First experience

All new members, your first post should be an introduction here;
Forum rules
Welcome, these forums are for professional and amateur scientists to discuss matters relating to radiation detection, and to facilitate friendly and honest collaboration. It is a requirement for all members to register using their real names and location. If your name has not already been used, feel free to use it as a login handle, otherwise add your name and location to the bottom of your posts as a default signature. Thanks for cooperating.

If you have a rad related business and want to post information about your products you may do so in the trading post, providing you first introduce your business formally in the introductions forum and add your business details in the signature sections so it is clear to everyone that your post has a commercial purpose.
Post Reply
User avatar
Posts: 1
Joined: 13 Sep 2015, 02:25

First experience

Post by mcmaster » 13 Sep 2015, 02:39

(repost from old forum)

First spectrum, calibration issues, etc: ... on20150909

Comments on the GS experience: ... evaluation

About myself: I have two interests in nuclear tech. First, I occasionally use the UC Davis cyclotron for electronics testing and want to better understand whats going on. Second, I am/was trying to get a SEM running and was looking into options for EDS/EDX.
John McMaster
Mountain View, CA

User avatar
Posts: 929
Joined: 27 Apr 2015, 11:40
Location: Sydney

Re: First experience

Post by Sesselmann » 13 Sep 2015, 14:29


Thanks for coming across to the new forum, I really need to change the "forum" link in the footer, but I can't for the life of me figure out where to change it, the Open Cart software is a web of interlinked php pages.

I read your detailed review of the GS-1100A2 and will make a few comments on the review, someone else might want to comment on your experiments.

I won't comment on the price other than to mention I use quality new components, give a generous $50 student discount to anyone on a tight budget (student or otherwise) and give a lifetime guarantee and after sales support.

On the issue of current draw, the GS is fully powered by the USB port on a PC which has a typical limit of 500 mA , this is for most detectors plenty, but for some detectors that are designed for NIM equipment, the voltage dividers on the PMT's might have very low value resistors (designed to maximise linearity) and for this reason draw too much current. This could cause a significant voltage drop. Also the user must understand that the 1 M ohm coupling resistor inside the GS forms part of the voltage divider chain, because the voltage setting on the switch assumes infinite impedance. True voltage on the BNC is VGS*(Rd/(Rd+Rl)) where VGS is the voltage setting on the switch, Rd is the divider resistance and Rl is the load resistor.

The GS-1100A2 has now been superseded by the GS=1102-PRO and for higher voltages GS-2002-PRO which comes standard with an SHV connector.

The manual has had a few updates since the one you are referring to, but it is still only intended as a quick setup guide, and not as a comprehensive introduction to spectrometry. Regarding the dogmatic writing style, I will try my best to make the next edition manual a little more pragmatic :)

“choo-choo 1 2 3: ??? Apparently a Disney TV show reference"

No, it's what musicians say when they test if the microphone is working ... "choo, choo, 123"

Height calibration / InTune
InTune is a nice little program, but it's no longer required for the pulse shape acquisition process in PRA.

The reference to sound card volume adjustment in PRA is a guideline for users to set the volume to such a level that the full range of the ADC is utilised. The arbitrary unit range is -100 to +100 arb.u. this is divided into 16 bits (65,536 bins when set to 0.01 bin size)) For a NaI(Tl) spectrum I usually suggest that the level is set so the background peak from K40 (1460 KeV) appears at around 50 arb.u. This would give the spectrum a range of 0-100 arb.u. = 0-3000 keV. The user may choose to increase or decrease the volume depending on which part of the spectrum is of interest.

There are several ways to select the region of interest;

1) select a bin close to the peak and press d, f or g on the keyboard, this will select a range of one two or three standard deviations

2) select a bin to the left of the peak and press "b" for begin, then select a peak to the right of the peak and press "e" for end.

Pulse pile up PPU, is a problem in all spectrometry systems, but is more pronounced in sound card spectrometry because the sample rate is lower, to solve this problem a clever system of pulse shape filtering is used to discard (hide) badly formed pulses. In PRA 10 and above each pulse is recorded and assigned a distortion value, and via the advanced pulse filter function, the operator can set the filter to an acceptable level of distortion, PRA will then hide any pulses which fall outside this tolerance. This is very different to the earlier versions of PRA, where the pulses were just dropped.

All NaI detectors are non linear to some extent (some more than others) this needs to be compensated in the software. PRA offers a simple linear calibration method or a multi point calibration method with interpolation, this system works well when you have many calibration points. It is also worth trying some of the other software programs, such as Becqmoni which allows a polynomial function to be used. Theremino uses a GUI type slide adjustment.

