Results on the LED Pulser System for the Hall A DVCS Experiment

Transcription

1 Results on the LED Pulser System for the Hall A DVCS Experiment Fernando J. Barbosa, Pierre Bertin Jefferson Lab 28 February 2003 System Description The LED Pulser System Diagram is shown in figure 1. Four LEDs and associated electronics are mounted on a plug-in board designed for easy replacement. Three of the LEDs are operated in a pulsed manner via a 3-bit binary counter while a fourth LED can be turned on continuously. A programmable clock source resides in the VME module and a narrow pulse of less than 10 ns is regenerated locally in the LED Pulser Board. In this way, the fast characteristics of the LED drivers are not affected by the length of cable between the VME module and the LEDs. Figure 1 The LED Pulser System Diagram The impact of background noise on the Photo Multiplier Tube s current (PMT) in a real experiment can be approximated by operation of the continuous LED. The 3-bit binary counter drives the three pulsed LEDs in a sequence providing for synchronously additive optical intensity at the PMT under test: LED1, LED2, LED1+LED2, LED3, LED1+LED3, LED2+LED3, LED1+LED2+LED3, all off. The LED drivers provide enough current sinking capabilities to take advantage of the full dynamic range of the LEDs. FACT technology was chosen for being radiation tolerant and for its high 1

2 performance. The optical intensity of each of the LEDs is governed by the bias voltage supplied from independent high stability, 12-bit DACs residing in the VME module. The BASE circuit board, in addition to holding the LED plug-in board, also contains a cold cathode fluorescent lamp (CCFL) and associated high voltage AC supply inverter. This lamp is only used when annealing of the PbF2 crystals is required. During this process, the HV supplies to the PMTs need to be turned off. To insure inadvertent powering of the CCFL during normal operation, an interlock input on the VME module and software controls are provided for complete user control. An auxiliary output and a status output are also available on the VME module: the status output can be used for control of the HV system and the auxiliary output is a general purpose TTL control port. The LED plug-in, Base and Transition circuit boards are mounted on a moving X- Y-Z stage that positions the LEDs and the CCFL in front of each of the PbF2 crystals. The PbF2 crystal array contains 11 x 12 individual crystals, each measuring about 30 mm x 30 mm x 180 mm, and each crystal is coupled to a Hamamatsu R7700 U PMT and associated amplified base. Optics The NSPBF50S LEDs are manufactured by Nichia and generate blue light at a nominal wavelength of 470 nm and 20 ma of forward current. This LED was chosen for its compact size and low intensity, which is required to prevent saturation of the PMT and base assembly. The output intensity of LEDs varies and Nichia characterizes its products into three nominal luminous intensity ranks: rank T for 232 mcd, rank S for 160 mcd and Rank R for 116 mcd. The LEDs used for this report belong to rank T. The relative luminosity can be varied by a factor of about 2.7 for forward current values ranging from 1 ma to 100 ma which correspond to forward bias voltages ranging from about 2.7 V to 5 V. Continuous forward currents must be limited to 30 ma to prevent damage to the LED, though. The temperature dependence of the relative luminosity is less than 0.2%/ C from 30 C to 85 C and at a nominal forward current of 20 ma. Neutral density filters are employed to minimize saturation of the PMT and base assembly and preserve the useful dynamic range of the LEDs. These neutral density filters have a uniform pass band covering the visible spectrum and are gel-film Wratten type by Kodak. Several tests were performed with filters of optical densities of 2.0, 3.0 and 4.0, corresponding to optical transmissions of 1%, 0.1% and 0.01%, respectively. Results Our tests were performed with a PbF2 crystal coupled to a Hamamatsu R7700 PMT and bias and amplifier base. The Wratten neutral density filters were placed between the crystal and the LEDs. Figure 1 shows the LED driver pulse at full bias of 5V. The pulse is 8.8 ns wide. Figure 2 shows the resulting PMT signal for a single LED at 5V bias, the PMT at 569V and with a Wratten filter with an optical index of 3.0. Note that the maximum output from the PMT base assembly is 2V. 2

3 Figure 1 LED Driver at 5V Bias Figure 2 PMT signal at saturation Figure 3 shows a single LED behavior at 2V bias, which is below the recommended level specified by Nichia. This shows that below the specified minimum bias level, multiple amplitude pulses are produced and as shown on the histogram on the left of figure 3. The PMT was biased at 689V and a Wratten filter with an optical index of 3.0 was used. Figure 4 shows the response from a single LED at a bias of 5V, PMT bias at 554V and with a Wratten filter with an optical index of 3.0. The full dynamic range of the system is being exercised. Figure 3 LED Under-Biased at 2V Figure 4 LED=5V, HV=-554, OD=3 For proper calibration of the PbF2 calorimeter, it is necessary to have stable output distributions for different number of photoelectrons. For a Gaussian distribution, the mean amplitude (µ) is represented by the number of photoelectrons (N) and sigma or RMS (σ) by the square root of N: σ/µ = N/N = 1/ N (1) For calibration purposes, it is desirable to have markers at 1000 and 2000 photoelectrons (γ e ) and from equation 1: 3

