We had some problems with our dome flats this afternoon - the chips appeared much darker at one end than the other (i.e. left to right asymmetry, not due to the 2 amps). We restarted SISPI and took the flats again, and they looked fine.

Tim adds:

I have noted something interesting this evening.  My program involves following the same field for several hours.  The exposures are only 20s long and so not long enough for guiding to engage and I have therefore turned it off.  Thus I am following the same field from ~2.5 hours east to ~2.5 hours west with open loop tracking and taking exposures approximately equally spaced in time.  I observe a significant and nearly continuous drift in RA (and a smaller wobble in DEC).  Moreover, if I try to reset the pointing nothing much happens.  I conclude, therefore, that the telescope is *not* tracking at the sidereal rate, but it *thinks* it is so that its pointing error accumulates steadily over my observing period.  Thus some pointing error seems to be accumulating while tracking, as well as, as we already know, in an offset and possibly in slew.

More info has been provided to the TCS team.