Improve accuracy

[Peuqui]

Hello there,

while I am up to build one myself and did some recherche about, I stumbled upon this website:
https://github.com/Deutherius/Filament-Width-Comp-Experiments/tree/main?tab=readme-ov-file
This guy made some really nice research about it and used this model:
https://www.printables.com/de/model/57699
which relys on this:
https://www.thingiverse.com/thing:4138933
As I am not an engineer, I think, that your design with a lever is superior to the sliding mechanism of the latter design in regard to friction. But, what is intriguing in my eyes, is that he uses two (2) hall sensors to increase sensitivity/accuracy.
What do you think about integrating two hall sensors into your design, will it work as the distance to the magnet in yours is made by a lever instead of a slider?
Do you think, it will improve accuracy?

Second thought:
The other design has a version 3, which uses pins instead of ball bearings to guide and measure the filament. In my opinion, this could improve accurracy as well, because ball bearings may have some inconsistent diameters. I ordered high precision ball bearings, which are much more expensive than the one sold at aliexpress in regard to not introduce more room for inaccurate measurements. But with non rotating pins, the measurement error isn't apparent.

Besides that, I wouldn't use the PCB, but instead connect the hall sensors directly to the MCU of the printer (BTT SKR 3 EZ in my case), as Klipper supports hall sensors directly as shown in the other design.

I would like read about your thoughts!
Best regards