a) Best-fit diskbb temperature vs eqpair soft photon temperature. The solid line is the linear best fit. It has a slope of 1.9. The dot-dashed lines correspond to the 1σ error on the slope i.e. 1.9 ± 0.5. b) Flux of the diskbb component versus diskbb temperature. The solid black line corresponds to the log-log linear best fit and the dot-dashed lines to the 1σ. The solid red line corresponds to the best fit assuming a law. c) Plot of R_proxy (see Sect. 4) versus the disk inner radius R_in deduced from our fits (reported in Table 4) and which assumes the X-ray luminosity ∝ Ṁ^α with α = 2.5(the results are very similar for α = 2 or 3). The dashed line corresponds to R_proxy = R_in. The blue points in each figure correspond to the 2008 Suzaku observation.