Will the real K-40 please stand up (please stand up, please stand up)

[stamasd]

And indeed almost completely linear at 670V.

I used Am-241 for the low end because the 88keV from Lu-176 was not clearly distinguishable. Even the Am-241 peak was unusual-looking, with a hump to the right side which I haven't seen before. The 26keV from Am-241 was not visible at all.

In the graph I used the automatically generated mean for the peak ROI (yellow) as done by PRA. My eyeball says the actual peak is more at 3.72 Arb.U but since for all other peaks I used the automatically generated ROI mean I did the same in this graph for the 59.6keV.

Note that the Am peak is stretched horizontally quite a bit (0.02 Arb.U/bin) to show its profile better.

[Geoff]

That bump should be there. It’s likely 125 keV. You can see there are more peaks to a 241Am spectrum hidden by the 59 keV in my discussion on energy attenuation:

viewtopic.php?f=5&t=384&p=1822#p1822

[Sesselmann]

Silviu,

The automatic selection of ROI in PRA works best when the SD has been calibrated correctly. Pressing the keys [d], [f] and [g] selects a region spanning 1, 2 and 3 standard deviations. The default standard deviation is 1 however it's not uniform across the spectrum, so to set it properly you have to manually select your calibration peaks and then update the arb.u and SD in the calibration settings, once you have done this the gaussian correlation will sharpen up and separate the peaks better.

To manually select a peak you select a bin and press 'b' for beginning and [e] for end, the region will show in yellow and the mean arb.u as well as SD will show at the top of the window.

Steven