[Frost] Some problems with g9a

[Michael Dugger]

Hi,

I believe that we need a systematic approach to finding and fixing 
problems with the g9a data. In particular, some of these problems include:
* Finding the CLAS momentum correction that feeds into the eloss program
* Finding any remaining problems with eloss
* Parameterizing the amount of leakage of Carbon into Butanol (and vice 
versa)
* Verifying that the tagger sag correction is correctly implemented

Looking at the energy correction (by way of kinematic fitter) on the
Butanol target is the worst of all possible worlds. We have an energy
correction that I believe is not to be completely trusted, large eloss
corrections that may not be correct, possible tagger sag issues, CLAS
momentum corrections (that feed back into the large eloss corrections),
and the precision of the vertex construction "leaking" Carbon into the
Butanol.

The reason that I keep saying that the kinematic fitter does not give good
energy values when the pull is not close to optimal is from the g8b
experience. In g8b, the energy differences given by the FSU kinematic 
fitter were ALWAYS close to zero. However, when we corrected the energies, 
the pulls were much better. This is why I say that the pulls are more 
meaningful.

One approach is to try and minimize the effects coming from eloss and 
isolate the tagger photon energy corrections. THEN we look at the CLAS 
momentum corrections. THEN we look to parameterize the Butanol/Carbon 
contamination. THEN we look for problems with eloss.

We can minimize the eloss effects by concentrating on the CH2 target with 
high momentum final state particles. This has the added benefit of "side 
stepping" the contamination issue. From this data, it should be possible 
to verify that the tagger sag is correctly dealt with. Once the tagger 
sag issue is addressed, then we find the CLAS momentum corrections and 
check that eloss makes sense for the CH2 target.

After the tagger photon energies and CLAS momentums are fixed, we 
parameterize the contamination, and move on to the Carbon and Butanol 
targets. At this point, we should be able to narrow most problems to be 
associated with the eloss program.

Currently, I need to finish up the g8b analysis. Assuming that Ross has 
the time, we can start working together on the energy/momentum corrections 
for CH2 when he gets back to ASU. This should take a few weeks to 
implement and test. The code I used to determine the tagger energy 
corrections was written for the PAW framework, and I have now moved on to 
ROOT for g9 data. This means that the simple iterative energy/momentum 
correction routines need to be converted to ROOT.

Potential problems: The simple iterative routines worked well with a 
Hydrogen target. It is possible that the Carbon background from CH2 will 
cause us difficulties. I can't know until we try. Also, I did not 
parameterize and test the momentum corrections for g8b. I simply used the 
corrections supplied by Chuck. This means that the momentum corrections 
obtained by the simple iterative approach might take us more time to 
develop.

If anyone else wishes to find and parameterize the photon energy and CLAS 
momentum corrections, please let me know. This is time consuming and 
boring stuff I would rather not do.

Take care,
Michael