I still haven't dealt with http://bolocam.blogspot.com/2011/07/additional-problems.html, or essentiall any of http://bolocam.blogspot.com/2011/07/minor-ongoing-problems.html. However, they're probably related to this problem: % CLEAN_ITER_STRUCT: There were            1 bolometers with high weights:       13.5047 indices:      142 or flagged indices          142 which indicates that in a simulation in which there can be no outliers (in terms of weight/scale), there is one being rejected as an outlier. That indicates that weights are being computed incorrectly, despite the fact that scales look right (so far): Relsens calibration: Scaled to bolometer #        0% RELSENS_CAL_PCA: There were            0 NAN scales and            0 very low scales% RELSENS_CAL_PCA: Scale Median/Mad:        1.0005510+/-     0.010990833 led to            0 scales set to zero for a total of            0 bad bolos% RELSENS_CAL_PCA: Scales avg+/-std =        1.0007304355062783+/-       0.0103252880998329Relsens calibration: Scaled to bolometer #        0% RELSENS_CAL_PCA: There were            0 NAN scales and            0 very low scales% RELSENS_CAL_PCA: Scale Median/Mad:       0.99955294+/-    0.0093122063 led to            0 scales set to zero for a total of            0 bad bolos% RELSENS_CAL_PCA: Scales avg+/-std =        0.9990063123041220+/-       0.0096663438876613