Hmm, this is interesting. For now, I have Intel Nuc Extreme, but with those being bought off by Asus, this might be something to consider for me in the future.
Aside from pure curiosity and or research purposes...
Is there any everyday/practical apps/uses for something like this, for the avg joe?
Building something for 500 that would normally be 10k, already has me intruiged.
I brought this up in a comment in another thread. I found this page because I wanted to see whether it was feasible to monitor nutrient concentration in an aeroponic solution. Typically the solution circulates between the tank and the root box, so the concentration in nutriments will decrease gradually. Near-IR isn't enough to figure out these details I believe (I'm not a scientist), but it is enough to get a fingerprint of the nutrition "equilibrium". These solutions are probably the most technical aspect of aeroponics, so I wanted to use commercial solutions bundled as three bottles you're supposed to mix in various proportions. I think that if you explore the concentrations space using such a spectrometer you may be able to train a model to interpolate the concentrations from spectrographs measured in "production".
Anyway I think this is best to address the shortcomings of circulating the solution, for instance use a tiny tank, circulate for a few hours/days, refill, dosing as you go. A spectrometer would be still be cool in order to get real-time concentration diffenrential for each solution component. Identifying individual chemical species could be accomplished I think using far-IR spectroscopy with a model trained on data measured using more expansive spectrometry techniques.
I haven't grown anything yet, I want to build a sap flow sensor first.
* any microcontroller with multiple ADC channels, for example ESP32 if you want it to be wireless as well...
* a heat generator (can be as simple as a transistor)
* a few thermistors, which you will have to calibrate (pretty easy if you have access to a thermometer, an electric kettle and ice, basically dunk the thermistor in a glass, put the thermometer in the glass, have the microcontroller output measurement number, and measurement value over serial port, now heat water to boiling point, make a table with a first column pre-entered with the thermometer gradation, then pour it in the glass, each time it passes a gradation you look at the scrolling list of values and note it in the second column of your table. Now you have a monotonically decreasing list of measurements, some of them marked as a thermometer gradation passing. Make a preadsheet and use the value of the fixed resistor in the voltage divider [the other resistor was your thermistor] together with the standard NTC or PTC formulas, then fit the free variables of the formula to the data... An acquaintance of me was starting to brew beer had all the components, but didn't know how to calibrate it, so I helped him out. Doesn't take long if you set your mind to it and properly prepare.)
Perhaps the difficulty is in the mechanical fabrication such that you thermally isolate from the phloem (which you puncture) and selectively measure temperature in the xylem?
Perhaps standard thermistors are too large. so an alternative would be a tiny diode, whose IV curve is temperature dependent. Added benefit is lower thermal mass, although I was surprised at the response time of the thermistor once calibrated (and filtered to result in step transitions when putting in and out of hot water and icy water).
EDIT: just making sure, what I'm trying to say is, don't be intimidated, and break up the task in smaller tasks, go ask for some help on ##electronics or so when you have sketched an initial plan, and listen to their feedback.
I used to work doing software for fruit sorting machines and we had a NIR Spectrometer to work out the sweetness of fruit, so when I'd go to actual sorting sites, especially in California which have these machines on a massive scale, I'd use it to sort out the biggest sweetest pieces of fruit.
I read a lot of articles of this nature, not necessarily in this specific subject area, and may be guilty of writing some myself. I rarely build anything exactly to its original plans, but I take away bits and pieces, theory, and techniques, and re-combine them for my own purposes.
The value of the whole thing hanging together as a product is that you know the person has worked out the bugs of the individual parts, to the point where the whole thing is testable and reproducible.
This is actually true of the scientific literature as well. People rarely reproduce entire research projects, but often borrow bits and pieces and adapt them for their own use.
I don't know enough to know if this one would work for this kind of usage but cheaper and easily accessible spectrometers would be amazing for harm reduction and drug testing usages.
You could run Quality Control on your amazon chemical purchases if you were willing to put in time to find methods using spectrophotometery in this wavelength range.
Just don't be too critical, or exasperated...
They'll quickly categorise it as rude/disrespectful.
Ignore any & all contribution, refund, & then bump you in an instant.
And they'll do all this in a ham-fisted & egotistical/condescending way.
My experience of their culture, is it doesn't bode well LT.
