Preliminary measurements
To perform the measurements reported here, a simplified electronics system instead of the standard DAQ chain was used: the avalanche photo diode (APD) is directly connected to a Charge Sensitive Preamplifier (CSP), with a sensitivity of 0.833V/pC and a light yield of 4.4 photo-electrons/MeV, which produces an output voltage step proportional to the charge created in the APD.
Step voltage passes through a transition card (T-card) to a digital oscilloscope for the measurement and the recording of the CSP signal amplitude. Typical shape of APD's output signal (screen capture from oscilloscope) is shown on picture below.
Since it is well known that APD gain is very sensitive to the temperature, the setup includes also a system for the control and monitoring of the APD's temperature during the measurements: the liquid from a chiller flows through a pipe inside a copper plate; the APDs are placed in direct contact with this plate and their temperature is continuously monitored by a thermocouple placed on APD surface. Use of this system allows a control temperature with precision of 0.1 °C.
A series of measurements were performed to explore properties of APDs. At the start the protocol of measurements was developed. After some series of measurements, followed by analyzing results by ROOT, the optimal set of bias voltages was found: {25V, 30V, 35V, 40V, 50V, 100V, 150V, 200V, 250V, 300V, 320V, 340V, 360V, 380V}. If APD doesn't reach gain M=50 at bias voltage of 380V, additional points should be added.
To understand how precise the measurements of output signal are, 2 APDs were measured 20 times with the same temperature conditions (T=25C). After analyzing received data, the distribution of results was found: it is less then 2%.
The main task of these tests is the study of the APD's gain dependence on the applied bias voltage and on the temperature: current experimental setup provides a good control of both variables, and allows accurate measurements in different working conditions. By this procedure it is possible to recognize those APDs which do not fit the EMCal requirements and to predict the behavior of APD's output signal during the ALICE operation.
APD's surface uniformity test
To verify that output is equal for all zones of APD's surface, series of measurements have been performed. First, the output signal, produced by central area was measured. For this measurement optical fiber, connected to LED, was placed at 1mm from the surface of APD (A). The section of the fiber is a circle d=1mm. In this geometry only central area was received light, delivered by fiber. Then the output from all APD's surface was measured. In this case fiber was placed at the distance of 10 mm from APD's surface (B). In this geometry light, delivered from LED by the optical fiber, was lightened all surface of APD (5mm*5mm). The third measurement was made to analyze output signal from periphery of APD's surface. For this test the geometry of previous measurement was used, but central part of APD was covered by opaque square (3mm*3mm) and output signal was received only from peripheral area (C).
For all geometries standard testing procedure was performed: 16 values of bias voltages from 25V to 420V were applied, corresponding output voltages were recorded and gain values for each value of bias voltage were calculated.
The results for first two measurements are very close: the difference in gain at all voltages is less than 1%.
The variation for third measurement is about 3%.
It was caused by very low output signal: it is comparable with noise in low voltage region, which defines the unitary gain and therefore made all results for this measurement less precisely.
Conclusion: all areas of APD's surface produce the same output signal at the same bias voltage.