Hello Everyone, I'm new to the forum but have been using Open Builds parts and systems for my last few projects. I am an engineering teacher and have a student who is designing a chess playing machine. The idea is that she will move her piece and then input the move to her program and the program will then respond with a move and move a physical piece on her chess board. I convinced her to use a platform similar to a CNC machine but instead of having the router/spindle, she would have a Z-Axis that will drop down, power an electromagnet to pick up a piece and then bring it to the desired move location and drop. Question is this - we are looking at using the Acro system (20x20) along with a C-beam linear actuator. Will the C-beam linear actuator attach easily to the plate that the Acro uses to attach the laser? Any other options for simple Z-axis which obviously will not need to resist lateral loading? I have seen some people talk about "Pick and Place" systems. Are there any systems setup for an electromagnet? Thanks! Alex Jones
The C-Beam may be a bit heavy for the Acro (acrylic plates) Its meant more as a 2-axes laser/plotter platform.
Thanks Peter - can you give a recommendation about how I can raise and lower the pieces by about 3" or so? Should we custom build a lightweight system with 3D printed plates? I don't think it makes sense to get anything beefier than the Acro system because it's not being subject to lateral loads. What do you think? Thanks! Alex
I would suggest a petal shaped grip lowered by a vertical piston with both being solenoid actuated rather than using an electromagnet. Much more elegant solution and more closely replicates moving them by hand.
You are on the right path.. I recommend you take a look at the video linked below, and his other videos. He is the developer of the Index pick and place (or co developer, not sure who all is involved). It is based on the Acro frame, and a ton of other modifications. OpenPnP – Open Source SMT Pick and Place
Interesting idea Rick, thanks! I did see a video from this person. I will look into it more closely. Thanks!
Another option would be a petal grip, as suggested by Rick, actuated by a RC servo, something like this.
Thanks everyone for your replies! Very helpful. We are going to go with the Acro system and custom make a Z-axis that will match our needs. We have access to 3D printers. Another aspect of this project that is still a bit unknown is how can we have a piece of software control the machine during chess play. So here is the flow events: 1. Machine is in home position awaiting a move. 2. Student makes a move - enters the move into the program. 3. Program processes the move and creates a counter move. Then essentially it spits out Gcode to tell the machine where to fetch the piece that it will move. (Also not sure how to acquire this step, but this might be doable with some programming) 4. Machine receives Gcode from the program and automatically runs to fetch the computers move. Then returns home. Now normally when I run CNC operations, I locate and fix my work piece, adjust the home position, load the Gcode, then run it. So if we transfer this workflow to Chess machine, the program would output the Gcode, but how can we set things up so that it is automatically loaded and run. I almost need a macro of some sort to do this operation. Any ideas? Alex Jones
Standard Acro frame with a servo- or pneumatic-operated petal grip (pneumatic would allow you to adjust operating pressure instead of building a spring-tension overtravel mechanism), mounted high up on pylons with a (long, another piece of 2040 V-slot) custom Z-axis so it can travel above the board and reach down, but also be able to move pieces over the top of other pieces. Similar to the style and operation of very large 5-axis gantry machines. An obvious issue that I see with a non-machine-vision method is that the pieces have to be where the machine is assuming they're going to be, not where they possibly could be. They can't be bumped slightly out of place, and if the machine "takes" one of the operator's pieces, that operator has to place ALL of their pieces, on every move, in the exact center of the square.Even if you solve that by using magnets in the board and pieces for indexing that can survive bumps and vibration, the machine still has to be able to take the pieces, and the petal grip is unlikely to survive trying to push a piece off of its magnet- perhaps two independent heads? grbl has some extra "coolant" IO and I think has some interesting compile time options around custom machine heads. With grbl you can stream g-code via serial, so you wouldn't necessarily need a "run" command so much as a subroutine that acts as a g-code sender, sending one line at a time and waiting for the "ok" return. Wouldn't be very involved, since there would only be maybe 5-6 total moves each turn. The technical details on interfacing are all on the grbl wiki.
Would suggest beveled recesses in the board. The pieces would thus be self centering or at least fall within the target range.
Rob - you raise some really good points. To this point, I have not really considered taking pieces and what that will look like. We definitely need to think about that. As for streaming g-code via serial - thanks for this tip! I will look into this. Rick - Bevels are a great idea - looks like we will be CNC'ing the chess board! Peter - I only know of MATLab, I have never used it yet. It's on my list of things to learn. I will look into this - thank you so much! Alex
