Strange Light Curves in Dwarf Novae Clothing

A bit of a follow on to Tom's previous post about Dwarf Novae targets here. The search for faint background Cataclysmic Variables such as DN at the prompting of GO Director Martin Still is a very interesting tangent that has developed at PH since the project's inception. Over 200 such systems similar to the SU UMa target known as NIK 1 (KIC 4378554) could be hiding in the Kepler FOV and thus far this effort has been marked with some early success involving several promising candidates.

Since the recent Q7-Q9 data release, the sharp eyes of several PH users have uncovered a host of light curves that seemed to resemble such Dwarf Nova type variables:

These looked promising at first, but some immediate problems quickly emerged. The events were too intense, possessed no superhumps typically associated with superoutbursts and more importantly they were present only in the Pipeline corrected Kepler data (PDCSAP_FLUX), and not at all in the raw light curves found under SAP_FLUX!

Though no apparent commonality is found across FOV co-ordinates or hardware modules, (and there are several possible explanations common to the pipeline reprocessing that remain to be hashed out), one clue that may offer some explanation can be noted in the August 1st Mission Manager update by Roger Hunter:

"The new data release includes data taken in Quarter 8. This was an operationally challenging quarter, as the spacecraft exhibited a new behavior that resulted in a safe mode event and an extended interruption in data collection at the beginning of the quarter. Mitigation of the new behavior -- excess noise on the perceived sun vector -- resulted in another science loss at the end of the quarter. Quarter 8 represents the winter season for the spacecraft, where it is looking over the sun, with relatively small margins against pointing errors. The mitigations against the excess noise proved to be effective this year, when the Quarter 12 winter season was executed without incident."

In the extreme cases, most of this growing body of 'fake DN's' do appear to be found in the Q8 dataset, which is an interesting coincidence.

When in doubt, always check the SAP_FLUX!