Minutes of the CALCOM meeting, January 16, 1998. =============================================== Agenda: K. Griffioen - Status of the CLAS Torus field analysis K. Mikhailov - Status of the straight tracks analysis for drift chamber alignment L. Elouadrhiri - TOF timing using the beam RF structure ------------------------------------------------------ The areas covered represent 3 areas where further progress is most pressing. I think it is fair to say that in all 3 areas significant progress has been made recently. This is especially true for the TOF counter timing analysis presented by Latifa E.. The TOF system calibration seems now well on its way to soon come close to its design values. For the magnetic field analysis it was decided that the William and Mary group now begins to tune the magnetic field simulation to agree with the measurements. The impact of this on the reconstruction of kinematical invariants such as the elastic peak position will provide important feedback to the W&M group for their analysis. The drift chamber analysis has also made significant progress. There is clear evidence for some misalignments in the system. Clearly, mmuch of our data analysis now hinges on the accurate understanding of the drift chamber geometry and their positioning in the magnetic field. Volker Burkert ========================================================================== Individual reports: ------------------- Keith Griffioen and Hovanes Egiyan ---------------------------------- I. Overview A. Hall probes are well-understood 1. planar Hall effect is properly calibrated B. We have a credible simulation program 1. rewritten by Hovanes 2. calculates magnetic fields and the Hall voltages for calibrated probes. 3. we compare measured and predicted Hall voltages because of the complicating planar Hall effects. C. We have a good mechanical model of the device 1. rms variations between survey data and model are 1.5-2 mm. 2. model has 14 parameters: 6 azimuthal angles (one for each cryostat), 2 calibration constants each for radial and polar motion (i.e. linear calibrations), 1 offset of the tube track from the pivot point, and 3 coordinates locating the mapper's rotational center. 3. survey data for each carbon-fiber tube are rotated through the azimuthal angle that aligns the plane of motion vertically. The offset of this plane yields x_0. The transformed data are fit to the equations: z = z_0 - s \sin\theta + r \cos\theta y = y_0 + s \cos\theta + r \sin\theta r = a_r E_r + c_r \theta = a_t E_t + c_t in which z_0, y_0, a_r, c_r, a_t, c_t and s are fit parameters and E_r (E_t) is the encoder value for radial (polar) motion. 4. some improvement is possible using separate coefficients for each cryostat (roughly 0.5 mm decrease in rms values) 5. no improvements are seen from quadratic calibrations of the radial or polar motion. II. The Problem A. Z_0 is roughly -1.5 cm. 1. the mapper origin is about 1.5cm behind the survey origin. 2. survey origin is 3260mm behind the front face of cryostat as specified in drawing 66210-E-01738. 3. this shift is to be expected: the mapper center of rotation was simply not at the target position. B. "Coil-Locator Maps" show a 2-3cm shift between simulation and data. 1. we scan radially outward and find the maxima in the magnetic field which should occur at the positions closest to the center of the coils. 2. the same procedure of finding field maxima is applied to the data and the simulation. 3. these two sets of points, which outline the shape of the coils in the cryostat, can only be made to lie on top of each other with drastic shifts of 2-3cm! III. Possible Causes A. There's a bug in the simulation (unlikely) 1. original program has been checked elsewhere 2. old and new versions agree B. There's a mistake in the survey calculations 1. survey data are reckoned for the probe position using 4 sets of cross hairs stuck on the carbon-fiber tubes and a slant-view target in the probe position. Absolute coordinates were calculated from shots of the CLAS frame. 2. one must go back to the raw data to check for errors C. There's a difference in survey origin and cryostat origin. 1. drawing 66100-E-0463A shows the cryostat origin 2. this origin was designed to be the same as the origin in drawing 66210-E-01738 3. Bernard claims it's unlikely that the two differ 4. It's difficult to find out for sure now D. Mechanical instability of the mapper 1. We did have problems with the device, and mechanical uncertainties are our largest source of errors. 