A problem with the temperature controller for H band at the northern observatory was noted by Mike Skrutskie:

> From skrutski@north.phast.umass.edu Fri Apr 3 10:10:52 1998
> Date: Fri, 3 Apr 1998 13:10:36 -0500 (EST)
> From: skrutski@north.phast.umass.edu (Mike Skrutskie)
> To: roc@ipac.caltech.edu
>
> The H-band temperature craziness in the north log for last night
> probably explains the variability in the H-band darks that
> were noted recently. The temperature controller in this band
> was not working properly due to a bad contact. I'll be very
> interested to see if there were any significant changes
> last night due to lowering the operating point by 4C (zeropoint
> may drop a few percent).
>
> Mike.

DARKS has processed this night (980403n) and the subsequent night (980404n) with the following results. The evening and morning dark medians for 980403n band H are shown in Fig.1. The black line shows the frame median as a function of frame number; the quadrant medians are shown as variously colored diamonds.

The corresponding information for J band is shown in Fig.2. The large difference in quadrant medians but relatively low noise for each quadrant is typical for the northern observatory's J band, and both dark sequences are typical of other nights' J darks.

Similarly, the K dark sequences are shown in Fig.3. These are typical for the northern K array, similar to J except for smaller differences between quadrants.

980404n had only morning dark sequences. The median plots for H and K band are shown in Fig.4. Both J and K are the same as on the previous night, so J is omitted here. H, however,looks much better behaved than on the previous morning, although it is still exhibiting its "rising" behavior, which seems to happen in both the morning and evening!