Few months ago I made a post about a liquid container I had built. It is a one liter steel can with two windows. I can use it for measuring transmission spectra of all kind of liquids. However it didn’t turn out as useful as I initially thought. Changing a sample liquid is too troublesome. This became apparent when we used the container for the first time. First filling it up with reference water. Then emptying it. Then filling it with first sample and then emptying it again and so on… All the time there were small changes in the beam alignment and splashes all over the working area. A drain valve would have been more than useful.
The lab manager I was working with proposed that the measurement should be made through a test tube? Yeah sure, I thought. Like that would ever work well. Too short a path length and bi-convex ray properties. No way!
But the idea stack in my mind and soon I was convinced that it could be a great add and first drafts were in process. And here it is, set up for measuring transmission spectrum through a test tube.
Key thing is to limit the beam to narrow center part of the tube. If part of the beam is allowed to travel through the edges of the tube then the effective path length decreases quite rapidly. If the beam is allowed to travel through the whole crossection the effective pathlength is close to 67% of that in the center.
Second improvement is the LED. Now I use one that is emitting “warm” white i.e. one that has more red wavelengths. I read an article about similar project (link) and their LED source inspired me to change into similar type of LED. The exact type is Marl NSPL500DS.
Third improvement will be adding second reference voltage to PICs ADC. This way I can eliminate the dc-offset consuming my valuable 1023 bits. Resolution should increase roughly 20%.
This change will be added to the schema schema soon (and corrections of obvious errors). So please check that also in coming weeks.
Experimental Photonics 