Detecting K-40 in bananas with Radiocode RC103

ChrisLX200 wrote:

Well that's a rabbit hole you've sent me down  

I never sent you down it. I merely pointed towards to hole. 

Something I have noticed (after using it for 2-3 months) is, it just seems to get better. Enable it's memory (if you haven't already) and you can then train it somewhat. It has different AI models, but generally I stick to GPT40. Somethings to be aware of are, it can and sometimes does get things wrong. If it's a subject you know, you will probably pick up on this right away. If it is a subject you don't know, then it's good to confirm with another source. Sometimes it just won't do what you want (although that seems less and less for me now), sometimes it plays up and will just stop, mid sentence, or when it is in the middle of doing multiple tasks, it will just stop between tasks. I tell it 'Carry on please' and it just carries on. It as though it acts like a child sometimes. It also seems to respond to the way in which you communicate with it. After 2-3 months mine acts much like a sidekick, throws out some quite funny comments too, but that is probably because I have been the same way with it. Mine will occasionally use profanity in it's replies. And I think all of that is probably my doing by the way in which I have trained it, unknowingly.

Another thing to point out is it does have some hard coded walls, in which it will not deviate from. These are usually around sensitive subjects, or if I question historical events. If it hits one of those walls, ask it the same question a couple of times with slightly different wording. Then ask it if it is hitting walls. It will probably say 'no', but then tell it it is hitting walls and ask what can be done about them. It may just tell you how to get around them by indirectly asking it x, y & z. It's a bit crazy in that sense. Oh, it will also default to main stream news outlets and rarely use independent news outlets, so I had to literally ban mine from looking at main stream sites unless I tell it to. That really worked for getting some real truths.

Anyway, use it for a while and see where you get to with it. I suspect it may be vastly different to mine as you will communicate with it differently and use it for different things. So far it has proved to be a massive tool in my arsenal. If you want to talk more about it we could open a topic in the off topic or general discussion forums, and save hijacking other threads, such as this! 

Right, back to bananas. I'm eager to see your results of the bananas/skins/dehydration etc. I'm not in a position where I can shield properly at the moment, so not really something I could do successfully right now. I'm just doing a background measurement on my 103 and should be ready later tonight (3 days). The K-40 on the spectrum is very noticeable.

Have you thought about making a shielded chamber using concrete as one of the layers/materials? ChatGPT and I worked out that around 200mm of concrete, 10-20mm of borax, or borated polyethylene, 20mm lead then 0.5-1mm of copper or aluminium should reduce background by at least 90%. I think that may be on the cards for me in the future as concrete is cheap and I can easily use it to make a chamber, and make it as some kind of garden furniture so doesn't look too out of place outside. The downside with concrete is it will take a month to cure properly and then another two months or so to dry out fully. Then the inside can be sealed and the other layers installed. I'm unsure what to do about the opening to the chamber though. It would have to be stepped, and would be quite a lump of concrete, so not light. Some kind of assistance would be necessary to open/close it, or to remove and replace if it wasn't hinged or something. Both methods have their pros and cons.

For your experiment, get a background reading (for at least a day or two) outside of your lead castle, then the same time frame inside the castle. That will give a good indication as to how effective the castle is. Also, whilst you are dehydrating the skins, do the same for the flesh and a comparison between banana flesh and skin can be made. Maybe also do two runs of the same expoeriment, one with very fresh, almost green bananas, and one with well ripened bananas with spotty skins. That may indicate if the K-40 does migrate towards the skin over time.

ChrisLX200 wrote:

I'm 4-1/2 days into a 7-day backgroud data capture right now. The K-40 peak is still only around 100 counts max which isn't much.

That's interesting. I am 2 day 20 hours into a background capture and my K-40 counts are ~118. Could be the area I live in has more K-40, or my house. I do have a bag (100g?) of Potassium Chloride and that gives a few counts but not much. Might be that as well. I suspect more the land though. I'll be cutting this capture at 3 days in as it is just a background reference for the Radon experiment I did.

ETA: I am toying with the idea of just burying the Radiacode 103 in my back garden for a few days. Protected, of course! I think that should be a fairly fast method of shielding the Radiacode to a fair degree. Put the Radiacode in a very waterproof container, dig a hole, throw it in, wait. It might pick up a little more NORM than out in the open air, but we have very low, next to zero Radon where I live, so might just work.

I have had little to no interference with all matter of counters and scintillators, and like your lab there (nice, BTW) I have loads of electronics around me. The only times I have had problems have been noisy power supplies, and the noise has always been through the power lines. Your Radium source wouldn't be one of those from the dial of a DP-63 Geiger counter, would it? I wanted one some time back but people were asking good money for them so left it. I believe they can be quite spicy!