A difference between the results of the rheoFoam and rheoHeatFoam

[Joe199999]

Hi guys
I solved a steady problem using the rheoFoam and got a good convergence. Then continued that using the rheoHeatFoam (by addition of related parameters like T, ...). But rheoHeatFoam changed the results of hydrodynamic parameters, like velocity, pressure, drag coef, etc about 2%.
No viscous dissipation and the properties are T-independent.
You can also check this using the tutorial problem named cylinder (of90/tutorials/rheoHeatFoam/cylinder/powerLaw).
Do you know why?