The plot posted in the previous hclog entry confused me, since it seemed to show the "hot" bellows thermometer (qw:bellowstemp3) as being similar in magnitude to the other thermometers.
However, I checked the archived values. It is indeed at a plateau of around 85 C, unlike the others, which are in the 34-44 C range. It seems the strip tool scale was set differently for that variable. If you look closely, you can deduce this from the size of the jitter in that signal.
I would have preferred all temperatures to be on the same scale on the striptool...