i'd like to make a rfid chess board that can record the position of the pieces in real time for game analysis,?

it's called a blitz board, when the clock is hit it records the position of the pieces. then the full game is converted by software to a game database.

 i have programming experience and electronics experience but it's been a while. trying to figure out where to start.?

sort by: active | newest | oldest

Hi all,

As I am still getting some response on my post 2 years ago, let me shortly expand.

I have finished the RFID board development about 2 years ago, some information is available on http://eitschess.de (in german but should still be easy to understand). The key challanges are the speed of reading a board by RFID (have a look on the RFID spec so I use ASIC chips), the 32 reader chips incl. appropriate coils, assembling the 8x8 antennas, the microcontroller programming (including some direct read of the antenna signal in order to save components). Getting a good spatial resolution for good off/on square detection is key, and requires a good understanding of the actual detection process. I can operate the board for > 100hours out of the 4 AA batteries which reside in the clock. So low power design is very important, if a separate power supply should be avoided. Putting everything in an wooden board at reasonable height requires everything staying well below 1cm in height, important for PCB, and circuit design as well as the jacks. I also have a modified chess clock, which delivers the time stamp after each move. If you are at this stage you are approximately 25% done, as the other 75% is consumed by the appropriate PC program, which reads the positions / moves from the board and generates a error free .pgn file. It is amazing how much mistakes are mady by the players, even if the electronics works perfectly. Currently I am working on a wireless version, as cables have been turned out to be the most likely source for errors. So overall it is a kind of challange if the entire system should work reliably.

For the future I would be very interested in building an "Open-Source/Open Hardware" state-of-the- art but cheap chess clock. If anyone thinks this is a good idea, please let me know (Felix@Eitschess.com).

@danzman: You need add the 32 tags, the coils, the wooden board and pieces, but I may have not fully understood your concept.

danzman3 years ago
exactly the point. They are expensive. If I can build one for less than $150 then I will do it. I have started the project using matrix driver and optorelay chips which activates each antenna of the antennae in the array of 8X8. I am using only one RFID module which is about $30. 64 relay IC at 50 cents each, A reel of 34awg magnet wire about $25. An arduino uno ($25). Some wires and cheap capacitors. I still have $38 to spend in my budget. That will be used to make the PCB. I can see that the major problem is programming. I do now have a problem with C++ or Pascal.
danzman3 years ago
Anyone interested in reviving this thread? I am currently working on a rfid system for chessboard.
fusion (author)  danzman3 years ago
i did a bunch of research on this and decided that it is far too advanced for me since i haven't really touched a soldering iron for anything beyond basic repairs for almost 20 years.

however, there are already boards like this, somewhat expensive though. Novag Citrine is actually somewhat affordable at ~350$USD but the best are DGT boards but those run over 700$USD. but take a look. the Citrine is what i'm saving for.
What about a webcam and image recognition ? 
fusion (author)  steveastrouk7 years ago
webcam wouldn't work because of portability unfortunately. i guess i never stated that the biggest thing is being actually useful, at some point this is going to be taken to a bar for use so i can analyze my games against stronger players.
I've been looking into this but would really like to just place 4 rfid readers into the base and read response and replying signal strength.

I really don't know how to even get started on this could someone point me in the right direction please.

I'm aware it might not work.. but I figure I can play with 4 and add more more if required.

I would like to read the results via usb

Hi all,
I do have exactly what you are talking about. It is a chess board with an antenna array in the board, and RFID tags in each of the pieces. It works great, including software for controlling the board and converting into .pgn.
Just give me a quick note if you are interested Felix.Fuernhammer@Eitschess.de
mjd5 years ago
How about a combination of reed switches and RFID technology. The reed switch would be used to identify squares that actually have pieces on them; then RFID could then be used to identify the pieces. You'd only have to poll the RFID antennas for squares with pieces actually on them, and by doing that you might be able to discriminate between pieces in close proximity to one another based upon signal strength or some other method I'm not informed enough to suggest.

I've long dreamt of building a similar device, with many of the same features you have suggested. I even thought it might be nice to use for across-the-board playing, where it would record both the move and the time of the move, and keep track of the timekeeping as well. It could store these and then send email of the game records, including timestamps for the moves, in this mode.
Taiku6 years ago
Did you ever find a way to this? I am interested in implementing some kind of rfid antenna array and wondering if you ever did that - or anything similar.
frollard7 years ago
RFID would work - but you'd need a coil in each of the 64 squares - not too hard, and then run the RF signal through a matrix multiplexer - a microcontroller pings each square one at a time, and they would be tuned to only work within a few milimeters - such that only the one piece would be read at once...

A version could be made similar to multitouch displays, where each piece has a unique barcode (or bocode) design on the bottom being read by a camera...

