Before clock-driven analog multiplexing techniques became the baseline approach to second-generation FPAs, a clever approach was invented in the U.K. that arguably might have qualified it as one of the first GEN2 devices. Since the benefits of high detector count were widely appreciated, there were various attempts to solve the detector access problem. One of the first attempts was the use of TDI in which the output of any given detector was delayed and summed with the output of a following detector as described previously in Section 5.4.3 for the HAC Discoid Serial Scan FLIR. However, with only bulky analog electronics, delay lines, and many discrete components, that task was difficult. Later, in the mid-1970s, the U.K. was trying to come up with their own version of a Common Module FLIR to duplicate the successful U.S. Common Module FLIR program. Tom Elliot, while at the British Defense Ministry’s Royal Radar Establishment (RRE), Malvern, later renamed QinetiQ, realized that TDI could be accomplished inside a single, extended, monolithic, photoconductive detector to make it appear to be many more discrete “virtual” detectors [32, 47]. The resulting serial output could then be coupled to a single amplifier circuit, thus effectively multiplexing the virtual detectors together. Some called his invention TED (for Tom Elliot’s Device), although the official name for it was Signal Processing in the Element (SPRITE). Figure 7-1 illustrates how it worked. The detector’s bias voltage was chosen so that photoelectrons, formed from a scene pixel, drifted at the same rate the image was being scanned. Accordingly, photo charges accumulated and remained registered with the same scene pixel across the whole detector. When they reached the end, they were read out with a preamplifier. Early versions made in 1975 were 50 µm wide by 1,000 µm long and so appeared to make it possible to have 20 virtual detectors contained within its length. However, charge diffusion resulted in an effective individual detector length of about 75 µm, or more, so this approach did not achieve the same resolution in the scan direction as was obtained by using discrete detectors. By 1980, arrays were built for parallel-serial scanning with eight detectors each of 800 µm by 75 µm. The latter configuration would thus effectively provide a total of about 8x10 elements in an 80-element array while using only one preamplifier and one postamplifier. An equivalent 80-detector U.S. GEN1 FLIR, for example, would have needed four 20-channel preamplifier PCBs and four 20-channel postamplifier PCBs plus a much larger power supply. So the U.S. GEN1 FLIR would have required a much larger package. SPRITE detectors were the basis of U.K. GEN1 Common Module FLIRs, although they might have qualified as second generation. The SPRITE-based U.K. common modules provided high sensitivity and high-quality imagery in a very compact package. Various versions were built and, by 1999, over 3,500 FLIRs were made based on the Class II configuration alone. They were used in the Falklands conflict and in the two Gulf Wars.
