Hi Tavi,
Your vertex cut which is now employed to mine as well seems quite
reasonable. I have Q^2<0.4 GeV^2 cut as well because both models have
this restrictions. I am currently trying to emply the timing cuts.
As for the asymmetry computation, I think you need to subtract the
background not only from the denominator but also the numerator as well.
In other words,
R_+ - R_-
A_{meas.} = -------------------------,
R_+ + R_-
with,
N_++ (hyd.) N_++ (empty) N_-- (hyd.) N_--
(empty)
R_+ = ------------------ - --------------------- + ---------------- -
------------------ ,
Q_++(hyd.) Q_++(empty) Q_--(hyd.)
Q_--(empty)
R_- = ( your answer ).
I think the left side is particularly vulunarable to the background,
which caused the "shift" you have created. Please see eq. (5.9) in the
Laurens van Buuren Thesis. Eq.(5.12) is only accutate if the background
has no asymmetries. In the unpol. runs, the left side creates non zero
asymmety! I do not know why!
Thank you,
aki
Octavian F Filoti wrote:
>
> Thanks Aki,
>
> Here is how I did it:
>
> Q**2 is in the range [0.08, 0.3} GeV**2
> I used only 'basic' cuts (i.e. electron (+ Cherenkovs),
> target (vertex) position (abs(zwl)<20.&&abs(zwr)<20.) and the
> full range of energy Ee' = [0, 1]GeV. Though I used also
> Ee' = [0.25, 1]GeV (this changes the behaviour at W > 1.3 GeV).
>
> The asymmetries were calculated as follow (see Laurens van Buuren
> Thesis):
> (there is no need to add sector index L,R)
>
> A^{meas} = ((R+ - R-) / (R+ + R-)) * ((R+ + R-) / (R+ + R- - 4*Rbg))
>
> where:
>
> R+ = N++/Q++ + N--/Q--
> R- = N+-/Q+- + N-+/Q-+
>
> and for
> Rbg (background) I used only the empty-target runs (almost all of them).
>
> That again, the inclusive cross section is:
>
> d^2Sigma/dOmega_e'dE_e' = Gamma{Sigma_T + epsilon * Sigma_L +
>
> h*Pz[ cos(theta_star) * sqrt(1-epsilon^2) * Sigma_TT' +
>
> sin(theta_star) * cos(phi_star) * sqrt(2epsilon*(1-epsilon)) *
>
> Sigma_TL']}.
>
> where:
> Gamma = virtual photon flux factor
> epsilon = virtual photon polarization
> Sigma_T + epsilon * Sigma_L = Sigma_0
>
> How did you get your asymmetries?!
> We should get together and talk about it.
> Have a nice weekend
>
> -----------------------
> Octavian F Filoti
> Nuclear Physics Group
> Univ. of New Hampshire
> 9 Library Way
> Durham, NH 03824
> phone: (603)862-1220
> FAX: (603)862-2998
> email: ofiloti@unh.edu
>
> On Fri, 11 Mar 2005, Akihisa Shinozaki wrote:
>
>> Hi Tavi,
>>
>> Please never mind my previous results. I found at least four mistakes
>> in my code. Both data and simulations were not right results. I was
>> basically collecting the "garbages" for the data. Also, the
>> simulations did not took the pi+ channel into account by my mistakes.
>> I am sorry if I made another confusions for you.
>>
>> So this is my another attempt for the p(e',e) results for you. This
>> time only trigger 7 is considered. For e' selection, Cerenkov hits
>> were required for the corresponding TOF counters.
>>
>> The file, asymmetryi.ps, shows the asymmetries from all the data
>> obtained from last October to December (48 deg. only). The data are
>> compared with my MAID and SL simulations. The simulated curves assume
>> the dilution factor = 0.5 with no asymmetry offset. The data make
>> good agreement with the simulations. It *seems* to me that your data
>> do not quite match with mine in the left sector since your data do
>> not go beyond the zero as much as my current result. I think we can
>> discuss this later.
>>
>> peei_md.ps and peei_sl.ps show the data distributions of W, Q^2,
>> Ee'(lab), and, Theta^e'(lab) . The blue (red) + points are the
>> counting rate in events per Coulomb in the electron left (right)
>> sector. The thiner lines are background (empty target spectrum). The
>> thick curves are my MAID or SL simulations. The data spectra in
>> peei_md.ps and peei_sl.ps are not the same because W<1.4 GeV is
>> applied for MAID while the SL simulation requires W<1.3 GeV. There is
>> no radiative process considered and this should be one of the sources
>> of the differences between the data and simulations. Since you are
>> doing with GEANT, you could tell me more on the characteristics of my
>> simulations.
>>
>> It is quite striking that the background rates are as high as the
>> foreground, which is the main source of the error in the asymmetries.
>> I am using the all empty target runs available but the statistics is
>> still about one tenth of the foreground.
>>
>> I think that is all for now. Thank you!
>> aki
>>
>>
>>
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:32 EST