In order to evaluate the stability of the spectrophotometer, 48 experiments were done without delay.
This experiment was repeated twice on
The variation for RGB is less than 0.1%.
The experiment was redone using a delay of 600s between each acquisition. The main variation is probably due to temperature changes. We observe that the changes can now reach up to 0.5% and are depend on the color.
In order to confirm the importance of the temperature, the experiment was repeated with the acquisition of the temperature.
We can clearly see an inverted correlation between the red variation and the temperature.
When repeating the same experiment blank (no sample) / sample (no sample) + blank / sample, etc.
When we insert / remove the cell (vary slightly the cell position), we observe the following values:
We observe that the orientation of the cell is critical. If we force the rotation of the cell so that it is parallel or in bias, we have the followings values:
Turning of 90° the cell also has a big impact depending ,I guess, on the quality of the cell. In our case (with cheap plastic ones) we observe:
It appears that when measuring the difference of absorbance between an empty tank and a tank full of water, the values were aberrants. Indeed the energy was higher with the water than without. One of the problems could be that there is a space inside the tank hole that could lead to refraction. We designed a part to hide this space using 3D printing.
We obtained theses results:
|Empty tank without part||R||G||B|
|Mean of energy||353033,4||317978,2||59185,9|
|Tank H2O without part||R||G||B|
|Mean of energy||472297,8||421643,8||59170,7|
|Empty tank with part||R||G||B|
|Mean of energy||327422,2||322457,5||397948,8|
|Tank H2O with part||R||G||B|
|Mean of energy||414108,5||422623,7||50511,6|
|Test orientation with part||R||G||B|
|Mean of energy||61153||141100,7||177537,6|
|Test orientation without part||R||G||B|
|Mean of energy||91207,2||256636,7||181855,8|
All our measures to test the influence of our part are stored here: