For all UH tests the detector temperature is held constant to <1mK rms, approximately a nominal temperature (37 K for JWST tests) throughout data taking. This level of control may be difficult to achieve even in ground based instruments and will increase complexity, weight and the potential for EMI in space applications. We have therefore characterized the shift in output signal with temperature (5V/5T) in 9.12 mK/min ramps of ± 50 mK and ± 600 mK about 37K and find significant effects not compensated by the reference pixels. We have subsequently developed a technique to correct such drifts to the 1ADU (~ 1 e-) level in post processing of the data.
Without active thermal control, temperature drifts in the JWST NIR focal planes may have amplitudes of up to ± 50 mK at rates up to 10 mK/min. We have evaluated output temperature drift under such conditions by running a series of standard 145 frame ramps in which the temperature is held constant at 36.950K for the two resets and the first part of the ramp, then ramped up 100 mK at 9.12 mK/min (55 frames) and finally held constant at 37.050K for the remainder of the ramp. This is followed by a similar sequence ramping back down 100 mK. The entire process is repeated five times to produce a set of five "up" ramps and another of five "down" ramps. The detector output for the five "up" ramps, corrected using the horizontal reference arrays (see Fig. 4) along with a histogram of the change over 100 mK for each pixel. From more detailed analysis we conclude that over 100 mk SV/ST is constant for each pixel with an array average value of 0.75 ADU/K or 3.25 ^V/K although values range from 0 - 1.5 ADU/K (0 - 6.5 ^V/K) for individual pixels. In order to meet JWST requirements without correction the focal plane temperature must be held constant to a few mK.
By ramping JWST-002 between 36.4 and 37.6K at 9.12 mK/min and taking full 145 frame ramps at each temperature, we have established linearity over this range and a template frame of SV/ST for each pixel. Subtraction of the scaled template to form individual frames of the 5 "up" and 5 "down" ramps eliminates drift between the plateau frames at each side of the ramp (see Fig. 5), but leaves an offset ~ 8.5 ADU (37 ^V) during the ramp. We interpret this as a time lag between the temperature of the mounting plate and that of the array. Delaying the temperature used for correction by 5 frames (60 seconds) relative to the plate temperature corrects this effect to within the noise as evident in Fig. 6.
We conclude that the SV/ST template technique combined with an optimized lag time will fully correct temperature excursions of at least ±50 mK at rates up to 10 mK/min. This holds promise of allowing the JWST NIR instrument focal planes to meet requirements without active temperature control.
Was this article helpful?