Available on CMS information server CMS NOTE 2003/000
May 1, 2003
The validation systemfor the Barrel Muon Chambers at LNL- The cosmic setup -
M. Bellato, E. Conti, M. De Giorgi, F. Gonella, A. Meneguzzo, M. Passaseo, M. Pegoraro, S. Vanini, S. Ventura
Dipartimento di Fisica and INFN, Padova, Italy
L. Berti, M. Biasotto, E. Ferro, M. Gulmini, G. Maron, N. Toniolo, L. Zangrando
INFN, Laboratori Nazionali di Legnaro, Italy
Abstract
The first design of a Barrel Muon Chamber composed of staggered layers of drift cells started in Padova in 1993 [1]. The Chambers were designed for the CMS experiment that will run in the LHC accelerator at CERN. The full production of these Chambers started two year ago in three European laboratories, and about 70 are being built in a large hall of the Laboratori Nazionali di Legnaro of INFN (LNL).
The INFN section of Padova is in charge of working out the mechanical design [2], the FE electronics and LV system [3], the HV system and the muon local trigger design [4] for the final chamber setup.
Moreover, before the Chambers are shipped to CERN, it is necessary to test and validate their overall functionality and stability, for which a proper test system has been designed and implemented.
This note describes the experimental setup of the chamber test area at LNL and the cosmic-ray acquisition and monitor system for the on-line data analysis.
1 Introduction
The first design of a Barrel Muon Chamber based on staggered layers of drift cells started in Padova in 1993 [1]. After eight years, the full production of the Barrel Chambers for the CMS experiment for the LHC accelerator at CERN finally started in 2001 by three European collaborators, namely Padova (sez. INFN), Aachen (RWTH), and Madrid (CIEMAT).
The muon chambers of MB3 type that form the third shell of the detector, and some special MB4 for the fourth shell are built in a hall of the Laboratori Nazionali di Legnaro of INFN. Before being sent to CERN, chambers must be tested in all functionalities for validation: after preliminary checks, a test with cosmic rays is necessary, because cosmic rays runs reproduce real working conditions and allow the on line recovery of construction problems.
A chamber for the Muon Barrel Detector [2] [5] consists of 12 layers of rectangular drift cells with a pitch of 42 mm by 13 mm. The cell layout aims at measuring the drift time of ionization electrons produced by incident muons. The specific electric field shape is assured in each cell by 4 electrodes on the walls and by the central wire anode; three different HV are needed to bias the cell and provide the linearity of the space-time relationship of drifting electrons. Each wire is read out by a specific designed FE chip housed in compact boards of 16 or 20 channels to accomplish the different chambers size [3]. The smallest chamber (MB1 type) has ~620 readout channels and the bigger (MB3 type) ~800.
The 12 layers of tubes are staggered by half a cell and grouped four by four to form 3 independent modules named SuperLayers (SLs). The two external SLs will be devoted to the precise measurements of incident particle in the bending plane and together should yield 100 micron precision in position and 1 mrad in direction; the other SL will measure the coordinate along the beam line. The wires information will be used in the first level trigger [6] so the relative position of each layer and of the wires in each layer should be less then one hundred micron. The chamber is assembled by gluing the 3 SLs together with an aluminum honeycomb plate to ensure the required stiffness and increase the lever arm between the two external SLs.
To assure a constant and precise drift velocity inside the cell for ionization electrons the ArCO2 gas mixture must be well-defined and highly pure; moreover, the chamber must work at a steady and controlled pressure.
The production and functionality of the chamber under construction must be monitorized. The construction and test of a large number (70) of chambers implies the implementation of robust and reliable tests to be used for all the 3 years foreseen for the production [7].
The same data acquisition and monitor system has been used in the test beam runs performed at CERN.
