Run 8 1-second scaler failures

 In a TOF presentation by Xiaoping Zhang, some fixed values of the RICH scalers appears to be present.

I was able to reproduce the effect from plotting the scaler rates from the same MuDsts, but not using the DB entries. It turns out to be only (and all produced) runs from day 47; all RICH scalers (I checked several) appear fixed at some values. This will affect the scaler-based SpaceCharge+GridLeak distortion corrections of the TPC. Either the DB should be used, or the data discarded.

ZDCx vs. (run-9000000+9*(run%1000)) from the MuDsts (1-second scalers):

ZDCx vs. (run-9000000+9*(run%1000)) from the DB (30-second scalers):

Runs from day 47 (below 48000 on the above plots) show the discrepancy. This is in addition to a "x2" problem seen in the 1-second scalers in the dAu200 (2008) of SpaceCharge for the Run 8 dAu data. Poignantly, this effect is also seen in the above MuDst plot for 1 second (a couple hundred events) from run 9052047.  Otherwise, the 1-second and 30-second scalers appear to match reasonably well.

It's already too late for the dAu data production (already mostly done), and it may be too late for the pp production from Run 8 (library finalization set for tomorrow), but it would be preferable to recover this data by implementing some kind of switch to use the DB's 30-second scalers when the 1-second scalers are failing. How to implement such a switch (if it's even possible to accurately determine when to toggle it) would take some careful thought and testing.