The heterogeneity and sodium void effects in ZPR-6 Assembly 5, a 2700-liter UC core, were analyzed using ENDF/B data. The cross-section sets used were generated for the homogeneous composition of this assembly, corrected for resonance spatial self-shielding in 238U via equivalence theory, and weighted with the fine structure of the fluxes in the unit-cell. Bilinear weighting with the unit-cell fine structure of the real and adjoint fluxes was also performed. The calculated critical mass was 18% greater than the measured value. The calculated keff of the as-built system was 0.988. The calculated ratio of fission in 238U to fission in 235U was, depending on the loading pattern in the unit-cell, 13 to 26% less than the measured value. The calculated ratio of capture in 238U to fission in 235U was 7 to 10% greater than the measured value. The measured reactivity difference between a small homogeneous sample and another matched plate heterogeneous sample was a factor of 4 greater than the calculated value, indicating the analytical tools used for treating the heterogeneity effects were not adequate. The reactivity difference between homogeneous and rodded samples was nil. The calculated reactivity worths of 235U samples are in a reasonable agreement with measurements. The predicted 238U reactivity worth was 15 to 20% greater than the measured value. The measured sodium-void coefficient was about 25% greater than the calculated value. The magnitude of this coefficient is quite sensitive to the particular loading pattern in the unit-cell and it is strongly dependent on the effective capture rates in 238U. The dependence of the void coefficient on loading patterns is predictable in direction but not magnitude. The void coefficient in a rodded sample was the same as that in a plate heterogeneous sample. The void coefficient in a homogeneous sample was 56% greater than that in the rodded or the plate samples.