PRA Tolerance Threshold

PRA, BeqMoni, Theremino, Fitzpeaks and beyond!
Post Reply
Posts: 24
Joined: 03 Sep 2016, 18:11
Location: Germany

PRA Tolerance Threshold

Post by ManuelDiaz » 10 Oct 2016, 20:40

Hi folks!

As I've commented in other videos, I am struggling with what I think is just noise (either that, or something is radiating my home haha, I'll find when I get some lead ingots this week!) and I've found that lowering the Tolerance Threshold to 1.5 or 1 I get rid of that strange peak, which leads me to believe that there is some noise causing it. By lowering the threshold to that value my CPS background drops from around 120-140 CPS to 60-80cps so obviously it's a lost but not so huge.

With that setting, my resolution on Cs137 at 662 kev seems to be quite good (PRA calculates around 6.5/6.6, I don't know if I can trust that setting) but I can't see the 32kev x-ray (I guess as my Cs137 comes from a very old spark gap it's just hidden in the rest of the background.

So I'd like to know, what settings of threshold do you use? Is it good to use such a low threshold or by doing that I'm killing part of the spectrum?

Thanks and kind regards,

User avatar
Steven Sesselmann
Posts: 674
Joined: 27 Apr 2015, 11:40
Location: Sydney

Re: PRA Tolerance Threshold

Post by Steven Sesselmann » 11 Oct 2016, 10:33


A shape tolerance of 10 will let almost all pulses through. and if you have a good detector and low count rate it can still produce a reasonable spectrum, but when you tighten up the tolerance to 1.0 or even less 0.7 you also tighten up your resolution.

You sacrifice counts for better resolution or vice versa, this is the penalty for doing spectrometry with a sound card which has a relatively low sample rate. If your sound card can do 192 kHz then you can get better results without losing too many counts.

PRA takes 16 samples of the pulse regardless of the sample rate, so at 48 kHz the sampling time is 16 samples * (1s/48kHz) = 333 µs, but at 192 kHz its only 83 µs.

If we did not amplify and stretch the pulse from the NaI detector it would only be a couple of µs, but we would have to sample it at some huge sample rate, which is perfectly possible but not with a PC sound card, also not necessary because of the low activity sources handled by most laboratories.
Steven Sesselmann | Sydney | Australia | | | |

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest