Figure 1: Signal flow diagram for the VLPC characterization test
stand.
Figure 2: VLPC test cassette: side and top view
Figure 3: Cold end detail of the VLPC test cassette,
showing hybrid holder
with optical fibers and VLPC contact springs
The fixture that eventually holds the VLPC's in
line with optical fibers carrying the light signal from the tracker, and serves
as enclosure for insertion into the cryostat, is called a cassette. There, the
plastic carrier serves both the purpose of mounting
the hybrid into the copper isotherm at the cold end of the cassette, and for
aligning the optical fibers with the VLPC pixels, which are glued permanently
into cylindrical slots in the Torlon carrier. For the characterization however,
the VLPC's are temporarily mounted in a special test cassette (Figures
2 and 3).
This test cassette does not require any soldered electrical
connections or gluing of fibers. The test cassette holds
8 hybrids (= 64 VLPC pixels) at one time. During the characterization, we
tested 7 new hybrids in each cool-down, using one slot to monitor run
to run light level variation.
The primary features of the test cassette are the following:
i) Light Distribution- Light from a red LED was used to pulse the VLPC's. A red LED was chosen because of its fast response to drive pulses. The VLPC's show relatively flat response in the visible spectrum, so the use of a red light rather than a green one should not compromise the measurements. The light from the LED is diffused by a scored piece of acrylic and brought to each VLPC by an individual optical fiber, whose diamond-polished end sits about 5 mm above the VLPC surface.
ii) Removable Torlon Carriers- The slots for the hybrids were designed for easy insertion and removal of the carriers. The bias supply and read-out lines are connected to the substrate by single wire, gold plated springs, which contact gold-plated castellations on the hybrid. The spring alignment with the hybrid turned out to be non-trivial, due to minute variations in the VLPC chip position on its substrate (the alignment between VLPC and Torlon carrier was guaranteed to 10 micron, but not the VLPC mounting on the substrate). Therefore the electrical contact was checked by a warm resistance measurement for each individual pixel, before the cassette was cooled down for VLPC characterization.
iii) IR tightening- Since the VLPC's are sensitive to IR radiation, care must be taken to shield the devices as much as possible from IR originating at the warm end of the cassette. In the test cassette, the optical fibers were fed through copper baffles, and layers of black felt were inserted between and around the optical fibers and electrical cables close to the hybrids, to block the IR light path from the warm end of the cassette. (In the final cassettes, the optical fibers, which themselves absorb IR light quite efficiently, terminate closer to the VLPC surface -about 50 micron at low temperature-, thus making IR shielding somewhat easier.)
A VME-based low voltage supply designed at Fermilab supplied the bias voltage to the VLPC's. The signals were read-out through a QPA02 preamplifier, also designed at Fermilab. This transimpedance amplifier has a gain of approximately 2000 and a rise time of about 5 ns [7]. The preamp output signals are then transformed from double ended 100 Ohm to single ended 50 Ohm before going into CAMAC 2249A ADC's. We read the ADC's through a PC. The data was transferred to a VAX 4000-60 for analysis.
The VLPC's were cooled using liquid helium through a flow cryostat system[6]. The temperature was computer controlled to a tolerance of about 0.1K.