My Experimental detection software

[Daniel]

firstly a big thank you to those who without their kindness I would have not been able to explore the hobby, help advice, equipment etc made available.

I have various detectors and software now and always thought the weakest link was my own self doubt on correct detection of multiple sources.

I also disliked the kromek kspec software that the multispec was £1200 and significant to a hobbyist with a kromek RA4S.
So I decided to try and build from what I have learned currently alpha state software with RA4S usb working and demo sampling. I also believe that sound card should be a viable input soon.

ge.png

Kind Regards

Daniel

[Sesselmann]

Daniel,

Writing your own software is a good way to learn about the art and get a much deeper understanding. Well done.

Steven

[Daniel]

Hi Steven,
Thank you for the kind words.
Sound Gamma Spec I believe to be much much harder - than pulling channels out of USB device, it's taken a fair while to get sound results tangible like with theremino, PRA etc.

I'm currently working On IAEA compliance and tangibility, for instance we don't want Am241 on one peak match.

I've got things like the software can tell user the highest channel and guide with visual indicator that they are on it, something the Kromek kspec I struggled with sometimes with lots of peaks becoming a distraction.

The work of Tomy and Theremino team etc I found most useful, by this I believe that the equivalent software offered was as good and sometimes possibly superior- but I personally didn't have confidence to use it and believe I had right results.

Currently while away on holiday running software and checking it's database identification operation, hopefully some operational videos on YouTube soon.

Have you heard anything from George Dowell Geo aka K0FF or Geo electronics? or the other regulars? that I guess have since retired?

Kind regards

Daniel

[Svilen]

I like your CRT green glow design, Daniel.

About Geo Dowell - I’ve seen many people asking the same question in several forums and discussion groups... A letter I sent more than a year ago was also never answered, although he was always very kind and helpful before. This isn’t his typical behaviour, and I can only hope I’m terribly wrong, but based on all of this, I would guess that George is, unfortunately, either incapacitated, or no longer with us. I’ll always remember him sending me extra stuff I never paid for.

[Daniel]

I like your CRT green glow design, Daniel.

About Geo Dowell - I’ve seen many people asking the same question in several forums and discussion groups... A letter I sent more than a year ago was also never answered, although he was always very kind and helpful before. This isn’t his typical behaviour, and I can only hope I’m terribly wrong, but based on all of this, I would guess that George is, unfortunately, either incapacitated, or no longer with us. I’ll always remember him sending me extra stuff I never paid for.

Svilen, I had read that the property had been sold, I kind of hoped that George had at best moved to a smaller / retirement complex.
Time is a funny thing, it has a way of running away with us, I had hoped to share my exploits with George- bit like a student graduating and coming back with a meaningful project.

The green reminds me of cathode ray scopes- my grandad used to use when I was a child.

I'm also looking at PT100 platinum temperature probe, amplifier and esp32 or similar that prints temperature via HID to be driverless temperature compensation

[Daniel]

Still developing this idea slowly, linear calibration -as don't trust cubic / quadratic yet.

I'm wondering ASIO Sound ASIO (Audio Stream Input/Output) is a computer audio driver protocol developed by Steinberg that bypasses the operating system's mixing layers to provide low-latency, high-throughput communication between software and hardware
v intel ?


PIC SCALER
2 INCH Nal
50kg plus lead pig
esp32 pt100 temperature via lan or usb serial


Sound card MCA · Expert Ident
Photopeak Lab 2026.09.06 · Daniel Young ·Live time: 15510.9 s
Detector: NaI(Tl) 2″
Source: audio
Temperature: 26.5 °C
FWHM: measured 8.58% at 662 keV (56.8 keV, 5 peaks)

Standards (public metrology )
Peak search: Mariscotti, Nucl. Instrum. Methods 50 (1967) 309 (2nd difference).
Decay data: IAEA Nuclear Data Services, X-ray and gamma-ray decay data (X-γ).
Uncertainty: ISO/IEC Guide 98-3 (GUM); IAEA TECDOC-1401 (2004) Quantifying uncertainty in nuclear analytical measurements.
Characteristic limits: ISO 11929:2019 (decision threshold LC, detection limit LD, coverage interval).
Energy and efficiency principles: IEC 61452 (nuclear instrumentation — energy resolution and efficiency).
Identification: IAEA laboratory practice — library-directed ROI, companion-line test, Compton-edge veto, Currie LC before a claim. Activity pattern: non-negative least-squares fit of ε(E)·Iγ photopeak templates (SNIP continuum).
Origin: IAEA soil practice — naturally occurring radioactive material (NORM: K-40, U-238 series, Th-232 series) is listed separately from anthropogenic nuclides (Cs-137, Cs-134, Am-241, Co-60). Be-7 is cosmogenic. Counts from 40–500 keV under Cs-137 are mostly Klein–Nishina Compton scatter (edge 478 keV), not extra photopeaks.
Line overlap: Ba-133 356.013 keV and Pb-214 351.932 keV (Δ 4.08 keV) are one peak on CZT/NaI/CsI. IAEA companion test: Ba-133 needs 81 and/or 303 keV; Pb-214 needs 242 and/or 295 keV or Bi-214 609 keV. A lone 352/356 keV peak is not an ID.
Conversion budget: u_rel 9.5% (ε 8% · geom 5% · Iγ 1% · E 1% · t 0.1%)

