DIY Fraction Finder on a budget

You’re the man @blackie. I also had memory getting full on my surface with 8gb of ram. Sampling rate was high as well.

4 Likes

Nice @blackie! I got my sparkfun order!

3 Likes

So did you guys ever get this working? This is to monitor column chromatography? So that when you see something absorbing in the visible region you know you’re collecting pigments that don’t belong?

1 Like

No, this is to monitor fractions during distillation.

I thought this was supposedly done with emission spectroscopy as Arometrix claims.

I wasn’t aware of any of the typical cannabinoids absorbing light in the visible region. The Cannalysis HPLC potency method uses UV-vis wavelengths at 210 and 228 nm.

2 Likes

I still need to pick up some UV LEDs.

I think this thread is relevant.

2 Likes

If anyone is curious how to use a UV flashlight to demystify your process, hit me up! I’m down for site visits too if you are in my area. I’ve used my method to optimize CO2, Ethanol, SPD, TFD, Rosin, Hydrocarbon, Soxhlet, SBD, Chromatography and Crystalization Each color means a very specific thing. I’ve built “fraction finders” from arduinos and 3d printed parts, and I found them to be a little too finicky. The automation will catch up, but it really needs a really resilient Programmable Logic Controller like a Allen-Bradley or something. For right now the best “fraction finder” an experienced human brain :grinning:

6 Likes

You can pick up a micrologix 1200 for a grand or so, it would be much more robust than an ardiono or cherry pie, unfortunately the software licensing is unbelievably expensive.

But some people don’t have that…

And right now I also have a camera hooked up to my raspberry pi, so I could request help from remote eyes, or just keep an eye on it myself if I have to step away.

You could also log everything from vac levels and temps, and whatever sensors you have, along with pictures if you want to. Log are useful when trying to figure out what works and what doesn’t.

In this case the arduino is just acting as a sensor. And a remote computer draws a pretty graph.

There is also the possibility of training AI to detect the different fractions. If a trained person can do it by sight, a trained computer can probably also do it. Google has a USB dongle that can be used with the raspi pi do vision based AI.

2 Likes

This is so freaking badass thank you for sharing! Would love to see the python code @CBNight

2 Likes

Is there a github repo with the python code?

@densone this seems like its up your alley

1 Like

It’s linked in the first post. It’ll be in a zip file.

3 Likes

So is this a fully functional build? Does the programming and all work? What still needs to be accomplished if anything to make this a working thing?

The hardware is functional. It shows a graph of the light spectrum with sensor reading. It’s missing some features that I wanted on it such as a chemical reference overlay, data logging, and easily adjustable sampling times. To integrate these features, you’d have to modify the existing python program or write a new one. It’d be a better option to just use matlab if you have a copy and just write your own code in that.

3 Likes

10,000 dollar reward for anyone who can duplicate a fraction finder and it’s accuracy.
As well as a offer for a full time position at summit research.

Bravo. Good luck :+1::grin::grin::kissing_heart::kissing_heart:

11 Likes

Now we’re talking!

4 Likes

good luck @CBNight. I would love for my brother to be in california working for summit. This would be super baller if the salary was like 150k a year and you only had to work like 30 hours.

Im sure youd rather come back to miami and do your own business. I guess playing with 300k-500k watches isnt fun anymore! hahahaha

2 Likes

I have it pretty good right now. I can’t complain. You definitely got that right! I won’t work for anyone again not unless the deal is too sweet to pass up. @Killa12345

6 Likes

So is SPDking any closer to pulling out his checkbook?

2 Likes