FULL DISCLOSURE: Some Problems seen in the Henderson Data
(you didn't really expect everything to work smoothly, did you?)

1. Our first data period (Electrical Shop, ES8100) was run with the wrong thresholds.  See this link, then this one for a discussion of threshold settings.  The default thresholds were much too low, letting in electronic noise and lots of accidental or fake coincidences.  Luckily, the written data includes information related to the size of the signals, so we should be able to figure out a way to correct these counts!  We will discuss this later.

2. Data was lost for our final underground location.  Typically our 2-3 week runs generated  15-20  data files (the datalogger was pre-programmed to open a new file every 24 hours).  But only one file survived from the final run.  Stranger still, it is a mere 10 minutes long.  That's not much data!  We'll be able to report a rate at that depth, but the low statistics mean our measurement will have a large error bar.  You'll understand how when we learn to calculate the expected errror on each of our readings.

3. A channel or two (of the 4 detectors we used) showed noisy behavior.  This may compromise the data from one of the two satellite modules (but its OK because we did have 2!).  Looking at the data, you may find periods of time where the count rate ran wild.  Look, for example, at these histograms from the RFHS files:
Channel 0 NoiseChannel 1 Noise
CHANNEL 0                                                    CHANNEL 1
 Channel 2 NoiseChannel 3 Noise
CHANNEL 2                                                    CHANNEL 3
Each histogram shows the hourly rate (counts/hour) over 10 days (480 hours).  Channels 0 and 1 relatively stable rates fluctuating around 1600-1900.  Channels 2 and 3 show discontinuous rates with huge jumps (with double the average rates of 0 and 1, and values as high as 4000!).  But notice how perfectly both 2 and 3 track one another.  I could imagine the module that stacked 2 and 3 together too near some electrical device (fan? transformer?) that occaissionally kicked in and generated all sorts of noise.  Perhaps the operating settings for detectors 2 and 3 were not optimized: they were overdriven at the high we used and/or the threshold settings were simply too low.  In any case, they are clearly oversensitive to environmental noise.  Your 1st assignment will be to analyse all the data files to identify which channels and which periods we can assume are good.

4. Data files written to the Base Station PC grew too large when the period between runs ran over 2 weeks.  The 32-bit processor hit an addressing limit just over 2-Gigabytes, and the DAQ program ended with an error.  The data collected to that point remained on disk, but the Windows operating system was unable to handle files of that size.  They could not be opened.  Transfered (through an iPod, which has much more memory than any of the portable flash drives) to a Linux machine, we were able to split the over-sized files into 20 100Meg files.  These files still need to be analyzed.