Index: /trunk/doc/design/Makefile
===================================================================
--- /trunk/doc/design/Makefile	(revision 8140)
+++ /trunk/doc/design/Makefile	(revision 8141)
@@ -1,3 +1,3 @@
-# $Id: Makefile,v 1.17 2006-08-04 12:50:49 eugene Exp $
+# $Id: Makefile,v 1.18 2006-08-04 12:55:02 eugene Exp $
 
 PDFLATEX = env TEXINPUTS=../../latex/inputs:$(TEXINPUTS):.: pdflatex
@@ -6,5 +6,5 @@
 help:
 	@echo "USAGE: make (target)"
-	@echo "  targets: srs ssdd scd all cdrresp pars"
+	@echo "  targets: srs ssdd scd all cdrresp pars comm"
 
 scd: ippSCD.pdf
@@ -13,4 +13,5 @@
 cdrresp: ippCDRresponse.pdf
 pars: ippParameters.pdf
+comm: ippCommission.pdf
 
 scd-draft: ippSCDdraft.pdf
Index: /trunk/doc/design/ippCommission.tex
===================================================================
--- /trunk/doc/design/ippCommission.tex	(revision 8141)
+++ /trunk/doc/design/ippCommission.tex	(revision 8141)
@@ -0,0 +1,281 @@
+\documentclass[panstarrs,spec]{panstarrs}
+
+\title{IPP Commissioning Plan for PS-1} % put in your title
+%\subtitle{Specification}
+\author{Eugene Magnier}
+\audience{Pan-STARRS Science}
+\shorttitle{AP Survey Requirements}
+\group{Pan-STARRS IPP}
+\project{Pan-STARRS IPP}
+\organization{Institute for Astronomy}
+\version{DR}
+\docnumber{PSDC-xxx-xxx}
+
+\newcommand\ugriz{$u^\prime g^\prime r^\prime i^\prime z^\prime$}
+\newcommand\grizy{$g r i z y$}
+
+\begin{document}
+\maketitle
+
+\section{Overview}
+
+The Pan-STARRS Image Processing Pipeline will undergo extensive
+commissioning tests in the period September - December, 2006.  With
+the staged arrival of the telescope, the first camera (TC-3) and
+optics, and the main camera (GPC), the commissioning will take place
+in a series of steps.  Initially, basic operating functionality will
+be demonstrated along with analysis throughput, without specific focus
+on achieving the final astrometric and photometric precision goals.
+Eventually, with the arrival of the GPC, the complete system
+requirements will be used as the commissioning criteria.
+
+\section{Pipeline / Infrastructure tests with Megacam data}
+
+Before actual camera hardware is delivered, the IPP will begin
+commissioning tests related to the infrastructure and interfaces.
+These commissioning tests may be performed with only a few IPP cluster
+nodes in the IfA computer room before hardware is deployed to Maui.
+Below, we list the commissioning tasks required to demonstrate the
+working infrastructure of the IPP.
+
+\begin{itemize}
+\item {\bf PanTasks / ippTools / Analysis Programs interactions :
+  Detrend Creation} The complete IPP pipelines for Phase 0 and the
+  Detrend Creation steps should be installed and set up on the test
+  cluster.  A complete megacam dataset must be run through the
+  pipeline.  This dataset will consist of raw bias, dark, flat-field,
+  and fringe images.  This commissioning task will demonstrate the
+  creation of master detrend images from raw detrend images.
+
+\item {\bf PanTasks / ippTools / Analysis Programs interactions :
+  Phase 2} The complete IPP pipelines for Phase 2 should be installed
+  and set up on the test cluster.  A complete megacam dataset must be
+  run through the pipeline.  This dataset will consist of science
+  images along with the master detrend images generated by the
+  previous task.  This commissioning task will demonstrate the
+  application of master detrend images to the raw science images and
+  the analysis up to object detection and photometry. (Phase 1 - guide
+  star analysis - is irrelevant to the Megacam test data).
