The current broken shutter code has two modes, reading and subtracting a bias and using an analytical model of the readout. The former has not been extensively used and is a nuisance as it cannot fit cleanly in the IsrTask as it has to hunt for a suitable bias exposure taken close to the data frame. The latter appears to work well.
We expect to replace the shutter within a few days, but I am not convinced that the problem won't reappear in the future; additionally the analytic version can be used from notebooks to produce a fast reduction on the mountain for e.g. "IFU images".
Please move the analytic code into the Task proper; we can postpone the possible removal of the bias-based code to another time.