From loveless@wishep.physics.wisc.eduMon May  8 17:48:22 2000
Date: Mon, 08 May 2000 17:31:53 -0600
From: loveless@wishep.physics.wisc.edu
To: leer@physics.purdue.edu, riterry@lynx.dac.neu.edu, evg@neu.edu,
    loveless@wishep.physics.wisc.edu, lusin@wishep.physics.wisc.edu,
    jhm@neu.edu, maeshima@fnal.gov, deartly@fnal.gov
Subject: thoughts of a run plan to the ISR

Hello,

   At the alignment meeting today it was difficult for me to 
understand what we need to do at the ISR to verify that our
alignment system is working well.  So I am writing down a set
of steps that seem necessary for the ISR test.  It is possible,
of course, to view the steps in a different order but since the 
data is all the same, we should get the same result.
   I will skip over some details assuming we all know the steps
involved.  If things are not clear, please let me know.  And
please let me know if there are other tests/measurements to make.

  ISR Run Plan  (proposed)

  1)  Set up all equipment according to the ISR drawing.

  2)  Make sure all equipment is operational.

  3)  Adjust the sensor locations so all sensors give good signals
	and are roughly in the middle of their range.

  4)  Once these conditions are satisfied, take the photogrammetry pictures
	and work out the locations of all the photog targets.

  5)  Calculate axial line

	Here we can start with the laser line as defined by the MAB and then
	compare to the photog measurements.  Or you can start with the 
	photog locations of the sensors and calculate the residuals for
	the line.  Either way you should get the same answer.  I am starting
	with the MAB data.

	*  Define this line with x,y,z points on each of the MABs.  This
		results in a 3 dimensional line in space.

	*  Using this 3D line and the data from each sensor (including the
		z location from the z link readout), position each of the
		sensors along this line.

	*  For each sensor calculate the 3D location of each of the photog
       		targets.  

	*  Compare with the photog measurements and compute a chi square
		(or a set of residuals).

	*  Using the photog measurements to define the sensor locations, 
		calculate the best fit axial line.

  6)  Calculate the SLM line

	Here we start with the drawing/calibration of the transfer plate
	and the known axial line.  Using this information and the data from
	the SLM sensors, position these	sensors on an SLM line projected 
	from the transfer plate.  Then compare with the photog measurements.

	*  Use the definition of the axial laser line and the drawing (or
		calibration) of the transfer plate to project a 3D SLM laser
		line.  Get the equation of this SLM line in 3D.

	*  Using this 3D SLM line and the data from the SLM sensors (readout
		x-y data as well as the data from the radial sensors) to 
		locate each SLM sensor in 3D along the SLM line.

	*  For each sensor calculate the 3D location of each of the photog
		targets.

	*  Compare with the photog measurements and compute a chi square
		(or a set of residuals). 

	*  It is likely that the SLM line as projected is not very good.
		We should refine it using the photog measurements and get
		a best fit SLM line.  Using this information we can check the
		calibration of the transfer plate.

  7)  Can we use the "other" SLM in the opposite direction to check this
	line?

  8)  Once the best fit axial line and the SLM line are defined, we should 
	calculate the angle between them and compare to the calibration of
	the transfer plate.

  9)  Understand how "real" life will work.  In "real" life there will not
	be any photog measurements that we can rely on.  The disks will have
	moved, the floor will have shifted, etc.  How sensitive are we to
	errors?  In point 5 if the MAB data is wrong, the 3D line is wrong.
	How can we recover?  Can we tell if the MAB data is wrong?  Can we
	define an axial line if the MAB data is missing (without photog
	measurements)?

  10)  Once the best fit lines are determined we should run the system for
	a lengthy period to look at stability and reliability of the data.

  11)  Develop a database for storing all the information and interface this
	database to the software analysis.  This should be done in
	collaboration with the link and the barrel groups so we can define 
	a consistent database.

  12)  Once the system is operating smoothly we should perturb elements and
	determine how to make the system more robust.  In "real" life, of
	course, there are 6 axial lines and we can cross check.  What can 
	we do at the ISR test to verify the axial line?

   There are other goals and tasks which Jorge has described in his memo. 
And this list of tasks can be made more detailed.  I would appreciate if
people could provide feedback on the details of this outline.

                                                Cheers,

                                                  Dick