High baseline "noise" with Theremino MCA. How to improve?

[stelios70]

I have two relevant devices. A Radiacode 103 and an Sohtech Orion IIIA PRO Gamma Spectrometer with an Alpha Spectra scintillation probe, 1.15" x 1.15" x 3" Na(Tl) crystal, FWHM resolution at 661 keV: 13.6% +/- 1%. The Orion with the probe came pre calibrated to Cs-137 31keV and 661keV peaks. The first problem that I had was that I was getting a very low cps in ambient light condition. It seemed that there was light leakage to the probe and after I covered the probe with aluminum foil, this problem was solved. The main issue now is that I am getting a lot of "noise" as the background cps tend to be around 160 and they go up to 250 when using a known source. That makes peaks very difficult to identify. Please see the background spectrum and the spectrums for Lu-176 (15 gramm of Lutetium metal) and Am-241 in Theremino MCA. For comparison I also have the spectrums of the same sources as recorded by the Radiacode software.
Any advise on how to improve the Theremino outputs and/or reduce background "noise" are highly appreciated.

[Sesselmann]

Stelios,

Thanks for registering and writing your first post, but please go and introduce yourself in the Introductions forum as well.

Frankly I had never heard of a Sohtech Orion IIIA PRO, so I looked it up and it looks like a nice copy of my GS-PRO.

I have never seen or tested this spectrometer, but the chances are that your problem lies with the detector, any NaI crystal with a resolution of 14% is pretty useless, most likely it has taken in some moisture and has a wet crystal.

If the problem was electronic noise you would see it as a tall peak in the first couple of bins on the left hand side of your spectrum.

When it comes to gamma spectrometry, the detector is always the weakest link, so you should spend most of your budget on a good detector.

Steven

[Mike S]

Stelios,

If you are trying to figure out a new detector, new MCA, and new software all at the same time you should brace yourself for a challenge. I have been there and made it through, but it can be frustrating at first.

Are you using a shield for your measurements? That will make a big difference in the quality of your spectra.

I have a Radiacode 103 also and it certainly is easier to use because it doesn't have the adjustments of the more advanced setup you have now. Since I am unfamiliar with your detector and MCA, I can only offer some suggestions for Theremino based on my experience with my GS-USB-PRO and various detectors.


1. Start with as little software distortion as you can. Turn off the equalizers and definitely turn off the IIR Filter for now. Resolution compensation and Gaussian deconvolution should definitely be off (and stay off in my opinion).
2. Adjust your audio input settings to get the best signal. At a glance it looks like you have a lot of bins and a lot of gain. I use x5 bins and an Audio gain of 1.0, but that will depend on your equipment. My goal is to get my spectrum so the energy trimmer is near the middle of the slider when a mid-range peak is aligned. Don't forget to check the microphone gain/trimmer in your operating system (I usually have it set at 100%, but sometimes I turn it down to make other adjustments work out). For all of this, the "view pulses" window can be helpful. Also, try setting your pulse polarity to "Auto"; I've seen it work poorly with the wrong pulse polarity setting but it worked well enough (surprisingly) that it wasn't immediately obvious. Because you have your minimum energy set to 22keV, it makes me wonder if you have a lot of noise that's making it harder to get clean pulses; I usually have my minimum energy set between zero and 5keV so I can see if noise is creeping in during pulse collection.
3. Once the gain is looking pretty good, see if you can dial in your baseline test settings. Your settings seem generally reasonable, but there may be benefit from making adjustments. For what it is worth, my settings are usually 80uS/240uS/20%/15% for my favorite PMT based detector. I would definitely try it with "Max Slope" turned on (I always have that on).
4. Once all that is as good as I can get, I collect about 1000,000 to 1,000,000 pulses from a strong sample and then start making very small adjustments to the "Energy Linearizer". It ends up looking something like the opposite of the response curve of my detector.
5. I will sometimes use the IIR Filter set to zero (it still filters) to apply some smoothing or so it matches the references (which always get what appears to be a IIR filter of zero). Keep in mind that the IIR filter can cause very misleading artifacts, even when set to "0", particularly if the total number of pulses collected is low.
6. In my opinion one should never use the "Resolution compensation" or "Gaussian deconvolution" features under the options tab. I just pretend they don't exist. If your spectrum is good enough for them to work then you probably don't need those features. When these features are used with poor spectra, they just make a deceptive mess (again, in my opinion).

You might try PRA or Impulse and see if you like that software more. I believe the pulse height detection algorithm and energy calibration method of PRA (and Impulse) are better than Theremino, but they don't have the same level of display flexibility (multiple references, easy to use isotope selection, fast and easy zoom and scaling). On the flip side, Theremino has many ways to make adjustments that (in my opinion) trash the spectrum (not permanently, you can back them out).

Good luck!