[mmaimcal] CRE 173 B7 cross-polarization

[Rob Reid]

Hi,

At 4:27 PM EDT on June 1 Steven T. Myers sent:
> 
> After reading it I still have no idea of what the level is
> in science terms.  At -16db (the new spec) does that
> mean 2.5% "leakage"?

That's my understanding.
 
> The xpol from the slant grid looks to peak at the -6db
> level in the beam, and is better within the 3db contour.

Note that plot was from a model, not measurements.

> They give no pattern for the IR birefringence.

I'm wondering, since it's a polarizing filter, if its leakage map is fairly
flat.  That would be good for wide field mapping, since on-axis calibration
would remove it, but bad for on-axis observations, since it would put leakage
where there usually isn't very much.

> (impact: if you want good polarization maps over a reasonable FOV then
> mosaic).

And as usual, accept that less of the primary beam will be usable than for
Stokes I.

> Also, since this seems to be due to optics, there is some hope it is stable
> (and thus calibratable).  Do they have any idea on stability?

The front page says this:

"We understand that the temporal stability of the cross-polarisation pattern
(on-sky) is more important than its raw instrumental value; the real impact may
be tied to calibration and observing procedures."

It's true that stability is crucial for correction to work, but I think we need
to draw a line around 10% leakage, preferably lower, where correction breaks
down.  This is because of
A. 2nd order effects, which may be correctable, but not without making
   someone's brain hurt, and likely not even then, and
B. Efficiency problems, because a sizable fraction of the light isn't going to
   where it's supposed to.  In other words, even if we can correct to get the
   right nominal values of I, Q, U, and V, in the current first order scheme
   the noise would go up, and
C. Even observations only requiring Stokes I, and even of unpolarized sources,
   would need full polarization correction.  With a heterogeneous array that's
   especially expensive.

> On Fri, 1 Jun 2007, Al Wootten wrote:
> 
> > Folks,
> >
> > I commend to your weekend attention:
> > 173.  	Band 7 cartridges: Cross-Polarisation
> > CRE Number :ALMA-40.02.07.00-095-A-CRE
> > (Attachments on edm)
> >
> > As Hans says, it seems we have no options on this one.  We should, however,
> > specify the compromises this CRE will force on ALMA polarization
> > measurements, for which B7 is the band of choice.  We can always be the
> > lone vote for a dissenting position.
> >

We could at least press them to make off axis xpol maps, including the IR
filter, a higher priority.  I'm mot sure what to make of Fig. 1.

-- 
Rob Reid		A NRAO ALMA PDF			@   @  @ @  @   @
1-434-244-6822	       	http://www.cv.nrao.edu/~rreid/  ^   ^  ^ ^  ^   ^