_logbook/analysis/20130121_ellipt…
1) I screwed up last week and instead of simulating ellipses oriented
parallel to detector pol. angles, I simulated ellipses parallel/perp. to
the (r,theta vector). Thus T leaked into E and B equally.
2) I was wrong in claiming that ellipses oriented parallel/perp. to the
pol. angles produce T->B leakage. As John pointed out, it in fact leaks
T->E.
The new posting simulates 1) ellipses oriented parallel to the pol angles;
and 2) ellipses oriented at 45 degrees to the pol angles. The leakage is as
expected.
A 10% elliptical sidelobe with 10% power and oriented at 45 deg. to the
pol. angle *does* cause noticeable spurious BB. We would not see this in
beam maps. I also had a quick look at the reobserved WMAP minus B2 maps,
and it looks like such a sidelobe would also *not* be noticeable in this
analysis. (Actually, we see a little bit higher rms in the WMAP minus B2
maps than would be expected from WMAP noise alone. I wouldn't go so far as
to claim this as evidence for a compact sidelobe, but it's interesting.)
However, the elliptical sidelobe does seem to cause TB correlation of a
form we do not see in the real data, so I think we can rule it out on these
grounds. Nonetheless, it's something to think about.
(Clem, it appears that ellipticity deprojection does *not* do anything to a
compact, 3 degree sidelobe with differential ellipticity.)
-Chris
--
**********************************************************************
Christopher Sheehy - Graduate Student - University of Chicago
University of Minnesota,
Room 220 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-625-1802 Fax: 612-624-4578 email: csheehy(a)uchicago.edu
**********************************************************************
Chris:
- Figure out if horiz/vert quadrupole leaks into E or B - reconcile with sim
results.
- Do subtraction of sidelobe versus non-sidelobe s+n T maps
- How much power in 3 deg sidelobe is allowed without distorting the T map
more than is acceptable given the excellent agreement between B2 real T map
and WMAP?
- (this could also be addressed by looking for slope in cross spectra abscal
style.)
- However based on what we're seeing in the B 2d pattern there is little
motivation to pursue quadrupoles much more... (Please dispute this if you
disagree.)
Randol:
- Please make radial profiles of B2 beam maps (and integrals thereof?) What
degree of monopole near sidelobe is allowed? (Maybe this is already done?)
Chris:
- Fix B-mode "noise" map in pager
- Is spatial variation of apodized B-mode maps consistent with apodization
alone? If not it could be telling us something important...
Jamie and Chris:
- Work on empirical T predicts B recipe
Jamie:
- Update near sidelobe monopole sim to make a uniform (jackknife evading)
effect
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
Chris monster pager:
- comp versus Keck maps
- multiple realizations
- different WMAP templates
- quad sidelobe sims and assumptions therein
- cross talk sim
- maps with no scan sync subtraction
Jamie T:
- update on monopole sims
Jamie and Chris:
- attempts to model predict TB from T
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
Hi John,
This post was discussed on the telecon. I commented that a uniform effect was
more relevant than the random pattern - which as you say will average down and
if strong enough cause jackknife failure.
However I am skeptical about making the effect fully differential and then
adjusting to max level which would evade existing beam map measurements. That
seems pretty artificial to me. Much more likely is that there would be such a
sidelobe whose differential component was many times smaller than its absolute
strength.
Generally I'd like to encourage more thinking about how our troubles could
result from non beam based effects. We can go ahead dreaming up very specific
jackknife evading sidelobe models which would also just evade existing beam
measurements. But supposing we find such a model - how likely is it to be
reality?
Clem
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
> Once we are messing with a non constant T->B leakage "smooth function"
> in fourier space we are talking about a beam effect.
I agree in the sense that such a leakage has a corresponding "effective
difference beam". However I have a feeling that such effective difference
beams do not necessarily correspond simply with actual physical difference
beams. There again given the constraint that the effect must be uniform across
the focal plane to pass deck jack then perhaps they do.
> Starting from the observed TB in the
> fourier space as you are advocating will be a faster way to find the right
> "beam effect", comparing to what Chris and Jamie are doing right now
> in real space.
So it is unambiguous? Does a given TB pattern imply one - and only one -
effective difference beam?
Clem
--
**********************************************************************
Clem Pryke - Associate Professor - Physics
University of Minnesota,
Room 313 Tate, 116 Church Street S.E. Minneapolis, MN, 55455
Tel: 612-624-7578 Fax: 612-624-4578 email: pryke(a)physics.umn.edu
**********************************************************************
Hi Jamie,
I don't know whether you discussed your monopole sidelobe sim posting on
the last telecon (I don't see notes on it). But reading your posting here
http://bmode.caltech.edu/~spuder/analysis_logbook/analysis/20130115_monopol…
I had expected you would inject a uniform amplitude A-B monopole
sidelobe into all pixels, not a random distribution as I think you have
done. We discussed that a uniform amplitude would not cause dk
jackknife failures. And obviously it will not average down, as I
suspect is happening to a large degree in your sim.
Quoting from your posting:
"I've completely made up this distribution; it is hard to tell in
Randol's sidelobe maps what the real level of sidelobe mismatch might be."
"It appears that adding a sidelobe does create a small amount of low l
BB power that does not deproject. However, we currently don't have
evidence that we have mismatched A vs B sidelobes, so while such a
effect can create B power, we probably need more evidence from sidelobe
maps as to its existence."
I think this is backwards. The existing near sidelobe analysis don't
offer positive evidence of a ~few degree monopole sidelobe, but what
upper limit it places is unclear--this has never been quantified. So
please plot the A-B beam for the (uniform, not random) monopole sidelobe
amplitude you choose in your sim on the same +/- 5e-3 colorscale and
axes as used in these postings:
http://bmode.caltech.edu/~spuder/analysis_logbook/analysis/20121217_bicep2r…http://bicep0.caltech.edu/~spuder/analysis_logbook/analysis/20121128_buddy_…
I suggest you choose an amplitude around the level so that it would just
start to be obvious on that colorscale if you added it on top of the
actual measured (tile-stacked) residual A-B structure (e.g. Abby's Fig
1). If you can get the data in that figure and actually plot the sum,
it would be great. I suggest trying 120 arcmin as well as 240 arcmin.
I think if you re-run your sim with these assumptions, it will show us
the signature of a common-mode monopole sidelobe leakage that is *at the
maximum level allowable by our current beam map analysis*. That's what
I want to see.
thanks,
John
--
___________________________________________________________________
John Kovac jmkovac(a)cfa.harvard.edu
Assistant Professor, Astronomy and Physics, Harvard University
160 Concord Ave rm 310, Cambridge MA 02138, 617-496-0611
