Minutes of Radphi Analysis Meeting, July 27 2000   -Richard Jones
-------------------------------------------------------------------

Present: Craig Steffen, Dan Steiner, Eric Koskinen, Cathleen Campbell,
	Russel Mammei, Erin Collins, Eric Walker, Brent Evans, David
	Armstrong, Elton Smith, Phil Rubin, Richard Jones

Notes:	Below are my notes of the presentations made at the meeting.
	Cross references are made to the to-do list at the bottom of this
	file using footnote references.

David Armstrong: reproducing Scott's spectra

	David now has code that can produce spectra similar to what
	Scott shows.  Right now he is using calibration constants (cc)
	from a file (not from the map) but pedestals from the map.
	He has included a fudge factor of 1.03 to get the mass scale
	right [1].  He is not sure yet that he is getting as many
	reconstructed omegas as Scott did.

Phil Rubin: looking at the LGD in Monte Carlo

	Phil has begun to look at the mass spectra from Monte Carlo
	and has seen some surprises.  He is submitting the simulation
	data to the standard cluster finder, and sees clusters being
	reconstructed in regions where the blocks are uninstrumented.
	Someone needs to fix this [2].

Phil Rubin: looking for tagger/LGD total energy correlations

	Phil has done a study of a dataset with selection criteria
	designed to select fully reconstructed events.  He then plots
	the total energy in the LGD versus the energy corresponding to
	each tagging counter inside a tagging coincidence window.  He
	sees a positive correlation between the two, although not with
	the expected slope of 1.  One suggestion was made that the MAM
	trigger is producing a bias that repells the LGD measurement
	away from the threshold.  The threshold is somewhat below the
	low-energy end of the focal plane, but perhaps not far enough
	below to affect the trigger efficiency there.

Everyone: special runs to be taken before end of period

	Following on the above discussion, it was decided that we want
	to have more data without the level-3 trigger, in order to be
	able to determine its efficiency.  The following list of special
	runs was identified as interesting.

	1) no MAM in trigger
	   Since the trigger rate will be far too high for the DAQ, we
	   will learn just as much if we turn down the current.  It was
	   decided to run at 2nA.

	2) no tagger in trigger (but levels 1-3 enabled)

	3) no UPV in trigger

	4) no BSD in trigger
	   This may require some rigging, because the tagger will be the
	   only positive condition in level 1.  The rate will have to be
	   brought down to about 1nA.

Craig Steffen: cluster cleanup library code

	Craig has extracted the code from his program that he uses to do
	cluster cleanup and has inserted it into the makehits library in
	cvs.  It is not called automatically, so the user has control over
	how much of this functionality gets applied [7].

Assignments:

1. [DA]	Oversee transfer of calibration task from Scott to Dan Steiner,
	check for convergence of results, check that correct mass spectra
	are produced from these calibration constants by the standard
	library code.

2. [RJ] Insert a check in makehits to suppress hits coming from dead
	channels (mainly a Monte Carlo issue) in all of the subsystems,
	particularly the LGD.

3. [RJ] Fix the Event display to work on both real data and Monte Carlo.

4. [RJ] Update the run logbook database, check into cvs, and add info
	regarding the whereabouts of the datat.

5. [PR] Send around a list of background channels that we need to simulate.

6. [PR] Work with Erin on a kinematic fitter.

7. [CS] Check in cluster cleanup code.

8. [RJ] Update runbook database, check into cvs, add information of
	whereabouts of data.