When fitting two spectra simultaneously, is there a way to vary one model, but keep the ratio of normalizations between different models the same?
If, for example, what you want is for this expression to be true:
(N_R1 / N_P1) = 2.0 * (N_R2 / N_P2)
and your source model definitions are as so:
source 1 = xsraymond[ray1] + powlaw[po1] source 2 = xsraymond[ray2] + powlaw[po2]
(so N_R1 is the normalization of ray1, N_P1 is the normalization of po1, and so forth), then the parameters need to be related in the following way:
sherpa> ray1.norm => 2.0 * po1.ampl * (ray2.norm / po2.ampl)
The help file on linking model parameters contains more information and examples.
