Cathode Preamp-Shaper

For the CMS endcap chambers, the cathode strips are in the radial direction and they provide the measurement of the bending (azimuthal) coordinate. This measurement relies on the interpolation of charges induced by the avalanche process on several (4-5) neighboring cathode strips (hereafter referred to as a cluster). The resulting position resolution is proportional to the percentage error of the cluster charge measurement. For chambers with strip width equal to the gas gap (twice the anode to cathode distance), this resolution is approximately equal to wD(Q)/Q, where w is the strip width, Q and D(Q) are the total cluster charge and it's error. For a cluster with N strips, D(Q)=sqrt(N)d(q), where d(q) is the noise of each strip channel. In the absence of RF pickup noise, d(q) is dominated by the electronics noise of the FET in the preamp.

To achieve the required per layer resolution of 150 micro-m for ME1/2 (=0.5cm) and 300 micro-m for rest of the endcap chambers (=1.0cm), the total cathode charge must be measured with an accuracy of <3%. The design goal is to achieve 1% in d(q)/Q for the entire endcap muon system. This calls for a state of the art low noise amplifier.

The intrinsic noise for the preamp-shaper depends primarily on three parameters: the input (strip) capacitance, the size (width and length) of the preamp FET and the shaper peaking time. Anticipating the high rate environment at the LHC - the estimated neutron and charged particle hit rate can be as high as 1 kHz/cm**2 at the designed luminosity of 10**34/s/cm**2 - the shaper peaking time is chosen to be 100 ns for fast recovery and for good two-pulse time resolution. These considerations together with practical choices for the FET size lead to the requirement that the rms equivalent input noise be less than 7000 electronics for the strip with the highest capacitance in the system (250 pF), or 25 e's/pF.

The required minimum signal to noise ratio of 100 can be met by operating the chamber with a gas gain such that the induced cathode charge by a normally incident minimum ionizing particle is about 7x10**5 electrons, or 112 fC. The overall DAQ system gain is 0.5 fC per ADC count. The electronics noise from the largest chambers will result in an ADC pedestal rms width of 2.2 counts and the peak of the Landau distribution will be located at around 250 counts (8 bits). The dynamic range is 4000 (12 bit ADC), which is necessary to accommodate the tail of the Landau distribution and variations due to chamber and amplifier gains. The gain of the trigger output of the preamp/shaper may be set higher in order to minimize any effects due to comparator offsets. These specifications are listed in the following table.

                  Cathode Preamp-shaper Specifications

Equivalent Input Noise (rms)            Cstrip x 25 e/pF+700 e (7000 e at 250 pF)
Shaper Peaking Time                     100 ns
Peaking Time of Shaped Chamber Pulse    170ns
Shaped Waveform:                        5 pole semi-gaussian w/ tail cancellation
Nominal Input Charge (i,ii)             112 fC  (=7000x100(S/N))
Preamp-Shaper Gain                      0.9 mV/fC
Non-Linearity:                          < 1%  0 - 1.5 volt
Dynamic Range:                          12 bits
Overall System Gain:                    0.5 fC/ADC count
Two Track Time Resolution:              125 ns
Gain Variation:                         <4% channel-to-channel <10% chip-to-chip
Calibration Precision:                  0.5% over full range

     Footnotes:

     i)  Corresponding to the average ionization deposition (Landau peak) for a
     normal incident minimum ionization particle. The operating voltage or the
     chamber gas gain corresponding to this input charge should be determined
     empirically.

     ii) The gas gain required is 700,000/(N_ion x f_att x f_ind x f_sh),
     where N_ion is the number of ion pairs generated by the traversing
     particle;   f_att is fraction of drift electrons lost  due to attachment
     to chamber gas molecules, f_ind is fraction of the anode charge induced on
     either cathode; and f_sh is fraction of avalanche charge collected by
     anode for a given shaping time. Simulation (CMS TN/95-194) of CSC response
     gives  N_ion = 180 (for 10 mm gas gap); F_att = 0.5;  f_ind = 0.41; f_sh
     (100ns shaper peaking time) = 0.19.  Based on these numbers, the gas gain
     is estimated to be about 10**5.

The cathode preamp-shaper ASIC ( Buckeye chip) has gone through 5 submissions. Each of the submissions is a four channel prototype ASIC. The foundry used is MOSIS-HP and the process is 1.2 micron CMOS with linear capacitors. The test result of the latest submission shows that the Buckeye chip has met all of the design requirements.