+
+\item {\bf PanTasks / ippTools / Analysis Programs interactions :
+  Phase 3} The complete IPP pipelines for Phase 3 should be installed
+  and set up on the test cluster.  A complete megacam dataset must be
+  run through the pipeline.  This dataset will consist of science
+  images along with the master detrend images generated by the
+  previous task.  This commissioning task will extend the previous
+  test to include astrometric calibration relative to the installed
+  DVO database.  It will also demonstrate the operation of the
+  photometric calibration relative to the DVO database, though the
+  reference data will not be appropriate to judge the photometric
+  results at this stage.
+
+\item {\bf Camera - IPP Image transfer} Test GPC-1 images will be
+  injected into the test IPP system using the Data Store mechanism.
+
+\item {\bf Otis - IPP Images} Test ISP images will be injected into
+  the test IPP system using the Data Store mechanism.
+
+\item {\bf Otis - IPP Metadata} Test metadata tables will be passed to
+the test IPP system using the Data Store mechanism.
+\end{itemize}
+
+\section{Camera in lab : Detrend Image Tests}
+
+Before either camera is mounted on the telescope, as set of Camera-IPP
+commissioning tasks may be performed using biases, darks, and
+(depending on camera lab options) rough flat-field tests.  These
+initial commissioning tests are performed before a bad pixel mask is
+generated.  The tests are conditional on the bad pixels not
+overwhelming the statistics.  The same tests should be re-performed
+after the bad pixel mask has been built with data obtained on the
+telescope.  When the mask is available, the success criteria can be
+set to more stringent limits.
+
+\subsection{Bias construction / application}
+
+Obtain a stack of ($N > 10$) bias images.  Build a master bias and
+generate residual images (automatically part of the detrend creation
+system).  Examine the detrend creation Q/A statistics: residuals per
+chip, residuals per frame, number of 5 sigma outlier pixels.
+
+\subsection{Dark construction / application}
+
+Obtain a stack of ($N > 10$) dark images for each of a range of
+exposure times (15, 30, 60, 120, 240 seconds).  Build master darks and
+generate residual images (automatically part of the detrend creation
+system).  Examine the detrend creation Q/A statistics: residuals per
+chip, residuals per frame, number of 5 sigma outlier pixels.  Model
+the trend of average dark current vs exposure time.
+
+\subsection{Rough Flat-field construction / application}
+
+Obtain a stack of ($N > 10$) flat images for each of the filters with
+a range of count levels.  Build master flats and generate residual
+images (automatically part of the detrend creation system).  Examine
+the detrend creation Q/A statistics: residuals per chip, residuals per
+frame, number of 5 sigma outlier pixels.  Examine the amplitude of the
+flat-field residual as a function of count level (accuracy of shutter
+ballistics model, non-linearity).
+
+\section{Camera on telescope : Detrend Image Tests}
+
+\subsection{Bad Pixel Mask Construction}
+
+Supply a series of flat-field images to the mask creation
+tool. Examine the statistics of the flagged pixels, and examine the
+coverage by eye.
+
+\subsection{Basic Detrend Construction}
+
+Rerun the tests above (bias, dark, flat) using data obtained on the
+telescope, and applying the mask created above.
+
+\subsection{Fringe master construction / application}
+
+Obtain a stack of many potential input fringe images: night-time
+images in the appropriate filters over a range of airmass and time of
+night.  Build master fringe images and generate residual images
+(automatically part of the detrend creation system).  Examine the
+detrend creation Q/A statistics: residuals per chip, residuals per
+frame, number of 5 sigma outlier pixels.  Examine the amplitude of the
+fringe residual as a function of count level, airmass, etc.
+
+\subsection{Non-linearity}
+
+Measure the non-linearity by obtaining flat-field images with a range
+of exposure times and flux levels.  Measure the non-linearity by
+observing star fields with bright stars with a range of exposure times
+from short to long exposures.
+
+\subsection{additional detrend characterization tests}
+
+\begin{itemize}
+
+\item bias levels as a function of detector \& dome temperature
+
+\item dark level as a function of dome light level
+
+\item flat-field consistency as a function of illumination level:
+   obtain a sequence of input flats for each filter over the entire
+   twilight period.
