NASA NTRS1992
Simulation techniques are used to obtain the X-ray variability power spectrum of unevenly sampled GINGA data from NGC6814. A simple power law is not an adequate description of the power spectrum, with the residuals showing excess power on timescales consistent with the periodicity seen in EXOSAT observations of this object. However the shape of the folded lightcurve is very different, with 3 main peaks, two of which are separated by an extremely sharp dip instead of the single peak and small harmonic structure observed by EXOSAT. Using the dip as a fiducial mark, a second GINGA observation of this source taken one year later is found to be consistent with being completely periodic and phase coherent with this first GINGA observation. Thus the period is consistent with being constant over a period of 6 years, but phase coherence is only maintained on timescales of approximately 1 year. Over 75 percent of the total source variability is due to the periodic component (r.m.s. amplitude of 36 percent). The residual variability can be described as the more usual 'flicker noise' f(exp -1.1) powerlaw. This shows no apparent high frequency break on timescales greater than 300 seconds. Subtle differences in the shape of the folded light curve with energy, and the very large amount of power in the periodic component suggest occultation as its origin, though amplification of variability from an X-ray emitting 'hot spot' at the disk inner radius through gravitational lensing is also possible. The former suffers from the very arbitrary nature of the periodic timescale, while the latter is unattractive as it cannot simply explain the lack of high frequency break in the residual power. That these models probably fail to provide an adequate explanation may be due to the added complexity of anisotropy of the X-ray emission, suggested by the discrepancy between the lack of soft photons implied by the flat spectrum and the copious source of soft photons available from reprocessing in the iron line producing material.