This is a pre-release which introduces a new function for count averages. It is relevant only if you want to show averages over hours or even days in long-term recordings.
The data were obtained with a FNIRSI GC-01 counter with the RadPro firmware, running overnight for 17h.
The first graph shows the time course for the regular CPM and CPS. But, made possible by the new functions, it also shows a CPH (per Hour) and a CPD (per Day) trace, though the letter is not visible as it is identical to the CPH trace, as less than a full day has passed. Last trace is CP(2min), which is like CPM but counting is over 2 min instead of 1 min, so count rate is twice as high.
CPH is clearly nicely showing an average of CPM.
To be sure the counter is working correctly, an inset is added showing the Poisson test curves for CPM and CPS, and both confirm proper working.
When you extend the counting period to 2 min, the resulting distribution of count rates is still of the Poisson type, as the 2nd graph shows. However, as the CPH (and CPD) data are normalized to CPM, these data are no longer of Poisson type!
Of course, there is no harm done. The Poisson test should establish that the counter is working correctly. Once this is done, you can do with the data whatever you like! Converting CPM data to Sievert / h data also renders them non-Poisson, and this is obviously also ok.
The way to use these new functions is shown in the 3rd picture. The function is:
COUNTRATE_AVG(<Variable>,<type>)
This function needs to be entered into the formula interpreter. The examples are: COUNTRATE_AVG("CPS","HOUR") recorded in variable CPM1st is the hourly average of CPS, normalized to CPM
COUNTRATE_AVG("CPS","DAY") recorded in variable CPM2nd is the daily average of CPS, normalized to CPM
COUNTRATE_AVG("CPS",120) recorded in variable CPM3rd is the average over 120 sec of CPS. It is NOT normalized to CPM! The resulting values are twice as high as for CPM, and the distribution is of Poisson type. Of course, you could normalize them to CPM by dividing by 2 (circled entry on the Graph Formula side). NOTE: this was NOT applied in the graphs above! The resulting distribution is again of non-Poisson type, but you would get an averaging on your CPM data.
The intended use is to set < Variable > to a GeigerLog variable holding CPS data. However, you can average whatever else you like, but you need to find the proper interpretation for the result ;-).
To use these new functions download the zip file and install as detailed in the GeigerLog manual.
This is a pre-release which introduces a new function for count averages. It is relevant only if you want to show averages over hours or even days in long-term recordings.
The data were obtained with a FNIRSI GC-01 counter with the RadPro firmware, running overnight for 17h.
The first graph shows the time course for the regular CPM and CPS. But, made possible by the new functions, it also shows a CPH (per Hour) and a CPD (per Day) trace, though the letter is not visible as it is identical to the CPH trace, as less than a full day has passed. Last trace is CP(2min), which is like CPM but counting is over 2 min instead of 1 min, so count rate is twice as high.
CPH is clearly nicely showing an average of CPM.
To be sure the counter is working correctly, an inset is added showing the Poisson test curves for CPM and CPS, and both confirm proper working.
When you extend the counting period to 2 min, the resulting distribution of count rates is still of the Poisson type, as the 2nd graph shows. However, as the CPH (and CPD) data are normalized to CPM, these data are no longer of Poisson type!
Of course, there is no harm done. The Poisson test should establish that the counter is working correctly. Once this is done, you can do with the data whatever you like! Converting CPM data to Sievert / h data also renders them non-Poisson, and this is obviously also ok.
The way to use these new functions is shown in the 3rd picture. The function is:
This function needs to be entered into the formula interpreter. The examples are:
COUNTRATE_AVG("CPS","HOUR")recorded in variable CPM1st is the hourly average of CPS, normalized to CPMCOUNTRATE_AVG("CPS","DAY")recorded in variable CPM2nd is the daily average of CPS, normalized to CPMCOUNTRATE_AVG("CPS",120)recorded in variable CPM3rd is the average over 120 sec of CPS. It is NOT normalized to CPM! The resulting values are twice as high as for CPM, and the distribution is of Poisson type. Of course, you could normalize them to CPM by dividing by 2 (circled entry on the Graph Formula side). NOTE: this was NOT applied in the graphs above! The resulting distribution is again of non-Poisson type, but you would get an averaging on your CPM data.The intended use is to set < Variable > to a GeigerLog variable holding CPS data. However, you can average whatever else you like, but you need to find the proper interpretation for the result ;-).
To use these new functions download the zip file and install as detailed in the GeigerLog manual.
Last edit: ullix 2024-06-15