Calibration
E(ch) = -6.373 + 1.06035·ch
Linear residual vs locks: max 0.00 keV (0.00%), R² 1.0000.
E(ch) monotonic: yes

Detection offset: -2.00 keV mean ΔE (-0.26% · RMS 10.54 keV · 17 lines · kind bent).
King line: Cs-137 662 Δ -0.01 keV. Gain scale 0.99742 (observed/library).
Calibration error vs IAEA X-γ: residual is a fraction of FWHM — fit is usable.

Compensation (as calibrated vs suggested)
As calibrated: keep the current E(ch). Use this for the reported activities.
Offset trim: set offset to -4.369 keV (+2.00 keV vs IAEA X-γ).
Bent: do not apply a single offset or gain. Recapture 3+ locks; check HV, PMT gain and sound-card AGC.
Temperature compensation is on — leave it on while trimming.

Audio / HV: Sound card looks healthy.

Mix %: activity fraction after ε(E)×Iγ (not raw counts). Line χ² checks IAEA branching.
Traffic light: IDENTIFIED ≥2 suit lines or a unique single; POSSIBLE one line ≥ LC; UNLIKELY below LC, missing companions, or Compton edge.

Origin (IAEA soil: NORM vs man-made vs cosmogenic)
Man-made: Cs-137 identified 88%; Sn-113 possible 85%; Ag-110m possible 50%; Lu-177 unlikely 28%
NORM (K-40, U-238 series, Th-232 series): Th-232 possible 54%; Pb-212 possible 54%; Ac-228 possible 33%; U-235 possible 29%; Bi-212 possible 25%; Bi-214 unlikely 88%; Ra-226 unlikely 9%
Cosmogenic: —
40–500 keV: 610515 counts (39.36 cps, 55% of spectrum).
Cs-137 Klein–Nishina Compton edge is 477 keV. Most of 40–500 keV is man-made Compton scatter from 662 keV, not extra isotopes. NORM photopeaks sit on that continuum and are listed above even when below ISO 11929 LD.
Photopeaks 40–500 keV: 52 keV 357σ, 210 keV 84σ, 187 keV 52σ, 249 keV 49σ, 398 keV 46σ, 398 keV 35σ.
Man-made lines rejected as NORM overlap: 208 keV sits on Ac-228 209 keV (NORM) — not Lu-177.

Mix Cs-137 6% · Ag-110m 23% · Ac-228 27% · U-235 41% · Bi-212 8% · Bi-214 5% · Ra-226 6%

IDENTIFIED Cs-137 man-made 88% mix 6% pass 3 suit / 0 close / 3 χ² — source Rel. Eff. · NaI/CsI Rel. Eff. — few isolated peaks
Why: 32 keV unique single line (1470σ)
31.8 X-ray I=6.9% suit 36.6 keV Δ +4.8 1469.5σ A=1.94e+2
36.4 X-ray I=1.4% suit 36.6 keV Δ +0.2 1469.5σ A=9.54e+2
662 γ I=84.99% suit 661.6 keV Δ -0.0 720.1σ A=7.05e+1

POSSIBLE Sn-113 man-made 85% mix 0% pass 1 suit / 0 close / 1 χ² — source Rel. Eff. · NaI/CsI Rel. Eff. — few isolated peaks
Why: 392 keV one line above Currie LC (ISO 11929)
392 γ I=64.9% suit 397.6 keV Δ +5.9 45.8σ A=4.34e+0

