Hi all, so far I was a very passive user here, reading the forums of this extraordinary community almost every day. I love that it is open source. For my maker radio show Netzbasteln (.de) I have been sourcing all the relevant parts in the last half year, iE using a self soldered shield, L298 driver modules and a welded steel frame (that can be used with plasma cutter later - we got one here already in our shared carpenter/locksmith workshops yard :). Will let you know when its about done and the radio show airs on german national public radio.
But coming to the point: as far as I remember there have been thoughts of how to detect faux / missing movements of the sled. What about adding an old optical computer mouse to the sled. So it would detect movement, at least when there are no cut-lines under the sensor and switch off the motors when the sensor data makes no sense.
I have no experiences in calibration so far, but maybe it could also help here?
Thirdly the mouse buttons could serve as some sort of hardware switches to pause the system and the wheel to adjust speed.
USB mouses should work together with ground control. PS/2 mouses have only two data pins and could even attach to the Mega.
Dont know if all this makes sense and can help something. But I wanted to share the idea.
Greetings from Berlin and thanks again for this excellent community work!
That’s an interesting thought. I would think you would want two optical sensors to help a) avoid loss of data from one sensor hiccuping, b) distinguish between overall sled translation and rotation around the center (I’m assuming triangular kinematics here), and c) a little better precision. On the other hand, if you have worked with a mouse in enough places, you’ll observe that optical mice don’t always pick up well, so I would wonder how well this would work on various materials.
That has been mentioned before,
which is a good idea.
The mouse then progressed to a usb camera
Which is still being developed
The trouble with using them for ‘normal’ robotics is that the camera that the mouse uses has a molded lens with a focal length of just a few millimeters.
But in this application, I’d expect to be able to maintain a decent proximity to the material you’re cutting. Optical mice need a bit of “texture” to the surface beneath them - so cutting smooth plastic or metal might be problematic - but for wood, it should work great.
Using two of them would resolve any rotation issues.
Using a camera could work too - but getting a short enough focal length lens and doing all of that heavy-duty image processing would be painful. The great thing about a mouse is that all of that stuff is already on-board.
At 400 to 800 dpi, even a cheap mouse is vastly more precise than you need - and for not much more $$$ you can get a “gamer” mouse with 1600 dpi.
The protocol used by a USB mouse is very simple - you can easily decode it with an Arduino.
remember, you are aiming for 1/64", so 800 dpi is only 12x your target, it
doesn’t take much error to add up to that.
also remember, you are cutting away the material, so you may have some time when
there is a significant gap from the mouse to the material.
And finally there is the matter of sawdust and what that can do to any attempt
to work with what you are sliding over.
Now, using a mouse to measure how much line/chain has been fed out instead of
using an encoder would be a very interesting option.