I am playing with an idea for a large format “board game” for larger groups of people. At the end of the day, players need to place a piece at certain spots (they are in a grid pattern, not hidden or anything). I would like to provide visual feedback of the placement and be able to track all the movements (thus having a gameplay recording). I also need to be able to place multiple pieces on top of each other. The number of board locations is about 90-100.
I came across these magnetic connectors with pogo pins. There are even styles with concentric rings, which would probably be even better:
Thus I am considering having a cylinder, which would have on one end a female and on the other a male connector. The board would have just female connectors. The piece would snap into place, and one can stack them as well.
The thing I am struggling with is how to approach it electrically. A hardcore approach could be having an MCU at every “port” (board location), which would be connected to two buses – one outside for pieces to connect to (I2C) and the other internal (like multiple I2C?) to report to the central MCU.
Or, assuming that every piece has a small MCU and I2C, having a chained 74HC595 multiplexers and adding something like a chip-select line to every port. So it would work like
for location in all_locations:
enable location
scan all possible I2C addresses of pieces
disable location
I dunno, it doesn’t look very elegant. Any ideas? Thanks.
Without understanding the entire “big picture” it’s hard, but I can riff a bit:
How about a row-column matrix, with a 2 circuit connector at each position?
You scan the board, and interrogate each piece by powering it, it responds by drawing pulses of current in a coded manner, telling the main computer what is at that position?
The stacked ones could “know” they were stacked by a feature (magnet?) and then would “hold off” their reply for a few mS to avoid messages colliding with the lower units message?
The 2-wire circuit would link through, up to the stacked unit.
It’s really hard to guarantee that these things will actually make contact, I see this as THE main, non-trivial problem.
Imagine having 100 connectors that ALL had to work, while being easy to plug and unplug…
Tough nut there…
How about a row-column matrix, with a 2 circuit connector at each position?
You scan the board, and interrogate each piece by powering it, it responds by drawing pulses of current in a coded manner, telling the main computer what is at that position?
That is a great idea; I completely forgot about current loops. Nice.
It’s really hard to guarantee that these things will actually make contact, I see this as THE main, non-trivial problem.
Dah, thank you for the feedback, I hand-waved this one off as “I have a magnetic pogo connector” it is going to be all right…
Each board position could have an antenna in it. For a proof of concept you could etch a PCB.
You’d then scan each antenna, switching each into a single RFID receive circuit in turn. You could do this with either an analogue switch, diode switching ( https://www.youtube.com/watch?v=YBNIq_d56sA ) or some transistors.
You’d need an RFID chipset that could deal with more than one tag at a time and hopefully also report RSSI (Received Signal Strength Indicator). RSSI would be useful to help work out how likely a particular token was to be on that square or adjacent squares, or you could use some aspect of the game play to position the board positions strategically. You could probably omit it from your proof-of-concept tho’.
You’d know which piece was on top of another piece by which order you first saw them appear on that square. This might involve some clever design of the board positions and playing pieces so that they stack nicely when placed casually or carelessly.
You might be able to do this with the cheap NTAG215 series of tags that are available in a convenient board-game-counter style package (https://www.aliexpress.com/item/1005006332360160.html). Most of the smarts for reading multiple tags at once is in the RFID chipset. The one in your smartphone probably can’t do it.
