CERN 99 Test Beam results

The test beam layout is given in the P2" 1999 CERN Test Beam Analysis Home Page (Setup) . The beam chambers BCX1-BCX3 were used to measure the X (horizontal) position of the 100 GeV/c muon track in each layer of P2". The resolution of the X plane of each beam chamber is about 0.06 cm (sigma). Using BCX1 and BCX2, one can predict the X position of the beam track in P2" with an accuracy of sigma = 0.26 cm. The position of the same track in BCX3 also can be predicted and compared with the measured one in BCX3. The corresponding residual BCX3(pred-meas) and BCX3 pull=residual/sigma(pred-meas) are plotted in Fig.1. Note that the fitted pull distribution has a standard deviation equal to 1, confirming that the single plane resolution is on the order of 0.06 cm. The cut -1.0 cm < BCX3(pred-meas) < 1.0 cm was imposed to select events for analysis. The angular distribution of the beam track (in the horizontal plane) and the X-distribution of the track in layer 3 of P2" after the cut are presented in Fig.1, as well. Since the

width of the anode wire group is about 5 cm, the beam barely covered 2 wire groups. The occupancy distributions for the wire groups in each layer of P2" (without the cut on the BCX3 residual but demanding the anode time be in the TDC window 2500 < TDC < 2750) are given in Fig.2. For the efficiency measurement one wire group (#23) from each layer was chosen as having maximum occupancy in each layer. The beam track was required to point to this wire group within a distance +- 1.5 cm from the center to ensure that the track did not miss it. The efficiency was defined as the fraction of such events having a hit from the wire group # 23 in the TDC window 2500 < TDC < 2750 (see Fig.3 where the TDC distribution is presented for wire group # 23, for all layers). This window corresponds to a 125 nsec time interval. Note that the majority of the data in the CERN '99 Test Beam Run was taken with the TDC range extended to 32 µsec. And one more piece of information - the CSC gas was Ar:CO2:CF4 40%:50%:10%.

- The results of the P2" single layer efficiency measurements for different conditions are presented below. An example of the efficiency and inefficiency curves vs HV shown to the right are from Fig.4b_3. The efficiency curve fit was done using the function:

Eff = 100. / (1. + exp( -(x - HV95)/B + C))

where C = ln(0.05/0.95), and the fitted parameters are the slope B and HV95 - high voltage, corresponding to an eff=95% ("knee point"). The values of the parameter HV95 are compared at different conditions (see below). Note that the actual HV settings were slightly different from the nominal ones used here. Their departures from the nominal HV were (in kV) -0.02, 0.02, 0.0, -0.01, 0.0, and 0.02, corresponding to layers 1-6. In the beginning of data taking, the actual HV settings were adjusted to have approximately the same amplitude of signal from the cathode front end. The parameter HV95 allows us to do this as well, but using information from the anode front end (on the condition of having one and the same anode front end thresholds).

The two thresholds used,Qth=20 fC and 30 fC, are nominal. Their actual values could be off by +- (5-10) fC. More accurate calibrations will be done later. Two different thresholds were used to see the shift in the value of the "knee" HV point. The 16 ch boards, mentioned below, have two 8 ch chips on them.

- Efficiency for 96 ch board, HV scan, GIF off, trigger-scint.
(HV = 3.1-3.9 kV, Qth=20 fC, Runs 440-478) :

Fig.4 - efficiency and Fig.4a -inefficiency, all layers
Fig.4b_1 - layer 1, Fig.4b_2 - layer 2, Fig.4b_3 - layer 3
Fig.4b_4 - layer 4, Fig.4b_5 - layer 5, Fig.4b_6 - layer 6
HV95 (kV, layers 1 - 6) = 3.31 3.25 3.23 3.29 3.30 3.31.

- Eff. for 16 ch board, HV scan, GIF off, trigger- CLCT(4/6)
(HV = 3.2-3.8 kV, Qth=20 fC, Runs 669-675) :

Fig.5a - efficiency and Fig.5b -inefficiency, all layers
HV95 (kV, layers 1 - 6) = 3.26 3.37 3.35 3.30 3.27 3.32.

- Eff. for 16 ch board, HV scan, GIF off,trigger -
CLCT(3/6)*scint., (HV = 3.3-3.8 kV, Qth=30 fC, Runs 717-722) :

Fig.6a - efficiency and Fig.6b -inefficiency, all layers
HV95 (kV, layers 1 - 6) = 3.36 3.40 3.42 3.40 3.34 3.39.

- Eff. for 16 ch board, HV scan, GIF K=1,trigger -
CLCT(3/6)*scint., (HV = 3.3-3.8 kV, Qth=30 fC, Runs 705-710) :

Fig.7a - efficiency and Fig.7b -inefficiency, all layers
HV95 (kV, layers 1 - 6) = 3.40 3.46 3.47 3.46 3.34 3.45.

- Table 1.The parameter HV95 vs Layer # for different conditions (see also Fig.8 at the right) :

96 ch board Qth=20 fC
Trigger Scint.
GIF OFF

16 ch board Qth=20 fC
Trigerr CLCT(4/6)
GIF OFF

16 ch board Qth=30 fC
Trigger CLCT(3/6)*Scint
GIF OFF

16 ch board Qth=30 fC
Trigger CLCT(3/6)*Scint
GIF K=1

- At Qth=20 fC, the plateau for the 16 ch boards starts at HV=3.35-3.40 kV. At Qth=30 fC, it starts at 3.36-3.42 kV.

- At GIF K=1 (maximum GIF intensity) the HV required to provide 95% efficiency is higher by about 50 v than the corresponding HV with GIF OFF. Is this a shift in the thresholds of anode (cathode) front end ? There is no such difference in layer 5, but this is due to the failure of the fit caused by a lack of data at HV=3.2 kV in these runs, see Fig.7a. Also with GIF K=1 the efficiency on the plateau is not 100%, it is less by 2-3%, see Fig.7b. This can be explained by the anode front end dead time of 100 nsec.

- The fact that all data were taken with the CLCT (some - with ALCT+CLCT) in the trigger could bias the efficiency measurements at low HV. It also makes it impossible to measure the efficiency of the CLCT (+ALCT) trigger itself. Instead, it would have been more desirable to have the CLCT and ALCT trigger decisions as latches, recorded by TDCs, and to trigger the data by the conventional trigger from the beam scintillation counters.