POSSIBLE Th-232 NORM 54% mix 4% check 1 suit / 0 close / 8 χ² — 239 keV net 25175 ≥ LD 1312 (ISO 11929 detection limit, total uncertainty)
Why: 239 keV net 25175 ≥ LD 1312 (ISO 11929 detection limit, total uncertainty)
239 γ I=43.6% suit 238.6 keV Δ +0.0 64.9σ LC 639 LD 1312 MDA 4.08e-1 Bq
338 γ I=11.3% miss — 0.0σ LC 681 LD 1399 MDA 1.98e+0 Bq
511 γ I=22.5% miss — 0.0σ LC 806 LD 1656 MDA 1.90e+0 Bq
583 γ I=85% miss — 0.0σ LC 761 LD 1562 MDA 5.11e-1 Bq
727 γ I=6.7% miss — 0.0σ LC 581 LD 1195 MDA 6.01e+0 Bq
861 γ I=12.4% miss — 0.0σ LC 623 LD 1280 MDA 4.33e+0 Bq
911 γ I=25.8% miss — 0.0σ LC 369 LD 760 MDA 1.33e+0 Bq
969 γ I=15.8% miss — 0.0σ LC 198 LD 409 MDA 1.27e+0 Bq

POSSIBLE Pb-212 NORM 54% mix 4% check 1 suit / 0 close / 2 χ² — 239 keV net 25175 ≥ LD 1312 (ISO 11929 detection limit, total uncertainty)
Why: 239 keV net 25175 ≥ LD 1312 (ISO 11929 detection limit, total uncertainty)
239 γ I=43.6% suit 238.6 keV Δ +0.0 64.9σ LC 639 LD 1312 MDA 4.08e-1 Bq
300 γ I=3.3% miss — 0.0σ LC 701 LD 1441 MDA 6.12e+0 Bq

POSSIBLE Ag-110m man-made 50% mix 23% check 2 suit / 4 close / 10 χ² 10.05 source Rel. Eff. · NaI/CsI Rel. Eff. — few isolated peaks
Why: 687/707/658/678/764/1384 keV seen; 885/937 keV companion missing
658 γ I=94.3% close 661.7 keV Δ +3.9 717.3σ INTERFERE A=6.33e+1
678 γ I=10.5% close 661.7 keV Δ -15.9 717.3σ INTERFERE A=5.87e+2
687 γ I=6.7% suit 698.8 keV Δ +11.8 504.2σ INTERFERE A=7.17e+2
707 γ I=16.4% suit 698.8 keV Δ -7.9 504.2σ INTERFERE A=3.03e+2
764 γ I=22.7% close 733.8 keV Δ -30.2 112.4σ A=6.63e+1
818 γ I=7.33% miss — 0.0σ
885 γ I=72.7% miss — 0.0σ
937 γ I=34.3% miss — 0.0σ
1384 γ I=24.3% close 1413.5 keV Δ +29.2 10.0σ A=2.43e+0
1505 γ I=13.02% miss — 0.0σ

POSSIBLE Ac-228 NORM 33% mix 27% check 2 suit / 2 close / 15 χ² 49.77 Ac-228 needs 338 / 911 / 969 keV
Why: 209/727/270/239 keV seen; 911/511/583 keV companion missing
209 γ I=4.6% suit 209.9 keV Δ +0.7 84.0σ INTERFERE A=1.08e+2
270 γ I=3.8% close 249.2 keV Δ -21.1 49.3σ A=7.11e+1
328 γ I=3.4% miss — 0.0σ
338 γ I=11.3% miss — 0.0σ
463 γ I=4.6% miss — 0.0σ
795 γ I=4.8% miss — 0.0σ
911 γ I=25.8% miss — 0.0σ
965 γ I=5.5% miss — 0.0σ
969 γ I=15.8% miss — 0.0σ
1588 γ I=3.7% miss — 0.0σ
239 γ I=43.6% close 249.2 keV Δ +10.5 49.3σ A=6.19e+0
511 γ I=22.5% miss — 0.0σ
583 γ I=85% miss — 0.0σ
727 γ I=6.7% suit 733.7 keV Δ +6.4 116.7σ A=2.21e+2
861 γ I=12.4% miss — 0.0σ

POSSIBLE U-235 NORM 29% mix 41% check 1 suit / 1 close / 4 χ² 72.66 NORM Rel. Eff. · NaI/CsI Rel. Eff. — few isolated peaks
Why: 1 suit / 1 close of 4
144 γ I=11% miss — 0.0σ
163 γ I=5.08% miss — 0.0σ
186 γ I=57% suit 186.6 keV Δ +0.9 52.3σ A=4.71e+0
205 γ I=5.01% close 209.9 keV Δ +4.6 84.0σ A=9.95e+1

POSSIBLE Bi-212 NORM 25% mix 8% check 1 suit / 0 close / 2 χ² — chain evidence for Ac-228
Why: chain evidence for Ac-228
727 γ I=6.7% suit 733.7 keV Δ +6.4 116.7σ A=2.21e+2
1621 γ I=1.49% miss — 0.0σ

