Phase 3 WISP arrays and BRITE arrays are currently being used extensively in many projection systems in many different facilities. These arrays have not been annealed at the factory, and previous tests with the arrays have revealed instabilities in the radiometric output when the arrays are driven at higher voltages. In some applications, the instabilities can be avoided by operating the arrays at lower voltages. In many KHILS applications, it is desirable to drive the arrays with the highest possible voltages to simulate hot missile targets. In one KHILS application (the KHILS VAcuum Cold Chamber, KVACC), the arrays are cooled to near cryogenic temperatures and then driven to high voltages. At lower substrate temperatures, the characteristic responses of the emitters change. Thus, it is important that the response and the stability of the radiometric output of the arrays be well understood for various substrate temperatures, and that the arrays either be annealed or operated below the voltage where the emitters begin to anneal. KHILS has investigated annealing procedures in the past, but there was concern that the annealing procedures themselves -- driving the arrays at high voltages for long times -- would damage the arrays. In order to understand the performance of the arrays better, and to reduce risks associated with driving the arrays at high voltages and operating the arrays at low substrate temperatures, a systematic measurement program was initiated. The radiometric output of new Phase 3 WISP arrays was accurately measured as a function of voltage and time. Arrays designated for testing were driven to the higher voltages and the radiometric output was measured for as long as two hours. Curves indicative of the annealing were observed, and it was determined that the maximum stable output without annealing was about 500 K (MWIR apparent temperature). Blocks of emitters were annealed and tested again. It was determined that stable output of as much as 680 K could be obtained with annealed emitters. KHILS personnel worked with Honeywell Technology Center (HTC) to establish annealing procedures that could be done by HTC in the future. Conclusions to date are that once the emitters are sufficiently annealed, their output does not change further with time, except for some small transient effects that will be discussed in the paper.