Comments

• The hot pixel finding algorithm has some difficulty dealing with on-board split-event reconstruction (SER). Multi-pixel telemetered events (items in the PIXELS table with SHAPE greater than zero, GRADE greater than one) cannot be compared directly to the thresholds set by the noise limit parameters. The charge-splitting ratios obtained from the CAL are therefore used to scale the thresholds appropriately for each grade. If a telemetered event passes this aggregate cut, each of its comprising pixels must be taken to pass as well, despite that the reality might be otherwise. This over-counting appears to be unavoidable, but hopefully over the course of many frames the true location of the hot pixels will become apparent.
• In HTR mode each telemetered ``pixel'' is assumed to represent activity in only one detector element. This is a good assumption in practice even though the on-board binning collects 74 rows together before each readout. Consequently it is possible that some of the telemetry will be mishandled by the hot column finding algorithm. In particular the algorithm is prone to under-estimate the average number of events in the more active columns. The impact of this potential error on the efficacy of the algorithm has not been investigated.