[Phys-L] Re: Swift's Self-Described Burster Detections.

[palmer at SFU.CA (Leigh Palmer)]

Brian Whatcott wrote:

> At 12:25 PM 12/17/2004, Leigh Palmer, you wrote:
> > I have been away for a while, lurking inattentively. We went down to
> > Cape Canaveral for the launch of Swift. For those who know me (or
> > David) I pass along these items:
> >
> > http://www.lanl.gov/orgs/pa/newsbulletin/2004/11/17/text02.shtml
> > (Read to the end. The article is actually about David.)
>
> I quote:
> " In the imaging equipment aboard Swift, 54,000 pinholes in a panel of
> lead
> the size of a full sheet of plywood produce an "image," actually
> thousands
> of overlapping images (approximately 30,000 of them). The Los Alamos
> software must unscramble those overlapping images and make one
> stronger,
> brighter picture from which the precise location of the gamma-ray
> burst can
> be found, while eliminating known sources and statistical variations"
>
> I haven't really read much of the referenced URLs, but I have some
> points
>  to make on the topic.
>    It is hard to imagine a more clumsy arrangement to be carried
> aboard an
>  expensive rocket payload than something which is essentially a 4 by 8
> ft
> sheet of lead with holes in it. (A glancing reflection gold-coated tube
> bundle would be more suited I suspect.  But then, I am no expert.)

The Burst Alert Telescope (BAT) on Swift is a wide field gamma ray
survey camera, analogous to a Schmidt telescope in the optical region
of the spectrum. It looks at 1.4 steradians of the sky at once. That's
more than ten percent of the whole sky! It is difficult to image gamma
rays even in a narrow field; grazing reflection doesn't get you very
much sky at these high energies.

A gamma ray pinhole camera can cover a large solid angle, but the image
would be very weak, just as it is in an optical pinhole camera. Making
the pinhole larger reduces the resolution of the image while increasing
the signal strength, so making the pinhole much larger than the size of
a pixel on the detector plane isn't practical.

Another approach would be to poke a second pinhole into the camera. The
signal strength is doubled, but the image looks funny. However, knowing
the orientation of the line connecting the pinholes and the distance
between them, one can compute an image from the pairs of points in the
funny looking image.

OK, let's poke a third hole in the camera. This time we have to be a
bit clever. It turns out that the maximum benefit can be attained by
placing this third hole so it is not on the same line as the first two
holes, and so that the interhole distances are all different. The
signal strength is increased by fifty percent, and it is still
possible, though somewhat more difficult, to deconvolve the image from
the data.

Since poking more holes in the camera increases signal strength it is
tempting to think that one can benefit further from poking even more
holes, and that is indeed the case, but only up to an optimum number of
optimally placed holes. That optimum number is reached when half the
area of the camera is perforated, and the optimal placement is called a
Hadamard mask or, alternatively a uniformly redundant array [q.G.].

David Palmer has been working with telescopes of this sort for almost
two decades. His PhD was done under the supervision of Tom Prince at
Caltech (now Chief Scientist at JPL), and his thesis object was the
supernova SN1987A. I have been told he designed the mask for BAT.

>   It reminds me mostly of the Halloween decoration I made for the
> house:
> I cut the base out of a pumpkin, and drilled one hundred holes in it.
> Into
> these I thrust the fairy lights from two strings from the inside, after
> pulping the pumpkin and swilling it with bleach solution. Pretty!

There is a pretty picture of Swift's aperture mask at
http://swift.sonoma.edu/images/multimedia/images/sc/swift_aperture.jpg
or, for your desktop,
http://swift.sonoma.edu/images/multimedia/images/sc/swift_aperture.tif
(6.5 Mb).

> I was particularly exercised by the literary treatment of a test
> sequence of running a CCD camera into a closed instrument barn-door.
> Yes, my dears, the King really IS naked!
>
> I am an opponent of high stakes science in space: I think it is
> unscientifically grand-standing and in fact, financially ruinous.
>     Britain and Italy have apparently signed on with others, but I hope
>  the ridiculous project will peter out under the weight of the
>  ever-growing deficits which fund it.

We could also save by ceasing support for nonessential Art Galleries,
Museums, and Symphony Orchestras. We would, of course, be a different
species were we to do so. While I am an evolutionist, I certainly hope
we don't evolve toward that direction. I also hope that you will
reconsider your view on generous public support of science.

> I don't believe that the raw data has been published for these
> "bursters":
> they are below the x-ray noise threshold, I presume, and  the events
> detected rely heavily on fore-knowledge of the events which are sought.
> In other words, the models which are used to extract the data thereby
> condition its discovery.
> Different model: different event.....

I'm unsure I understand this point, Brian. Swift scientists will
certainly be able to discover unanticipated new science. In fact David
has spent more time programming it serendipitous faculties than its
primary mission functions. That's what's meant by the "lots of knobs"
metaphor in the article. David will not be coming home for Christmas
this year because of the flood of interesting data now coming in, even
before Swift is all switched on. I think he's been home for 42 straight
Christmases, including even the one we spent "at home" in Cambridge,
England.

You may be heartened to learn that Swift's GRB data will not be
embargoed; it will be made public immediately on the Swift web site.
That is a break with previous practice as exemplified by Hubble's one
year proprietary embargo on scientific data. The investigator who got
the data could use it exclusively for that period.

Leigh

(Pardon the neoclassical pretension [q.G.].  It stands for [quid
Google])