4 1000 γ e : 1/ N = 1/ (1000) = = 3.16% 2000 γ e : 1/ N = 1/ (2000) = = 2.23% The results presented below will be helpful in determining the best set of conditions required for proper calibration of the PbF2 calorimeter. The mean (µ) and the RMS (σ) measurements were directly obtained from the oscilloscope statistical facilities for a large number of samples. Table 1 shows the results for a fixed PMT bias of 690V with a Wratten filter with optical index of 4.0. The LED bias is varied between the specified limits suggested by Nichia. Figure 5 shows the dependence of the number of photoelectrons on the LED bias. Note that very few photoelectrons are available with a filter with an optical index of 4.0. The PMT high voltage value of 690V has been previously determined to provide good performance from this particular PMT. Table 1 System Response with Neutral Density Filter with Index of 4.0 Single LED Operation LED bias(v) HV (V) Mean(µ) Sigma(σ) %(σ/µ) # γ e Findex Number of Photo-Electrons PMT Bias = -690V Filter Index = 4.0 (1/10000) #PE LED Bias (V) Figure 5 Number of PhotoElectrons Versus LED Bias with a Filter Index of 4.0 Table 2 shows the results for various PMT high voltage bias and LED bias combinations with a filter index of 3.0. For each LED bias setting, the PMT high voltage bias was adjusted to provide the widest dynamic range of outputs without saturation. Figure 6 shows how the number of photoelectrons for each PMT high voltage bias and LED bias setting combination. 4

5 Table 2 System Response with Neutral Density Filter with Index of 3.0 Single LED Operation LED bias(v) HV (V) Mean(µ) Sigma(σ) %(σ/µ) # γ e Findex Number of Photo-Electrons Filter Index = 3.0 (1/1000) #PE LED Bias (V) Figure 6 Number of PhotoElectrons Versus PMT/LED Bias and a Filter Index of 3.0 Table 3 shows the results for various PMT high voltage bias and LED bias combinations with a filter index of 2.0. Again, for each LED bias setting, the PMT high voltage bias was adjusted to provide the widest dynamic range of outputs without saturation. Figure 7 shows how the number of photoelectrons for each PMT high voltage bias and LED bias setting combination. Table 3 System Response with Neutral Density Filter with Index of 2.0 Single LED Operation LED bias(v) HV (V) Mean(µ) Sigma(σ) %(σ/µ) # γ e Findex

6 Number of Photo-Electrons Filter Index = 2.0 (1/100) #PE LED Bias (V) Figure 7 Number of PhotoElectrons Versus PMT/LED Bias and a Filter Index of 2.0 It can be observed that the largest range of photoelectrons is obtained with a neutral density filter with an index of 2.0. However, using a filter with an index of 3.0 may provide better results, as the PMT high voltage bias is closer to the actual operating conditions. Also, the lower sensitivity of the number of photoelectrons to the LED bias voltage and operation of the four LEDs during normal pulser usage may be additional advantages obtained with a filter index of 3.0. Figure 8 shows the response to operation of the complete pulser system with a filter index of 3.0. All three pulsed LEDs are enabled with a bias of 4.075V and at a trigger rate of 1KHz. The PMT high voltage bias was set at 547V to maximize the output response without saturation when the three pulsed LEDs are on. The VME plug-in module provided the LED bias from 12-bit DACs. The pulses with the lowest amplitudes correspond to any single LED being on (3 out of 7 events), the middle pulses correspond to any two LEDs being on (3 out of 7 events) and the highest amplitude pulses correspond to all three LEDs being on (1 out of 7 events). The linearity of the system is evident. Figure 9 shows all the circuitry employed in the LED pulser system. The board on the left is the Transition Board, which is required for cable routing on the moving X-Y-Z stage. The board in the middle is the Base Board with the Pulser Board installed on its lower right corner. The CCFL tube is the white rod shown on right of the middle board with its high voltage inverter assembled towards the top. The cable connector is installed on the back of the board. The board shown on the right is the VME Flex I/O plug-in board, which is shown removed from the Flex I/O mother board and without the front panel hardware. Figure 8 Pulser System Operation Figure 9 Pulser System Circuitry 6