There are many confounding factors that influence the system noise and hence the detectability of transits. The purpose of the tests was to measure the effects of these factors, identify the optimal operating conditions under the influence of each factor and show that when all of the effects are taken together that Earth-size transits can be reliably observed. The Testbed Facility incorporates the ability to measure the following effects:

Factors Investigated

1. Spacecraft jitter motion: Although the spacecraft is far from the influence of the Earth, which causes atmospheric-drag torquing, magnetic torquing and gravity gradient torquing for Earth-orbiting spacecraft, the spacecraft is still subject to torquing due to the solar wind and solar radiation pressure. The spacecraft is quite rigid with only one articulated device, a high-gain antenna and the only deployable part being an ejectable cover. Therefore the calculated spacecraft jitter is quite small, being estimated to be on the order of 0.01 pixels ( 3 sigma rms). Hence motions of this size and larger were introduced into the tests. The facility can introduce motions up to 500 millipixels using piezoelectric transducers.
   
2. Dynamic range of sensitivity: The sensitivity to transits must include the range of stellar visual magnitudes of 9<m_v<14. Stars covering this range of brightnesses are included in the tests. Additionally, background stars as faint as m_v=19 are also included.
   
3. Double stars: Due to the "soft focus" of the photometer optics, background stars adjacent to the target star overlap with the target stars. Spacecraft jitter causes the photometric aperture used for the target stars to move. The background stars are typically five stellar magnitudes fainter and for those near the edge of an aperture, their flux could cause an apparent change in the brightness of the target stars. The simulated star field includes double stars to simulate this effect.
   
4. Smearing: During CCD readout when the image is shifted to the output amplifier, flux from other stars within the same CCD columns contributes noise to the measurement, since there is no shutter in the system. The effect depends strongly on the brightness of the intervening star. Thus the simulated star field includes stars of differing brightness in the same CCD column.
   
5. Field rotation: Every three months the spacecraft is rotated 90° to compensate for the apparent motion of the Sun. To demonstrate that moving the star field to a different part of the CCD does not produce a different photometric result, the simulated star field can be either rotated or translating.
   
6. Operating temperature: The CCD operating temperature affects the dark current. At the end-of-mission (EOM) the dark current may no longer be negligible due to radiation damage. The CCD operating temperature can be elevated to simulate the higher EOM dark current. The CCD operating temperature is proportionally controlled to ±0.027 °C.
   
7. Amount of defocus and size of the photometric aperture: The amount of defocusing and the size of the photometric aperture affect the precision. The effect of defocus on noise is measured to determine the optimum focus. The optimum aperture size as a function of stellar brightness as it affects the measurement noise is determined.
   
8. Bright stars and saturated pixels: The few very bright field stars that are in the large field of view monitored by the Kepler Mission photometer could cause problems well beyond their nominal PSF due to the blooming of saturated pixels. Based on star catalogs, there 15 stars brighter than 6th magnitude. (About half of these can be placed between gaps in the FOV as shown earlier.) Though rejection of a few CCD columns from the data processing is acceptable, it is important to determine the number of nearby columns that are affected by the highly saturated columns. The test measures the number of columns that are not usable due to the presence of a very bright star. The Facility produces m_v=4 bright stars by piping in additional light to various points in the FOV using fiber optics.
   
9. Long duration testing: To demonstrate the ability to maintain long term relative precision, tests of up to two weeks in duration were run. During this time several five-hour transits were injected at various intervals as well as several twelve hour transits.
   
10. Cosmic ray hits: The effects of cosmic ray hits on CCDs is a well understood problem both in terms of their long-term degradation on the CCD performance and the spurious effect on the data measurements due to the charge added to the impacted pixels. These effects are introduced into the data stream by software modeling.
    
11. Compound test: All of the above tests are designed to isolate the effects of individual confounding factors. In flight, all of the confounding factors act at once. In the compound tests, the expected level of each of the confounding factors are simultaneously impose for durations of up to two weeks.