Warning: prior to Monday 18th June 2002, there was an error on this webpage with the generation of the new ESDs, please review if you have been using this page prior to this date Advice on summing multiple powder diffraction files, normalizing the data to look like a fixed count time datafile, and re-generating the resulting ESDs on each data point. Convert the data files immediately to (4 columns of values) : X Y (real counts) sigma Y (=sqrt(Y)), and normalisation factor (e.g. a standard monitor count - could be beam current, typical time per point). To create a single normalised powder diffraction file which displays as though the data had been collected with a fixed count time (with appropriate ESD's for each data point), you simply then add all of: the Y values up, the sigma squared values up, monitor count up, and square root the sigma^2 to get esd again, and finally then divide Y and (new) sigma by the monitor value. If you sum up and normalize 2 points - each having 100 counts - using the above - . Sum a point on 2 scans: Scan one: Counts = 100 Sigma Y = 10 Monitor = 1 Scan two: Counts = 100 Sigma Y = 10 Monitor = 1 Summed Counts = 200, Summed Sigma^2 = 200, Summed Monitor = 2 Square root the Summed Sigma^2 = 14.14 Summed Counts = 200, New Sigma = 14.14, Summed Monitor = 2 Now divide Y and (new) sigma by the monitor value to give the merged and normalised value, Merged/Normalized Counts = 100 , New ESD = 7.1 In the above case, the ESD has improved by root (2), in this instance by counting twice (as to be expected).