We also ran some guider tests at the start of the night. We learned the following: a) running the sextractor based Guider from the ram disk did not reduce the jitter. Santi’s initial analysis showed the same behavior we saw in the past. Whatever causes the variations in sextractor processing time doesn’t seem to be caused by disk IO. b) Santi’s new guider sort of worked in the PyGuide only mode but finding guide stars was not as reliable as with the sextractor based guider (this is just our impression – a more careful study is needed). We noticed a very long interval (30 seconds or more) between the first image (2kx2k) and the start of the guider sequence. It is amazing that this worked at all after just a day of programming but as is it is not yet usable. We tried lowering the threshold to 3 (sigma) detection but this did not make a noticeable difference and there also seems to be a coding problem setting this variable. c) We also tried the combined mode that uses sextractor for guide star selection and pyguide for the centroids measurement. This did not work – no crash but for some yet unknown reason no guide stars were found in this mode. There were also periods where the guider was oscillating back and forth. They forced us to to give up on the deep z-band supernova field. There were also large guider excursions on some images while on the survey field at RA = 65 deg DEC = -55 deg.