DEVELOPMENT OF NARROWBAND PHASE DISCONTINUITY SENSING
ALGORITHM ON NIRC2
1. Gary Chanan, Principal Investigator
Agusti Pinto, Co-Investigator
2. Purpose of test: development project
3. Test number: not applicable
4. Test duration: 10 hours (can be spread over several nights)
5. Brief description of test:
PDS (the narrowband version) has been used successfully on NIRC/Keck 1
for many years as a complement to PCS phasing. PDS does not currently
run on any Keck 2 instrument. The case for PDS on Keck 2 is not quite
as compelling as for Keck 1 because on Keck 2 one can quickly access
PCS even when the infrared instruments are being used (because on Keck
2 they share the same secondary). Nevertheless, there are several good
reasons for wanting to have PDS working on Keck 2: (1) PCS is currently
vulnerable to a catastrophic failure of the VMS hardware. Tbis situation
is likely to continue for some time. (2) A detailed test of PCS/PDS
differences can be performed better on Keck 2, since one can rapidly
switch between the two instruments; on Keck 1, by contrast, comparisons
cannot even be made on the same night. (3) It is not currently under-
stood what is limiting the accuracy of PDS on Keck 1 to about 40 nm.
One possible explanation is poor sampling of the image. We can thus
use the better sampled NIRC2 to resolve this question.
Ideally we would like to use a narrowband filter at around 3 microns
for these tests. [On Keck 1, PDS uses a pah filter with a central wave-
length of 3.31 microns and a FWHM of 0.063 microns.] This wavelength
represents best compromise between short wavelengths (to minimize the
background) and long wavelengths (to minimize the effects of seeing
and of segment aberrations). However, even if such an optimal filter
is not available, we can still use one of the existing narrowband
filters at a shorter wavelength to test the acquisition scripts,
and to quantify the effects of segment aberrations (and possibly
seeing) on the performance of the algorithm.
The test will consist of the following:
A. Run PCS to phase the telescope. Keck 2 will not have been
phased for two months prior to this date, so this part of the
test is in the interest of overall telescope performance,
independent of whether or not any PDS development is done.
B. Verify that the desired defocus distance and other PDS acquisition
parameters are correct.
C. Starting with a PCS-phased mirror, run PDS to test for consistency
with PCS and for convergence.
D. Starting with a PCS-phased mirror, introduce piston errors
within the capture range of narrowband PDS (400 nm) and attempt to
recover the correct phase.
E. If time permits, repeat step D with different starting configu-
6. Reasons why night sky is needed:
We have completed a new narrowband PDS code and have thoroughly
tested it by means of numerical simulations. We cannot make further
progress at this point without on-sky tests.
We emphasize that the continued development of PDS (both the narrow
and broadband versions) is highly labor intensive. We are fortunate
that most of these costs are covered under the terms of Pinto's Fulbright
Fellowship from the Spanish government. Pinto has devoted the first
six months of his fellowship to offline work on these algorithms and
is now at a point where telescope time is essential for further
progress. Since he will only be at UCI through January 2004, it
is very important to secure telescope time as soon as possible.
7. Plan for data reduction, etc:
The reduction and analysis of PDS data for Keck 2 are virtually identical
to those for Keck 1. The main differences involve the different
image scales and array sizes on NIRC and NIRC2, and some differences
in the acquisition scripts. All of these changes will be straightforward
8. Preferred dates:
September 16-18, 2003. Pinto will be on Keck 1 on September 13,
so these nights on Keck 2 would represent an efficient use of our
limited travel resources.
9. Progress report:
PDS has been used successfully for many years on Keck 1, but has
not yet been attempted on Keck 2.