+
+\item flat-field consistency as a function of internal vs external
+   illumination: obtain a sequence of input flats with dome shutters
+   opened different amounts
+
+\item flat-field consistency as a function of time: repeat the
+   flat-field construction process over N (> 5) photometric nights
+
+\item flat-field consistency as a function of solar angle: obtain a
+   series of flat-field sequences at a range of Alt/Az positions (Az =
+   0, 90, 180, 270; Alt = 45, 60, 80).
+
+\end{itemize}
+
+\subsection{Fringe construction tests}
+
+\subsection{Fringe consistency tests}
+
+\begin{itemize}
+\item fringe as a function of time-of-night
+
+\item fringe as a function of airmass
+
+\item fringe as a function of lunation
+
+\item fringe evolution twilight-to-nighttime
+\end{itemize}
+
+\subsection{Skyflat vs Domeflat comparison}
+
+Build master detrend images using both dome and sky flats.  a 
+
+input data: stack of N (5 - 10) skyflat images
+
+process: build a master flat, generate residual images (automatically
+part of the detrend creation system)
+
+test: residuals per chip, residuals per frame < limit, number of
+5 sigma outlier pixels.
+
+\section{Astrometry and Photometry}
+
+\subsection{Phase 2 / Phase 3 Astrometry solution tests}
+
+\begin{itemize}
+\item observe a high-quality astrometry field
+
+\item astrometry convergence for a range of stellar densities
+
+\item astrometry accuracy: perform a series of dithers and demonstrate
+   the astrometry solution
+\end{itemize}
+
+\subsection{Build the Flat Correction Frames}
+
+\begin{itemize}
+\item dither sequence in dense stellar field
+
+\item demonstrate photometric consistency within the dither pattern
+\end{itemize}
+
+\subsection{Photometry Calibration (Zero-Point) Checks}
+
+measure zero-points (using photflat) on megacam standards fields
+
+\section{Phase 4 Tests}
+
+\section{AP Survey Preparation Tests / Research}
+
+\subsection{Skyprobe Pre-AP Survey}
+
+Skyprobe observations in all 5 filters over 6x8 patch takes 30 minutes
+(45 seconds per point, one sweep in all filters).  The observable sky
+in one night is $\sim 25,000$ square degrees (ie 15\% less than 3pi),
+taking roughly 4500 pointings using the base overlaps.  The entire
+observable sky can be observed in 47 hours, or about 5 x 9 hour
+nights.
+
+In the pre-TC-3 / TC-3 era, we can perform a complete sweep of the
+observable sky every month. \note{dark half of the lunation only?}
+Between Feb-Jun, we can obtain 5 complete passes, including at least 3
+on even the portions which are in conjunction with the sun at some
+point in these 5 months.  The cost would be 25 photometric nights
+between Feb-Jun (16\% of the available nights).  This dataset will
+allow us to:
+
+\begin{itemize}
+\item build a complete skyprobe photometry dataset for the full 3pi.
+
+\item test the static sky construction process before GPC arrives.
+
+\item study the transparency stability of the sky at Haleakala.
+\end{itemize}
+
+\section{Transparency / Airmass Tests}
+
+we need to understand the stability of the atmosphere on Haleakala.
+Two effects which must be distinguished are variations in the airmass
+slope and the stochastic variations in the absolute transparency.
+
+we can follow a single spot on the sky (eg, Dec = 0 or 35) from the
+beginning of the night to the end of the night, for 9 - 11 full hours
+of transit.  we can obtain a 30-second image continuously (expose,
+readout, expose).  We need to do this on a photometric night (if the
+weather deteriorates, break and do some other tests).  More than one
+of these would be nice, at least one per filter.  the resulting
+photometry will illustrate the transparency of the sky versus the
+consistency of the airmass slope (thus, more than one night would be
+good, esp for g').  The measurement does not need the best possible
+flat-field image: keeping the position stationary reduces the
+sensitivity to the flat-fielding.
+
+To measure the stability of the atmosphere at high zenith angles, we
+can sit on the north pole field for several hours on a photometric
+night, performing 30 second continuous exposures.
+
+These two experiments would cost 5-10 photometric nights.  They can
+also be used to demonstrate the astrometric stability across a range
+of image qualities (high-to-low airmass).
+
+\end{document}