I’m always on the search for something impactful, something that really matters. This time I stumbled upon Wildlife Protection Solutions. I reached out to Eric and one of their initiative gets my attention – the water poison detection for preventing poaching. The poachers use poison to contaminate waterholes for the easy kill: temic & karbadust in South Africa, and cyanide in Zimbabwe. The main victims are lions, elephants, and vultures. As usual, I am always pumped by projects with potential for real-world tech solutions, so I jump right into it and start the research.
Researching the unknown
I started with the NanoStat, a device Eric mentioned. It is open-source which is always good news – it allows modifications and it is low-cost. The idea is to replace the sensor with one that can detect the poison. We focus on temic and karbadust as Cyanide is too dangerous and Eric’s team is based in South Africa where poachers mainly use temic & karbadust so it makes more sense for the field tests. Meanwhile, I also found this interesting article on Cyanide detection using light so will keep it here for future reference.
My chemistry knowledge is zilch, so I hired someone competent to do the research. The report is very discouraging. No sensor can detect these chemicals and all of the commercial devices are quite pricey and can do only one-time probes so become an impractical choice. We need something relatively low-cost with 24/7 monitoring. Modifying the existing commercial devices would be too much work and quite expensive so I don’t even explore this idea.
At this dead end, Ilia, my hardware engineer buddy, comes on board with fresh ideas. A brilliant and very modest guy! One of his recent projects is a brain-to-muscle signal-measuring prototype, which he believes could one day enhance athletes’ performance by linking dietary patterns to brain signal intensity. Plus, he’s a boxer – a great anti-poaching backup plan if the tech doesn’t work out!
Ilia found this article which leads us to the QCM sensors. A promising new path! We just connect the sensor to the NanoStat device and all good right? Well not really, the answer from the NanoStat developers is that the QCM needs MHz frequency waveforms, which the NanoStat is not able to produce easily. Back to the drawing board! This is when I came across an article from Rens J. Horst and his prototype using QCM sensors Yes! Finally some good news! I also found OpenQCM – a commercial open-source device which uses the same technique. I contacted them and apparently, they used Rens’ work in their initial development. I love the open-source world. Sharing is caring!
ChatGPT also joins the research to give us more insight into how this wizard tech works. QCM works by binding chemical molecules and measuring frequency changes. According to ChatGPT, we are on the right track. Simple and brilliant! We are happy and ready to purchase the parts and start building! Just when I am about to hit the buy button I get a reply from Rens. He thinks that it would be impossible to create a film on the QCM sensor that will bind to the carbamate molecules. The OpenQCM team confirms it. I am ready to sh**t myself!
As an alternative, Rens mentions the UV-VIS technique which uses changes in the light spectrum to measure the presence of given molecules. With this new direction, I come across AS726X and the more production-ready rodeostat. Before going this road Ilia thinks regardless of the small chance we should still give the QCM sensors a try so I place an order. It is only around 50eur and a very low effort to test it.
With this mixture of excitement and disappointment, we move to prototyping. Time will tell if wasted our time!
Prototyping with little hope
The QCM sensors arrive and it is time to do some very simple tests with glucose. These sensors are tiny, fragile and impossible to solder so it requires some innovators thinking. One sensor is collateral damage in the process. This is where we are now. More updates to follow!
Good news, bad news!
Had another call with Eric recently and waterhole poisoning is no longer an issue so time to say goodbye to this project.
We walk away with few lessons learned, and the bitter disappointment over the unfinished business!