2. Our problems were largely in azimuthal uncertainty. It's unlikely that a 2-3cm shift along the beam axis results from any mechanical problem with the mapper. E. The coils aren't where they're supposed to be IV. A Solution? A. First we must go back to the raw data of the survey and check all details. B. Second, we will modify the coil positions drastically (2-3cm) as our data shows and generate a field map accordingly. If this improves the tracking in CLAS, then Cause III.E becomes more probable. K.Mikhailov:Drift chamber alignment issue. ----------- The time to distance function have been extracted from the September's run 6112(B=0) and presented for region 2 and region 3 of drift the chamber. Spatial resolution is about 250 microns for the linear part of time to distance function for both regions. Time based reconstruction of straight tracks was used for the drift chamber alignment. We calculate angles for the projection of the track to the mid-plane of drift chamber. Three parameters defined geometry of drift chamber. There are angle theta minimum with respect to the direction of beam, X and Z coordinates. All parameters are considered in Bogdan Center(B.C.) reference frame. We used fast electron identification(eid0 code written by A.Vlassov) for sampling elastic events. We need it for better separation of low theta tracks(electron) and high theta tracks which corresponds to elastic scattered proton.Low angle tracks was analyzed to adjust X coordinate of B.C. while high angle track provides alignment for Z one. Alignment was produced in a three steps. First step is the minimization of the angle between parts of a track which are independently fitted in the region 2 and 3. Theta min of the region 3 is fixed at that step and we define correction for the theta minimum of the region 2. On the second step is defined corrections for X and Z of each region. We use combine tracking (region 2 and region 3) to check these values and to obtain mean values for X and Z near zero. Third step is the correction for the region 3 theta minimum, which was fixed on first step. We rotate region 3 around B.C. on the small angle and repeat first and second steps. These three steps give us the minimum misalignment and we had three corrections(theta minimum, X and Z coord. of B.C.) for two regions. We plan to continue alignment procedure. Next time will include region 1 alignment. First preliminary results of alignment for region 2 and region 3 sector 1 and 4 in mid-plane is presented(see table 1). Table 1. B.C. shifts REGIONS #2 #3 Sec 1, X[cm] 0.870, 1.188 Sec 1, Z[cm] -0.349, -0.341 Sec 1, theta[degrees] -0.152, 0. Sec 4, X[cm] 0.803, 1.090 Sec 4, Z[cm] -0.149, 0.100 Sec 4, theta[degrees] -0.090, 0. Analysis was made by stand alone DR package(Nikolai's code) in RECSIS framework for September's run 6112(B-field = 0, full target). We see significant improvement for vertex reconstruction using these parameters in comparison with ideal geometry(see /home/kmikhail/CALCOM/fig1.ps) N.Pivnyuk and A.Stavinski took part in this analysis too. Latifa Elouadrhiri, Volker Burkert and Stepan Stepanyan ------------------ Rf structure and TOF calibration We Studied the beam-time structure during the December e1 run at 2.4 GeV. Using the difference between each TOF counter and the rf signal from the injector. After correction of the TOF TDC calibration with a factor of 1.077, we obtained a very clear 2ns structure of the beam. From the beam structure we determined the TOF resolution of a single TOF counter to be of an average of 140 ps. We used the rf structure to determine the constant delay of each counter with respect to the rf. This constant delay is the sum of two terms, one is multiple of 2ns, due to the 2 ns structure of the beam, and an additional term of less to than 2 ns. In a first step we used the rf structure to determine the term below 2ns with a precison of about 50ps for the panel 1 (electron) strips. Then using inelastic data (scattered electron together with produced hadron) we determined the term of multiples of 2ns for each counter. Using this procedure we measured the calibration constant for each counter in each one of the four sectors: 1,4,5 and 6. the procedure shows an excellent results, a very good separation between proton and pion and a good signal for Kaons.