I am aware the soviet reactors were cheap to build in part because they lacked the safety features of western reactors (I think the RBMKs laccked a reactor conainment vessel) and that candu reactors, which are prbly as safe as other commercial western reactors, have a slight positive void coefficient. The Soviet pre-Chernobyl RBMKs, however, had a very high positive void coefficient that made the reactor difficult to control at low power output when the trubines are to be kept operating at normal speed, a condition created in the experiment preceding the disaster; I think the difficulty in controlling the reactor in the conditions described is because coolant flow rate was slower in those conditions than at full thermal and electric power. The control rods used (to shutdown the reactor in an emergency, which was attempted when the reactor's power was spiking as a result of voiding) were also graphite tipped, so the bottoms of the rods moderated the reactor and displaced neutron absorbing light water; the rods fractured as a result of the ensuing power spike, which caused more coolant voids and prevented the insertion of more rods, more core damage and the disaster followed. Since the common isotopes of carbon are more massive than hydrogen-1 or hydrogen-2, a greater moderator to fuel volume ratio is generally needed in graphite moderated reactors than in light or heavy water moderated reactors, a bigger reactor vessel requires a bigger reactor containment vessel. Graphite moderator needed can be reduced by using more enriched uranium; Could lowering moderator to fuel volume ratio also decrease void coefficient? After the Chernobyl disaster, the Soviets reduced the void coefficient by adding more absorbers to the RBMK reactor core, and compensated for this by increasing loaded fuel enrichment from 2% to 2.4%. They also started using emergency control rods that were tipped at the bottom w/ materials that had less of a moderating effect or absorbed more neutrons.