2 Requirement and the LNL cosmic setup.
The setup allows to verify the correctness of the following items once each chamber is assembled:
- electrical connections
- noise rate
- electronic performances
- cables and electronic delay
- detection efficiency
- time resolution
- uniformity of chamber response
- layer (and superlayer) mechanical accuracy
All tests must be performed under controlled gas and environmental conditions, defined working pressure and monitorized LV and HV settings.
The strict requirement depends on the fact that FE electronics and HV distribution to cells sit inside the gas volume, and any misfunctionality of any part may imply the opening of the SL itself. Furthermore, in order to fulfill a common quality level in all different production sites all detectors must have a minimum guaranteed performance on the items listed above.
In Figure 1 a schematic of the system is introduced. It is composed by the High Voltage power system controlled via net by a PC, the Low Voltage power system with slow control functions and connected to the same PC via serial line, the TDC modules read via net by the DAQ system, the plastic scintillators which trigger cosmic-rays, and the gas and environmental parameters monitor system. Finally, a third PC allows the visualization and on-line analysis of the acquired data.
Figure 1: Schematic of the chamber validation system.
Figure 2: Photo of the chamber test area at LNL.
In Figure 2 a photo of the test area is shown: an MB3 chamber can be seen over the metal structure also housing the groups of scintillators, while on the left we can see the rack of the control system and gas distribution, the two DAQ racks, the trigger system and the rack HV and LV power sources, finally in the middle the controlling PCs and DAQ.
A remarkable cabling and grounding job of the test area has allowed to obtain a reliable experimental setup, not affected by external noise and disturbance, which guarantees repeatability and reliability of the validation tests.
Some results about the quality of the MB3 chambers obtained from the data collected using the experimental setup described in this note are presented in [8].
3 Gas System and Environmental Monitor
The long life foreseen for the CMS detector and the necessary gas purity require a continuous gas flow supply system. In effect, for high detection efficiency oxygen concentration must be kept to a minimum and the same goes for nitrogen, in order to maintain stable the drift velocity inside the cell. Oxygen and nitrogen come from environment diffusion, so the measurement of the first (easier) automatically informs about the second.
Also, all validation tests must be performed under controlled gas flow and environment conditions (temperature, atmospheric pressure) and at defined and stable working pressure.
A whole rack housed the gas distribution and control system and the monitor of the environment parameters.
The gas employed is a mixture of 85% argon and 15% CO2 available in premixed bottles of 50 l at 200 bar, with a total impurity concentration certified below few ppm.
3.1 Gas distribution and control system
The gas system for chamber tests consists of a line with maximum flow of 2 l/min with electronic control and two emergency valves in order to limit overpressure to maximum 50 mbar. The complete outline can be seen in Figure 3. The mixed gas enters a Bronkhorst flow regulator [9], and is sent to the chamber after a safety valve. The second safety valve is placed downstream the chamber. The gas exhaust line is splitted in two. The first main branch goes into a regulator valve [9] by means of which we can control the absolute gas pressure in the chamber and keep it steady during data acquisition. The second line feeds the oxygen meter.
Figure 3: Schematic of the gas distribution and control system.
The oxygen measurement is made using a Zirconia Oxygen Analyzer [10]. The Zirconia sensor measures the oxygen content in the gas and emits a mV signal: the cell voltage rises logarithmically as the amount of oxygen in the gas falls, allowing the measurement of very small amounts (down to fractions of ppm). A Digital Display Unit amplifies and linearizes the logarithmic output of the Zirconia sensor.
The system also implements a second line with a manual flowmeter for leakage tests.
3.2 Environmental and gas monitor
A system of analog sensors monitors all interesting environmental parameters: temperature, humidity and pressure. A custom rack box powers the sensors, amplifies output signals and transmits them to the 16 input-ADC board with 12-bit resolution (National Instrument PCI6023 [11]).
The same electronic box receives signals from flowmeters, pressure control valve and the oxygen meter, which after conditioning and buffering are sent to the acquisition board too.
All information is published by means of an Apache Web Server by a software developed with Labview that regularly reads the analog channels of the ADC board.
In Figure 4 below the schematic of the system is shown.
