Excellent coverage of the entire 16-month 1996–97 outburst cycle of GRO J1655–40 was provided by the Rossi X-ray Timing Explorer .
We present a full spectral analysis of these data , which includes 52 Proportional Counter Array spectra from 2.5–20 keV and High Energy X-ray Timing Experiment spectra above 20 keV .
We also include a nearly continuous All-Sky Monitor light curve with several intensity measurements per day .
The data are interpreted in the context of the multicolor blackbody disk/power-law model .
The source exhibits two principal outburst states which we associate with the very high and the high/soft states .
During the very high state , the spectrum is often dominated by a power-law component with photon index ( \Gamma ) \sim 2.3–2.7 .
The source exhibits intense hard flares on time scales of hours to days which are correlated with changes in both the fitted temperature and radius of the inner accretion disk .
During the high/soft state , the spectrum is dominated by the soft thermal emission from the accretion disk with spectral parameters that suggest approximately constant inner disk radius and temperature .
The power-law component is relatively weak with \Gamma \sim 2–3 .
During the last few observations , the source undergoes a transition to the low/hard state .

We find that a tight relationship exists between the observed inner radius of the disk and the flux in the power-law component .
During intense hard flares , the inner disk radius is observed to decrease by as much as a factor of three on a time scale of days .
The apparent decrease of the inner disk radius observed during the flares may be due to the failure of the multicolor disk model caused by a steepening of the radial temperature profile in the disk coupled with increased spectral hardening and not physical changes of the inner disk radius .
The distortion of the inner disk spectrum by the power-law flares indicates that the physical mechanism responsible for producing the power-law emission is linked to the inner disk region .

Assuming that our spectral model is valid during periods of weak power-law emission , our most likely value for the inner disk radius implies a _ { * } < 0.7 .
Such a low value for the black hole angular momentum is inconsistent with the relativistic frame dragging and the ‘ diskoseismic ’ models as interpretations for the 300 Hz X-ray QPO seen during some of these RXTE observations .