Simulation studies of solar energy systems have traditionally been per
formed using hourly data because data for shorter time periods are gen
erally unavailable for extended periods. If simulations using hourly d
ata are to provide accurate results, the variation of radiation within
an hour must have a negligible effect on a system's performance. This
paper examines minute radiation data from three locations. The distri
bution of minute data within the hour is found to differ from previous
ly presented distributions of daily radiation in a month. Comparisons
of calculated and measured minute radiation on a tilted surface show s
ignificant differences which are attributable to inaccurate estimates
of the magnitude and distribution of diffuse radiation on a minute bas
is. Even with the same tilted surface radiation, significant differenc
es in simulated performance using hourly and minute-by-minute data are
found to exist for a simple photovoltaic system.