I don't need perfection. I just need a bunch of data so I can use machine learning to solve for the optimum set of calibration values. I won't use it all the time because any one pic from the camera will be less than informative as you mentioned. Simpson is much much much higher res than the webcam. However I can use it to calibrate and get a flat first layer and square up my prints when I use 1000's of calibration pics.Through the beauty of statistics crappy transducing becomes good.
I will mount the cam off axis and safe from the heat.
I am going for cheap so I will use a cheap webcam. The bowden tube flexes very gracefully. I was thinking about getting an 8 conductor ribbon and spiraling in just like I do it now. We will see.
I am in the testing phase now. If it does't work then oh well.
I will mount the cam off axis and safe from the heat.
I am going for cheap so I will use a cheap webcam. The bowden tube flexes very gracefully. I was thinking about getting an 8 conductor ribbon and spiraling in just like I do it now. We will see.
I am in the testing phase now. If it does't work then oh well.