On the night of July 19-20 I obtained a series of 8 images in two separate groups (2x 4 images with 10s separation each, the two groups being one minute separated). On two images, both part of the 2nd group, the object is too faint to be seen. Hence, the result was 6 segments of brightness behaviour. This yielded the following brightness diagram:
The data fit a period of 61 seconds. Over the past year, the period was as follows:
25-27 Aug 2009: 71 seconds (see here)
19 Nov 2009: 62 seconds (see here)
2-9 Mar 2010: 88 seconds (unpublished)
20 Jul 2010: 61 seconds (this post)
Clearly, there are regular changes in periodicity, without a clear direction.
One peculiar observation in the current series, is that of an anomalous bright spike in one of the "valleys" of the brightness profile. The satellite is all but almost invisible in most of the image, except for a brief spike right at the end of the exposure (EF 100/2.5 Macro USM, 1600 ISO):
I have never observed this before. It looks quite specular, i.e. due to a glint on a flat reflective surface.
As a comparison, here is the object during it's brightest appearance:
I'll keep periodically revisiting the USA 144 decoy periodicity.