[b1_ana] (no subject)

[Dustin Keller]

Some thoughts,

In response to Wallys questions a PVDIS asymmetry contribution as 
described in their paper seems to require some non-zero value to the 5
  spin-dependent structure functions listed for the PV differential cross 
section which should not show up for the real spin-1
target.  It can show up in the photon-Z interference but with our energy 
we are safe to say that the PV asymmetry is zero.  His question may have 
more to do with what is the best way to measure the PV asymmetry, not 
sure.

The experiment as we have written it up requires an unpolarized beam.  If 
we have decreased our chances of approval by that request I would like to 
understand the details behind that.  Otherwise this is a single
configuration change for the preparation of the experiment.  We need
true unpolarized beam in-order for our approach and method to be valid.
If it turns out that we can not have unpolarized
beam we can calculate the contribution from the additional background but 
the degree of beam polarization (and error) can easily overwhelm what we 
are trying to look at.

In general the orientation of the field should always be chosen along the 
q-vector when possible.  We have known and understood this for a while. 
We chose to orient the field along the beam line for simplicity and ease 
of configuration.  This is fine but this leads to issues that may be 
caught by the PAC so having an answer in mind is probably a good idea. 
When its is not possible to orientation the field along the q-vector it 
best to select the closest possible orientation to the q-vector with the 
quantization axis still in the scattering plane.  This concern only hold 
relevance for our highest x points, and even along the beam line the 
theta_d and phi_d are small.

For Jaffe the virtual photon is set along the quantization axis which is 
chosen to be the z-axis.  This is true throughout the formalism and was 
setup that way for simplicity.  This implies that corrections must be made 
if the virtual photon points anywhere other than along the z-axis.  The 
observable Azz is more robust but still requires corrections.

The effect on Azz to a first approximation is a reduction of the effective 
polarization under a Wigner rotation of (3/2cos^2(theta_d)-1/2).  This
assumes that for our larger x points the Azz is dominated by
the T20 contribution.  This approximation is better for smaller theta_d.

In addition to a correction to the polarization when the momentum transfer 
vector is not aligned parallel to the deuteron polarization axis
  then phi_d and theta_d are non-zero likely leading to a small 
contribution from the target-only vector asymmetry for our higher x 
points.  Though the Arenhovel formalism is for electro-disintegration if 
we assume a static deuteron the polarized inclusive cross section
  can be integrated leaving the only kinematic dependence on the out going 
nucleons contained in know form factors.  This leaves us with the capacity 
to estimate the contribution from the target-only vector asymmetry. 
Estimates indicate using the present configuration as mentioned in the 
writeup to be still smaller than the error previously outlined.  Clearly 
it is best to be open to changing configuration and modifying run plans to 
minimize all such effects.

The systematic error table pointed out by Steve for their proposal was 
taken from a thesis on R.  The way these errors are presented in the 
thesis is quite different then how they are presented in the proposal that 
Steve pointed out.  I think they are referring to these contributions as 
point to point errors only because in the cross section it is not critical 
to separated these types of contribution from those that propagate 
relative to the observable.  In other words it doesn't hurt them to 
over-estimate.

For us its very important to separate these as we have done in the writeup 
of the proposal.  There are some very small contributions that can be 
added to our relative uncertainties, but they do not add any overall 
uncertainty.  But we can add them to acknowledge the comment.
The only other contribution I can think of  to the absolute point to point 
effects would be correlation
terms.  Analytically there is a correlation between effective polarization 
uncertainty and the change in Azz from drifts.  I think this type of 
description is redundant for the systematic uncertainty estimates in the 
proposal.  I have added some relative contributions to our table, but I 
think modifying the proposal
is optional in this regard, see attached.  We can also split the table 
into different contributions as Oscar suggested.

A lingering issue has been the contribution to the BCM drift.  The number 
I used in the proposal was just an optimistic guess.  Pengjia gave me some 
data on the calibration over the g2p/gep runs.  There was a period of six 
day for which the calibration drifted 0.02% which is of the same order I 
used in the estimate on the proposal.  All other BCM calibration drifts 
where larger but a least this is an indication that ~0.01% is not 
impossible.


dustin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Azz_error.tex
Type: application/x-tex
Size: 1098 bytes
Desc: 
Url : https://mailman.jlab.org/pipermail/b1_ana/attachments/20130521/c5963ae7/attachment.tex