Figure 4: Schematic of the environmental and gas monitor system.
4 HV Power system
The high voltage system is based on a CAEN SY1527 [12] power supply system housing an A876 master board that controls an A877 remote board. The system used is shown in Figure 5 below.
Figure 5: Schematic of the chamber HV power and distribution system.
The A876 master board is able to control and supply up to four independent A877 remote boards. The board delivers four groups of high/low floating voltages: a positive high voltage in the range from 0 to 4200 V, a negative high voltage in the range from 0 to –2200 V, and a dual low voltage for remote boards powering. A common communication bus links the remote A877 to the A876 master board.
The A877 board delivers high voltages to the electrodes (anodes, cathodes and strips) of the muon chambers. The board provides 12 groups (MacroChannels) of floating HV outputs, each supplying two anode lines (up to +4 kV), one strip line (up to +2 kV) and one cathode line (down to –2 kV) per each layer. Anode, strip and cathode voltages are linearly programmable with 12 bit resolution; they are independent, but the anode voltage cannot exceed the corresponding strip voltage by more than 2 kV. Each macrochannel powers one layer of Phi or Theta SL.
Table 1: Channel characteristics of the A877 remote board.
Max output current (short circuit) / 100mAVoltage ripple / < 10mV
Voltage monitor accuracy / ± 0.5V ± 0.1% of setting
Current monitor accuracy / ± 10nA ± 3% of setting
The three cables out of the A877 board enter a junction box where they are split and adapted into a couple of HV custom connectors for each SL in order to match the internal granularity that is:
· 1 connection per 8 cells/layer for Anodes
· 1 connection per 16 cells/layer for Strips
· 1 connection per 16 cells/layer for Cathodes
The system is controlled via Ethernet, through the TCP-IP protocol and with a Windows NT PC. A software, provided by CAEN, allows access to all parameters, monitor voltages and currents, while saving the values on files.
5 LV Power and Slow Control system
The low voltage (LV) system has been specifically designed and realized in a custom rack box to meet the front-end electronics (FE) requirements: linear and floating power sources, noisily reference voltages for the thresholds and the complete slow control functionality to monitor and set FE parameters.
A PC controls the whole module via serial interface.
5.1 LV power and slow control hardware implementation
The system (Figure 6) is composed of two functionally and electrically independent sections: a low voltage source section to power the front-end electronics and a slow control interface.
The low voltage source section is composed of linear and modular powers, which supply the 5 V and 2.5 V required by front-end electronics boards. A further couple of modules generates the +5 V and –5 V, which power the test pulse box (see section 6). All power sources are equipped with crow-bar protection and output current limiters. The voltage is adjustable to accomplish the voltage fall along the cables. All voltages are floating and the reference ground point is the chamber aluminum structure.
The monitor and slow control system are implemented by a custom board based on a PIC 16C73 microprocessor [13] that realizes the I2C interface, integrates an ADC, and supplies an 8 bit logical port with I/O direction of each line individually configurable. The board generates moreover the threshold voltages for the FE electronics of the three SLs, using an 8 bit DAC with I2C interface and operational amplifiers for buffering. A common 1.5 V reference voltage is generated together with the independent thresholds with 100mV of full scale. In order to minimize all possible interferences the I2C bus of the PIC is optoisolated in both directions. A custom board supplies this subsystem with the necessary power.
The system is controlled by a PC running Windows NT via serial interface; obviously, the serial bus is also optoisolated, guaranteeing optimal isolation of the whole module.
Figure 6: Schematic of the Low Voltage power and Slow Control system.
5.2 The slow control software
A C++ program running on a desktop PC using Windows operating system controls all test operations on front-end electronics. On startup it finds the current hardware configuration, i.e. the number of 16 or 20 channels FE boards in each superlayer, which correctly respond to the computer request. Each board is represented by a led, whose colour represents its status. A configuration text file is used to set the serial port, data directory, PIC reference voltage, and threshold setting.