UNLIKELY Bi-214 NORM 88% mix 5% pass 1 suit / 2 close / 10 χ² 1.23 chain evidence for Ra-226
Why: chain evidence for Ra-226
609 γ I=45.16% close 625.6 keV Δ +16.3 387.5σ INTERFERE A=8.37e+1
768 γ I=4.97% miss — 0.0σ
934 γ I=3.19% miss — 0.0σ
1120 γ I=14.78% miss — 0.0σ
1238 γ I=5.79% miss — 0.0σ
1378 γ I=4.1% close 1349.8 keV Δ -27.8 9.3σ A=1.45e+1
1408 γ I=2.5% suit 1413.4 keV Δ +5.4 10.1σ A=2.43e+1
1509 γ I=2.2% miss — 0.0σ
1730 γ I=3% miss — 0.0σ
1764 γ I=15.17% miss — 0.0σ

UNLIKELY Lu-177 man-made 28% mix 0% check 1 suit / 0 close / 2 χ² — 208 keV sits on Ac-228 209 keV (NORM) — not Lu-177.
Why: 208 keV sits on Ac-228 209 keV (NORM) — not Lu-177.
113 γ I=6.2% miss — 0.0σ
208 γ I=10.4% suit 209.9 keV Δ +1.6 84.0σ INTERFERE A=4.79e+1

UNLIKELY Ra-226 NORM 9% mix 6% pass 1 suit / 2 close / 7 χ² — NORM Rel. Eff. · NaI/CsI Rel. Eff. — few isolated peaks
Why: 186/242/609 keV seen; 352 keV companion missing
186 γ I=3.533% suit 186.6 keV Δ +0.4 52.3σ INTERFERE A=7.60e+1
242 γ I=7.19% close 249.2 keV Δ +7.2 49.3σ INTERFERE A=3.76e+1
295 γ I=18.28% miss — 0.0σ
352 γ I=35.34% miss — 0.0σ
609 γ I=45.16% close 625.6 keV Δ +16.3 387.5σ INTERFERE A=8.37e+1
1120 γ I=14.78% miss — 0.0σ
1764 γ I=15.17% miss — 0.0σ

[Sesselmann]

Daniel,

Lot's of information, but I don't really understand the question here.

Steven

[Daniel]

Hi Steven,

I've always doubted that the results that were accurate in my case, possibly the calibration of the equipment or temperature drift.

Possibly that in certain cases say theremino identification tool by our Associate Tomy that poor calibration, drift and "enthusiasm" on my part would come up with reports of things that simply were not there.

Because of potential calibration errors and the fact commercial accreditation on such software must not give false positives-the software is out of the reach of most hobbyists, schools etc.

My goal was to take say Geo's Cs-137 soil sample, make the software aid calibration and use IAEA lab grade identification maths on a known sample.

I've yet to get to the point of sample activity, but this also can be factored into to the software. I'm going to use Geo soil and smoke detector as reference point.

Wrapping up and trying to answer my own question " is it possible to have hobbyist Gamma spectrometer that is equivalent IAEA lab compliant to the standard - even if it would never be accredited as user can calibrate"

The second problem since it is a decade plus research is finding someone in a group that can look at CSV or SPE and report and say if it is tangible.

AI says the reports and results are tangible, but I simply don't trust it is telling truth.

I did think of emailing our retired friend from scionix perhaps as Geo isn't available to ask.

Apologies if this is something not suitable, but I'm not sure where else to ask.
Kind regards

Daniel

[Svilen]

Hi Daniel,
I think you may be expecting too much from a hobby NaI system. It is useful for identifying strong, well-separated gamma emitters, but it can’t realistically match IAEA-lab performance in calibration, linearity, detection efficiency. While a Cs-137 soil sample will be recognisable with only a rough calibration, temperature drift, efficiency and over-enthusiastic peak identification can always create misleading results - this is also part of the fun ;)

[Daniel]

Hi Daniel,
I think you may be expecting too much from a hobby NaI system. It is useful for identifying strong, well-separated gamma emitters, but it can’t realistically match IAEA-lab performance in calibration, linearity, detection efficiency. While a Cs-137 soil sample will be recognisable with only a rough calibration, temperature drift, efficiency and over-enthusiastic peak identification can always create misleading results - this is also part of the fun ;)

Svilen, I agree and low level exempt samples are more difficult, I have a few tricks such as ashing a sample and a lead pig for NORM AND TEMORM

Maybe a polite email to a professor may get a response, one of the problems is actually having someone who can look and say great or- rubbish.

I don't trust AI marking my work.