Most calculations were done with version 4631, but for some calculations, versions as late as 5271 may have been used for calculations in the refereeing process. No significant differences between models in those versions were found.
The initial white dwarf models without accretion were made using the make_co_wd (0.6-1.0 Msun) or make_o_ne_wd (1.1-1.34 Msun) test case in $MESA_DIR/star/test_suite. To get different masses, the initial mass and wind efficiencies were adjusted on a trial and error basis until the ending model had a mass just slightly above (about 0.01 Msun or less) the desired mass.
Note: these inlists can be found in the "inlists" subdirectory.
A non-accreting WD with a mass higher than the desired mass and a core temperature higher than the desired 3e7 was loaded into the first inlist, inlist_co_nova_1. This inlist would turn on weak mass loss and evolve the WD until it was at the desired core temperature or at the desired mass, whichever came first. If the minimum mass limit was met first, it was then reloaded with mass loss shut off until it was properly cooled.
The second inlist would then load this model and rescale the mass down to the proper mass (if it was not already at the exact mass), and irradiation would be turned on at a depth close to where we would expect hydrogen burning to occur on a steadily burning model (some trial and error was needed here, both in the depth and the intensity of the irradiation).
The third inlist would relax the irradiation to an even higher level and accrete solar composition material at a low rate through a few flashes, stopping during near the end of a flash to ease the transition away from irradiation.
The fourth inlist shuts off irradiation and relaxes the accretion rate to a value that would be expected to yield steady burning. After one or two flashes, steady burning should begin and after a few time steps, the model is saved for use in exploratory runs.