Agreed about the reed switches - and only reading all of them once the time button has been pushed.
This would work, but I believe it would be expensive to purchase a 64-signal multiplexor. I suppose you could make your own (likely what DGT has done), but I wouldn't even know where to begin with that :(. Out of curiosity, have you found somewhere to purchase cheap RFID components for this application?

I am currently working on building a DGT-like board as well, although I've decided to use reed switches instead of RFID. As someone mentioned above, there will be challenges to this method that would not exist in RFID, but most of these can be resolved in software.

My goal is to have the board create a .pgn file of a game and send that file to a laptop via USB in real-time. The biggest hurdle I see is how to handle pawn promotion (for example, how will the program know if a pawn has been promoted to a queen, or a rook, or a knight, or a bishop?) and how to know when the game is over and mark it properly in the .pgn file (1-0, 0-1, 1/2-1/2). If anyone has some ideas on this, please share them.
orksecurity7 years ago
Frankly, I'm not sure RFID is the way to go. I'd suggest simply using reed switches under the board and magnets in the pieces, and tracking lift-and-place to keep track of which piece has been moved where. MUCH easier and cheaper to build, and no harder to program.

RFID really is not the right answer for all tasks. Or even most tasks.
fusion (author)  orksecurity7 years ago
to further clarify, http://www.chesshouse.com/ProductDetails.asp?ProductCode=DGT123&utm_source=googleproducts&utm_medium=free_feed&utm_campaign=comparison_shopping_feeds

is what i'm looking to do, however, i only need to record the moves. i don't need interface or wireless (but that would be a great upgrade)

the NOVAG citrine is a great product, but i don't need the built in chess computer. i'd just like to be able to automatically record games as they are being played.

the problem with using reed switches is that some players (myself) sometimes push the pieces, which is habit that is unavoidable. so, say you are making the first move from e2-e4 the comp will recognize e2-e3 and e3-e4, of course this could be resolved in software, but i think that the addition of the clock with turn buttons would solve that problem. just have the comp save the full game position at each clock press and look for changes.

(sorry, i tend to solve problems by explaining them in detail)
kidmosey fusion7 years ago
You could just wait for the piece to be at rest for a second or two before recording a move.

Adjust the timeout if players are used to half-second moves.

(sorry, i tend to solve problems by explaining them in detail)

Oh, please don't apologize. it's quite refreshing and out of the ordinary.

My own view is the problem is one of proximity. How big will the board be? Afaik, you're well within the capture range for rfid on a standard sized board, so it'd be hard or impossible to discriminate between squares at that scale...
Wouldn't it be nice if every one asked questions with too much detail!!!
It would be nice if most people asked questions with some detail. Everyone providing too much detail would go way beyond nice, IMHO.

On an unrelated topic, please note that orksecurity has joined the Reed Switch Nation. Resistance is futile.
I was going to point that out but I knew that that would be redundant.  By the way I justifiably used reed switch in at least one answer today.
fusion (author)  seandogue7 years ago
looking at the DGT board in action, (i lost the youtube link) it must be rfid. as you place pieces on the board it recognizes the piece and color.

biggest reason i'm thinking about this is because i finally got to take a look inside a chronos clock and it's just a panasonic chip and a circuit board plus the battery box and touch button assemblys

one of the things about the DGT board is that it is used for on screen demonstrations as well, but it only holds 500 moves in it's internal memory, without all of the other stuff in there i should be able to kick that up a few notches.
I hope you can make this work but the only way I see to make the rfid work is to have an antenna under each space then detune the antenna so that only the one used is active or at least those adjacent to it.  But then you have to have a descriminator circuit to tell which one is the strongest so you can tell without error which is the block the piece is sitting on.  Most rfid systems that I've seen work over a range of 3' to 50' or so.  You want this one to work over a range of 1-2".

Or maybe all of the spaces are touch sensitive.  Then which ever one gets the piece put down on gets read and the strongest rfid would be the one that is recorded.  THe magnet and reed switch could be used for that.

I'll have to think about this some more.
a pair of back to back reeds in board-square, pieces are encoded by color using a simple bar magnet? (white, N up, black, N down)???

yikes 27 reed switches
fusion (author)  Re-design7 years ago
i was thinking of a redundant system like that. still thinking about it. tweaking the antennas to a short range is simple enough.

that must be what the DGT board does because the board recognizes the pieces by color

maybe the magnet could be used to focus the signal, i'm going to pass this idea along to a few other people.

the squares on a standard board are 2.25 inches so there is a little more leeway, we used to tweak the receivers on our car alarms so you had to be within 1 foot of the car. hmmm... 2mm antenna??

from wikipedia "Milimeter accurate location sensing can be achieved by adding a micrometer wide photodetector and performing a hybrid RF-Optical communication with the RFID tag. This is called Radio Frequency Identity and Geometry (RFIG)."