In the mid May of 2004, we began work with the linear arrays. The electronics were finished by June, and before mid June we had already collected and processed data for the 1.7 micron array's first test run.

The chip is mounted to the board via two 14 pin connectors split in half and soldered so that two rows of fourteen are ~0.8 inches apart, the distance between the detectors pin rows. In order to create the +3.5V DEADPOT, the +5V coming in from outside is split to run through a voltage divider that is soldered onto the board. Since we cannot place the suggested  200 Ohm resistors close to the source of the clock signals to clean up the pulses, we instead place them on the board by their respective pins. Wires are then wrapped to the posts and soldered to connectors that interface with the outside world.

Two video lines, 1A(even) and 2A(even), are run out on a 25 pin connector, wrapped with grounds. Because of restrictions caused by pre made Dewar electronics, only four Video wires could be brought out. While wiring to the chip include these four wires (1&2A[even] and 1&2B[even]), the connector on the other side of the Op-Amp cabling only has two wires. More may be added in the future, however this is not necessary for testing in our current set up. These wires connect to a breakout box, and E-Series connector for a NI-DAQ card from National Instruments (PCI-MIO-16E-4) which interfaces with LabVIEW 7.0. The wires are connection in the box so that they are differenced before their values are inputted into the  acquisitionprogram.

Two clock lines are also run from the back of a Linux box through the Clock/Bias line into the Dewar and out another 25 pin connector. The C program used (lin.c) sends pulses out of the parallel port in sequences outlined by hexadecimal values (pin 1 is value "1", pin 2 is "2", pin 3 is "4", pin "4" is "8", pin 1 and pin 2 is "3", )etc. The CLOCK line is run between both CLOCK pins (odd and even) and the LSYNC to the LSYNC pin. From the box another line, TRIGGER, runs to the breakout box for the NI-DAQ card to allow the LabVIEW program's "DAQ AAssistant function to operate via the number of trigger signals. For the fourth of the 512 array we used a TRIGGER of 232, which doubled every pixel output and added "virtual" pixels to the end of the readout. The data goes through functions in the LabVIEW program to condense the double values into singles and readout the waveform.

Here are some results:

As you can see in this image taken at diode value 0.591V (just below room temp), pixel 101 is very prominent (in reality, this is pixel 202, since we are only looking at even values). This pixel has a peak in every test image we took that is easy to distinguish from its neighboring values. The data we used begins with the next image, which is taken at 0.614V, or 12°C.

The coldest temperature we exposed the chip to was nearly -201°C, which resulted in the following output:

What we see here is mostly dark current - that is, if the chip is still behaving properly, then this would be the effects of dark current. Let us take a look at something close to 180K (-93°C), where the planar array still gives valid data, and after which data points to breakdown. The closest to this is -80°C.

It appears that the chip is still very active at this temperature, unlike its planar counterpart. However, we noticed what appears to be a "light leak" at long integration low temperatures. This could be equivalent to the component we noticed in the 320x240 array, but it is difficult to determine, as this area of the array, though to the right and therefore the end of the data set in this graph, is in the center of the opening.  Perhapslack of packaging reduces shielding from such aanomalies Here is one of the data sets where it is evident, at -190°C:

And at -127°C

Comparing to the earlier graphs of -201°C, breakdown must occur after -190°C, as the chip becomes unresponsive and the artifact disappears.

Here is the dark current curve for the array, using high to mid-range temperatures: