Why Tolerance Correction in PRA ?

PRA, BeqMoni, Theremino, Fitzpeaks and beyond!
Post Reply
User avatar
Steven Sesselmann
Posts: 613
Joined: 27 Apr 2015, 11:40
Location: Sydney
Contact:

Why Tolerance Correction in PRA ?

Post by Steven Sesselmann » 01 Oct 2017, 08:36

Many of you might have noticed a setting in PRA called Tolerance correction, which by default is set to zero, and a recent post in Detectors prompted me to explain what this is. This feature did not exist in the earlier versions of PRA, it is something Marek introduced in one of the newer versions, I can't recall exactly which one.

Essentially Marek noticed how there was a strong correlation between distortion (malformed pulses) and pulse height. Let me explain, PRA compares each incoming pulse to a mean pulse sample and assigns to it an arbitrary number which is a measure of distortion, this number is saved along with the recording file. Marek noticed that the higher the distortion value was, the further to the left in the spectrum it would appear, and so he introduced tolerance correction, an algorithm which compensated for this.

The benefit of tolerance correction is that it moves the under-sampled pulses to the right without changing the number of counts.

In the following screen shots I have recorded a spectrum of an electronic pulse generator, with pulse shape threshold set to 10. With this setting no counts have been filtered all 90,658 pulses are counted.

Leaving the pulse shape threshold at 10, I incrementally increased the tolerance correction, and monitor how the resolution improves. Both standard deviation and resolution improves without any loss of counts.

Finally I achieve almost the same result as I do with pulse shape filtering, but without losing 50% of the counts.

The level of tolerance correction required for your system will vary from setup to setup, so you have to play around with it and find out what works best for you.

Note that tolerance correction appears to do nothing when your pulse shape threshold is set high, this is because the malformed counts have already been dropped, so set the shape threshold to 10 if you want to try this.

Steven
Correction-RAW.png
RAW Import
Correction-20.png
Correction = 20
Correction-40.png
Correction = 40
Correction-60.png
Correction = 60
Correction-80.png
Correction = 80
Steven Sesselmann | Sydney | Australia | gammaspectacular.com | groundpotential.org | rephopper.com | beejewel.com.au |

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest