Show HN: RadiaCode – Python Library for RadiaCode-10x Radiation Detectors | Hacker News Reader

Show HN: RadiaCode – Python Library for RadiaCode-10x Radiation Detectors

github.com

42 pointscdump09 comments

Hi HN,

I want to show you my project: RadiaCode, a Python library to interface with RadiaCode-10x radiation detectors and spectrometers. With RadiaCode, you can:

- Collect real-time radiation measurements - Analyze spectra for insightful data interpretation - Control device settings via USB or Bluetooth - Explore a web interface example for remote monitoring

It's available on PyPI, open-sourced under the MIT License, and you can find the code with examples on GitHub.

Open articleView on HN

> Analyze spectra for insightful data interpretation

I can't see much in the way of code to do that.

How can I calibrate for and lock onto the K-U-Th peaks.

Where do the stripping ratios go?

Can I generate a NASVD kernel over time and landscape an enhanced sharpened spectrum run over time to clearly see changes from radon gases rolling in, new decay products from (say) a waste catchment breaching and leeching closer to my location.

We had that in gamma spectral analysis libraries in the 1980s and 1990s.

eg: Grasty & Minty (1995) https://www.ga.gov.au/bigobj/GA7667.pdf covers a lot of the basics for setup, calibration and aquisition. (There are other tomes for analysis).

This library is designed to obtain spectrum data from the Radiacode device and manage calibration coefficients for the spectrum. It serves as an interface to the hardware, providing raw and calibrated data, but does not perform advanced spectral analysis itself. The provided examples demonstrate how to use the library for basic data retrieval and calibration.

Any higher-level processing, such as peak identification, stripping ratio adjustments, or NASVD-based spectrum enhancement over time, would need to be implemented on top of this library or using its output in external analysis tools. The data it provides can be integrated into custom pipelines for spectral analysis, environmental monitoring, or historical trend detection, much like the analytical techniques used in past gamma spectral analysis libraries.

I welcome the new library! I have used a RadiaCode 103g for more than a month. I also have a more conventional NaI:Tl 5cmX5cm scintillator+PMT in my basement with 95 lbs of lead shielding (as lead shot). I'm painfully familiar with detector drift and recalibrations required with conventional (amateur) detectors. In my opinion the RCs are game changing: their calibration ranges from good to excellent, in that I can import a spectrum into the (exceptionally useful free Sandia National Labs) tool InterSpec, extremely quickly identify peaks and the relevant radionuclides. The RC is very compact with a well-defined detector volume (because it IS relatively small). If there is a weakness, it is that it's mostly targeted toward running on a smartphone and support for iOS vs. Android has lagged a bit, though very recently an excellent update has appeared for iOS. I have tested in on a variety of spectra (137Cs, 232Th, 60Co, 133Ba) and for long background counts and am quite happy with the resolution and stability. I have yet to use its interface to GPS data (I already have a bGeigie Nano built from a Safecast kit). Its impact is that almost everything one might like--count rates, dose rates and net dose, data logging, and GPS interface is all there in a sub pocket-size unit.

1. RadiaCode is a repo for running IOT connected detectors from a specific vendor

2. Detectors are judged based on the mass and resolution. The resolutions for each vendor specific variant are at least 50-200% higher than what is possible with 'passable' material. Additionally the spectra shown indicate a very small detector mass, meaning these would not be much more sensitive than GM counters, and possibly less reliable in other metrics.

a. Why are the resolutions for CsI(Tl) 9.5% and not closer to 5%? (in this case, lower is better)

b. Are the sensors calibrated for temperature and other systematics? how long do the calibrations last and to what % of energy resolution?

c. Are the sensor responses linear? The documentation [1] shows that these can be very high error rates (up to 40% for "some isotopes"). Such high error rates makes this hard to see how dependable or useful the device would be.

[1] https://coda.io/d/radiacode-cyberspace_dTtFtorf-v-/Quick-Gui...

Thanks for your detailed comment! Just to clarify, I'm not affiliated with RadiaCode; I've simply developed an open-source Python library to interface with their detectors, making it easier to collect, analyze, and visualize data

Be careful of trade marks! If you are looking for roles check out some of my past submissions and our website.

I own one, it's a $300 Gamma Spectrometer that fits in your pocket. Closest thing I own to a real life Tricorder.

I've tested it out on a smoke detectors. Successfully identified AM-231 and its decay chains from like a foot away. Tested it out on my dad's basement, and it was able to pick up the whole radon decay chain (Thorium, Polonium, Lead, etc.)

It's not a laboratory grade tool. You can point it at radioactive stuff and the spectra lines up perfectly if you Google the charts. There is a community of folks who walk around with them to identify radioactive deposits. I own a GM counter and it's not even in the same league in sensitivity or capability, it would be like comparing a flip phone to a smartphone.

Unfortunately 300$ price point brings up a lot more questions about utility to me.

You can have all the things I mention above at a slightly higher price point.

It is not hard to make these and the really hard part is guaranteeing the quality so they aren’t treated as novelty toys.

> c. Are the sensor responses linear?

No. Not even professional geophysical survey spectrometer sensors + interpretation software have a linear response to increased sources.

Two main reasons:

* The electronics to count scintillations (energy bursts at some X keV or MeV energy level) get flooded and overwhelmed; there's a "dead zone" following each detection count, as events increase in number, the percentage of dropped | undetected events increases.

See: Grasty & Minty (1995) page 32 Sec. 4. DATA PROCESSING SubSec. 4.1 Dead−Time Corrections

  Gamma−ray spectrometers require a finite time to process each pulse from the detectors. While one pulse is being processed, any other pulse that arrives will be rejected. Consequently, the ‘live’ time of a spectrometer is reduced by the time taken to process all pulses reaching the analyzer. For large volume airborne gamma−ray spectrometers with their associated high count rates, the dead−time can be significant and corrections must be made, particularly when measuring on calibration pads.

~ https://www.ga.gov.au/bigobj/GA7667.pdf

* Stacking signatures from sources creates complex full spectrums; teasing full spectrums apart to extract component source decay signatures is messy and involves a degree of probablistic guesswork .. not very linear in a real world enviirnoment.

eg: The Cobalt 60 Gamma Spectrum has multiple peaks

The 1332 keV Co-60 peak bleeds into the "fat" K40 Potassium 1461 keV peak, throw in varying amounts of naturaly abundant Uranium and Thorium

https://www.gammaspectacular.com/blue/gamma-spectra/Co-60-ga...

https://www.gammaspectacular.com/blue/gamma-spectra/k-40-gam...

https://www.gammaspectacular.com/blue/gamma-spectra/norm-spe...

and it becomes clear that correctly identifying which elements are responsible for an observed full spectrum is a challenge with often multiple solutions.

The great challenge is correctly "guessing" the original proportions of commonly abundant radiation sources that share common daughter decay products. These are the sources that are listed as having a greater degree of error.

Reply on news.ycombinator.com