The software author of PRA is Marek Dolleiser, a link to his his web site can be found on my home page under software. He is always prepared to listen to suggestions and may implement new ideas if they are of a general nature.

Even the detectors I make myself have some non linearity, the reason is simply that a PMT can't supply enough current to the lower dynodes when a large pulse flows through. To a certain extent this is improved by adding capacitors to the bottom three dynodes, and even this will fail if the count rate is high, because the capacitors won't have enough time to recharge between pulses. This brings us back to what I mentioned further up, where high end NIM systems have low value resistors (100's of ohm instead of M ohm's) in the divider, allowing more current to flow through the divider, thereby feeding the lower dynodes more current. That said, most of the people using the GS are students and universities, who mainly work with low activity sources.

Posts: 56
Joined: 29 Apr 2015, 06:23
Location: Växjö, Sweden

Re: First experience

Post by pietkuip » 14 Sep 2015, 08:52

If there was lead shielding, the peak at 87 keV ("what is this?") is probably K-alpha x-rays from lead.
Pieter Kuiper, Växjö (Sweden)
Course lab: 3 mCi neutron source; five 3" NaI(Tl) detectors, CdZnTe detector (Amptek); lead bricks, two GDM 20

User avatar
Posts: 89
Joined: 27 Apr 2015, 03:49

Re: First experience

Post by brehwens » 15 Sep 2015, 04:33

Hi John!

Ok, lots of info to go through, so I will just mention what popped up first in my mind.

I see you use one of those "square" detectors (Bicron/Alpha spectra). I have had 10 of these, and first of all they are very moderate in resolution, often in the range of 9% FWHM @ 662 keV. This leads to broad peaks and difficulties separating adjacent peaks. Importantly, I found these detectors to be quite nonlinear by default. This is probably due to their indended use in a vehicle portal monitor, and they are most likely designed to perform optimally for this task. I found that modifying the VD (cutting the detector top open is required to do this) can improve this, but I don't think the juice is worth the squeeze.

All PMTs are nonlinear to some extent, but given the correct voltage I have never experienced that the GS (I have used both the GS1100A and GS2002pro) has limited what I can get out of my probes (and I have used around 7 probes with the GS, with voltages ranging from 600 to 1300V, and crystals between 6mm and 76 mm thick). It is much more a software issue, since the softwares are free and advanced functions are time consuming to implement. If you want to go hardcore, give fitzpeaks a try. No limits there, but a steep learning curve.

Second, I think you have some serious problems with your setup. The very last spectrum on your page, titled "background" does not look like a normal background spectrum. In fact, it looks very much like the spectrum you get in PRA if you set the lower threshold to "0", meaning that the leftmost peak you see is not a real peak but PMT noise. In fact, setting the lower threshold to "0", and collecting a "spectrum" is a good way to see where to put the cutoff. But this cutoff must be set, or the PMT noise will be a very large part of the counts you are getting.

A normal background spectrum would have a steep increase in counts from appr 20 keV up to 120 kev or so, and then a smooth slope down. Depending on your detector and the background and measurement time, you might see some peaks at 511 keV and most prominently also 1461 keV, but the building material of your house could also leave its mark in the spectrum.

I agree that the manual could use with an update, but I don't really think it is that complicated to get a GS and a functioning probe up and running. When I started out, with zero previous experience, I had watched Steven's Youtube setup tutorial and that was essentially enough I think.

Many of the settings in PRA are not related to the pulse height histogram. If you focus on those settings, the audio in settings, and making sure you get a good audio signal you should be fine. Laptop internal sound cards can have lots of noise, and I guess this applies also to stationary computer sound cards. An external USB sound card (creative X-Fi Go!) for 40€ was perfect for me, I used this for well over a year with excellent results.

As for the cost: You get a quality built device with lifetime support, and the choice of four free softwares. What can be done with this device can be seen on the forum. Sure, you can build a similar device yourself if you know how. But given that the device pays not only for itself, but for hours of R&D, troubleshooting, testing, I think it is a good deal. Call Canberra/Ortec and tell them you want to interface a "simple" NaI detector to a computer, and see what figures come up...

Not without reason a substantial number of Steven's clients are universities and research labs...

Sure there are things that can be improved! But I know Steven is always open to suggestions, and also has the ear of Marek Dolleiser (PRA creator) so I think that there are every possibility to contribute with ideas and suggestions.
Karl Brehwens
Eskilstuna, Sweden.

Setup: 5 cm lead castle, 2mm copper lining. Gamma Spectacular Pro 2002, Sound Blaster Live! 5.1,
Primary detector: Scionix refurbished 2x2" NaI(Tl) well detector, 7.5% @ 662 keV

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest