This paper describes a case study investigating the effects of simplifications in a Monte Carlo fixed-source supercell model used to predict the graphite nuclear heating contribution made by an inserted gray control rod in the Dungeness B Advanced Gas-Cooled Reactor. These simplifications are known to produce excess reactivity in similar calculations with neutron multiplication included. The investigation compared the results of a series of calculations performed on two-dimensional supercells using fixed-source (MCBEND), reactor physics (WIMS), and criticality (MONK) codes. The effects of actions to reduce the reactivity of the model to a k_∞ of unity have been investigated, including increased leakage, reduced enrichment, and increased burn-up. The results have been compared to assess the consistency of three codes in addressing the issue of excess reactivity and to assess the suitability of the original nuclear heating calculations. The importance of core modeling when applying fixed source solutions to in-core dosimetry problems is discussed.