Thanks for the question! The laser sensor is a Keyence IL-065 (www.keyence.com/products/measure/laser-1d/il/models/il-065/). It's designed for measuring the distance between the sensor and a flat surface. It's repeatable to 0.002mm which makes it very effective for adjusting the orientation of the vacuum gripper to get the LCD perfectly flat but the laser beam it uses to take these measurements is about 1mm wide (photo here: shorturl.at/krALR). We tried finding the edges of the LCD by pulling the LCD out from underneath the laser beam and detecting when the depth measurement increased as the beam went over the edge. Unfortunately, this method was only repeatable to about 0.3mm. I think this is because the sensor reading changes continuously as more and more of the wide beam falls off the edge of the LCD. Another advantage of the limit switches is that they cost next to nothing so we can use three of them to detect the edge in three different places which saves a bit of time compared with having to move the LCD to three different positions under the same sensor. The next thing I would like to try is replacing the limit switches with a machine vision camera. It's not super urgent because the limit switches work well but it would save a little time and would be fun to figure out.
Excellent 👍
Thanks for sharing this! Why use switches instead of the laser to detect edges?
Thanks for the question! The laser sensor is a Keyence IL-065 (www.keyence.com/products/measure/laser-1d/il/models/il-065/). It's designed for measuring the distance between the sensor and a flat surface. It's repeatable to 0.002mm which makes it very effective for adjusting the orientation of the vacuum gripper to get the LCD perfectly flat but the laser beam it uses to take these measurements is about 1mm wide (photo here: shorturl.at/krALR). We tried finding the edges of the LCD by pulling the LCD out from underneath the laser beam and detecting when the depth measurement increased as the beam went over the edge. Unfortunately, this method was only repeatable to about 0.3mm. I think this is because the sensor reading changes continuously as more and more of the wide beam falls off the edge of the LCD.
Another advantage of the limit switches is that they cost next to nothing so we can use three of them to detect the edge in three different places which saves a bit of time compared with having to move the LCD to three different positions under the same sensor.
The next thing I would like to try is replacing the limit switches with a machine vision camera. It's not super urgent because the limit switches work well but it would save a little time and would be fun to figure out.