On the use of CHARYBDIS
                     
                         T.Reichelt
                         
                          7.7.2000
                     
     In the quasielastic scattering of polarized electrons on the
     deuteron,the emerging neutrons carry a polarization in the 
     scattering plane:the sideways polarization Px,which is pro=
     portional to the ratio Gen/Gmn and the longitudinal polariza=
     tion Pz.Passing the field of the Charybdis magnet a spin pre=
     cession occurs. 
     The NPOL polarimeter is sensitive to the sideways projection
     of the polarization vector.The spin precession has the result 
     that the measured asymmetry is proportional to a linear mixing 
     of Px and Pz.
     
     The Gen experiment consists in measuring two asymmetries with 
     two different field settings,i.e. two different precession ang=
     les.The ratio of these two asymmetries is also the ratio of the 
     underlying polarizations P1 and P2,provided beam polarization 
     and analyzing power of NPOL did not change.In this sense the 
     experiment is self-calibrating.
     
     Being eta = P1/P2 and X1 and X2 the corresponding precession
     angles,g=Gen/Gmn can be calculated from these quantities:
     
     
         g = K*(eta*sin (X2) - sin (X1))/(cos(X1)-eta*cos(X2))
         
     where K is a kinematical factor given by the electron kine=
     matics alone. 
     
     If X1=0 and X2=90 degrees,g=K*eta and we get the situation de= 
     scribed in the proposal.It should be noted here that the error
     propagation of X2 to g disappears in first order,if X2 is chosen 
     to be 90 degrees +- X0,where tg(X0)=Px/Pz,the sign depending of
     the sense of rotation.In this case the complete polarization vec=
     tor - as composed of Px and Pz - is precessed to the sideways di=
     rection,not Pz alone as in the case of X2=90 degrees.For the 
     Galster parametrization X0 is close to 11 degrees,nearly indepen=
     dent of Q**2.

     One reason to run the magnet at two different fields both being
     different from zero could be the possibility to prevent charged 
     particles from entering NPOL,i.e. by using Charybdis also as a 
     sweeping magnet.On the other hand the two precession angles must 
     always clearly be different from each other,otherwise the feature
     of self-calibration is lost.
          
     The flux of charged particles entering NPOL consists mainly of 
     pions,muons as their decay products and electrons/positrons as 
     remnants from the electromagnetic shower emerging from the Pb ab=
     sorber situated in front of the magnet.A closer look to the cor=
     responding spectra shows that if for instance charged particles 
     up to energies of 400 Mev for the Q**2=1.15 GeV**2 point could 
     be preventd from entering NPOL,the pion fluxes would be reduced 
     by a factor of ten,the muon fluxes by a factor of hundred and the 
     electrons/positrons would almost disappear.

     Assuming that the magnetic gap of Charybdis acts as a collimator 
     for particles emerging from the target,a field length of 0.48 Tm 
     is estimated to be sufficient for setting a threshold of 400 MeV 
     and removing to a large extent charged particles from the NPOL 
     entrance.The price to pay is an additional sytematic error in g 
     due to the error propagation from X1.
     
     In the following table five different configurations of X1,X2 are
     chosen for Q**2=1.15 GeV**2 and the Galster parametrization is used.
     A field lenth of 0.48 Tm corresponds to X1=21 degrees.dg/deta is a 
     measure of the propagation of eta to g.The last column shows the 
     systematic error in g due to the angular errors in first order.Here 
     a relative error of 1% is assumed for the integrated field length 
     and the precession angles.
     
     
     Nr.      X1       X2       eta       dg/deta        delta g/g 
     
     1         0       90      -0.19        0.60           0.3 %
     
     2         0      101      -0.19        0.63           0.0 %
        
     3        11      101       0.0         0.62           1.0 %
     
     4        21      101       0.16        0.63           2.1 %
     
     5        41      101       0.50        0.72           3.9 %