[d2n-analysis-talk] More on Sign; Small Changes to A1 Results

[Diana Parno]

Hello all,

I've been corresponding with Jin about the sign convention for our  
beam helicity signal. It seems that both Matt's quasielastic 3He  
measurement and the similar Transversity measurements agree: if the  
Moller measures a polarization with positive sign, then the adchel==1  
signal in the LHRS and in BigBite corresponds to positive electron  
helicity, and the adchel==-1 signal in the LHRS and BigBite  
corresponds to negative helicity.

However, Jin reminded me of another salient fact. The electron spin  
precesses by a huge amount -- many full revolutions, and indeed over  
10000 degrees in some cases -- as it passes through the accelerator  
and into Hall A. (See Grames et al. in Phys Rev. STAB, volume 7,  
042802 (2004) for a nice reference.) The exact amount of this spin  
precession varies greatly depending on configuration changes, e.g.  
changes in pass. And, though we've been referring to our 4.74-GeV  
dataset as the "four-pass data", it really isn't. The first part of it  
(target S=0 and S=90 measurements) were indeed at four pass, but the  
second part (S=270) was at FIVE pass, after MCC removed several  
cryomodules from the circuit and reduced the energy added per pass.  
The beam energy was the same, but the accelerator configuration was  
very, very different.

A configuration change like this can result in a helicity flip -- i.e.  
the relationship between the actual helicity and the helicity signal  
in the hall is inverted, even without a change in HWP status. It turns  
out that we have three pieces of evidence that this happened between  
the S=90 and S=270 measurements, concurrent with the pass change:

1. The Moller measurements before the pass change (March 6) and after  
the pass change (March 12) reported polarizations with different  
signs, indicating a helicity sign flip.
2. The Compton also saw a helicity sign flip at the same time. I  
brought this up in my 18 November progress report, but we decided it  
was probably a cabling change in the Compton. However, it coincides  
exactly with the pass change.
3. Our transverse production asymmetry itself: this asymmetry should  
flip signs if the target spin flips and the electron helicity  
convention remains the same. However, we saw no sign flip. We thought  
this meant that the asymmetry was consistent with zero; instead, it  
seems to be real.

Together, this adds up to a change in our helicity convention between  
the S=90 dataset and the S=270 dataset. The S=270 asymmetries I showed  
yesterday therefore had the wrong sign; since A_perp takes its sign  
from the S=270 data, it does too. This also had the effect of diluting  
A_perp, since the S=90 and S=270 asymmetries partially canceled each  
other.

Luckily, once I knew about the sign flip, it was a very easy thing to  
fix in my asymmetry code. I've attached four new figures showing the  
effect of this correction on our data. First are the raw asymmetries  
for all three target spin configurations; next, the physics  
asymmetries (incorporating dilution factor and target/beam  
polarizations); next, A1 and A2; and finally, A1 in the context of pre- 
existing data. A_perp contributes more to A1 than it did before.

This does not affect the calculation of the kinematic variables and  
coefficients used in the computation of A1 and A2.

We will need to remember to be careful about this in the five-pass  
dataset. We should be able to use the Moller to set an unambiguous  
sign for each beam configuration.

The original sign convention problem (the one we applied Matt's  
quasielastic asymmetry to) is probably worth keeping in mind for  
future polarized experiments (i.e. the 12 GeV d2n experiment), too.  
Taking some dedicated sign-determination data would save a lot of  
ambiguity and work here. The standard way to do this is by arranging  
with MCC to take data with a very high charge asymmetry. The charge  
asymmetry can then be measured in the spectrometer and polarimeter  
DAQs, so as to set up an absolute truth table that can be used for the  
rest of the experiment. (And, as a bonus, check the logic on any  
helicity-gated scalers.)

Best,
Diana

-------------- next part --------------
A non-text attachment was scrubbed...
Name: raw_asys.png
Type: image/png
Size: 8399 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/d2n-analysis-talk/attachments/20110312/3d1b82de/attachment-0004.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: phy_asys.png
Type: image/png
Size: 8806 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/d2n-analysis-talk/attachments/20110312/3d1b82de/attachment-0005.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: apar_aperp.png
Type: image/png
Size: 8735 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/d2n-analysis-talk/attachments/20110312/3d1b82de/attachment-0006.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: a1_in_context.png
Type: image/png
Size: 13453 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/d2n-analysis-talk/attachments/20110312/3d1b82de/attachment-0007.png