Numeracy Game Sensors

First, do you mean near field communication.

How much money do you want to spend, per sensor?

What will you use for "tokens"?

More detail, please.

Especially how much do you want to spend?

Thank you for the reply,

I was looking at near field communication as it seems to be more widely used these days, but was also wondering if this was the best way of doing things?

I could have a token made out of plastic with the data contained inside it (I have a few contacts that could help me out here)

Once that data is set I just need to have it read from a reader in the grid.

So say we have 3 tokens all with a value of 1, two tokens placed on the first sensor would give a read out of two, and one token on the next sensor would read 1. etc

I need the sensors to add the tokens up but was not sure how this is possible?

Cost wise, I would not want to be spending more than £5 per sensor whilst I'm working it out

Thank you again

Maybe some kind of force sensors to detect the weight of the tokens. You could use fishing weights for the tokens. The sensors would automatically add up the tokens.

https://en.wikipedia.org/wiki/Force-sensing_resistor

I did consider the weight option, but was hoping that all tokens would be the same just different colours.

I was also wondering if touch sensors would be able to do the job.

All rfid readers transmit an energy pulse that triggers the tag to respond with it's unique code number. Placing two tags together would cause both tags to be triggered by the same pulse and the data returned would be intermingled and unreadable. Rfid will not work in your application. I think that your £5/square for sensors is unrealistic at the development stage but may be OK for production quantities if you read capacitance or resistance values embedded within the tokens. Consider a simple capacitive keypad array as a starting point. The individual key positions are bussed to a multiplexer so there only needs to be one processor not one for each square which reduces the cost. Keypads only read 0 or 1 so the electronics downstream of the multiplexer has to be much more sophisticated to distinguish values. Capacitance would automatically sum if the distance between the token and the sensor is uniform so one token dropped onto the one below won't work either. The previous suggestion of plugging resistive tokens into holes in the square is probably the most reliable route. Say each square is has sockets to accommodate up to 10 tokens. You need to multiplex each socket individually so that is 90 multiplexed reads each cycle. Unused sockets read as open circuit and occupied sockets might have values of 9kΩ down to 1kΩ to represent the different numbers. The token resistance values (signal) should be far enough apart so any resistance tolerances plus circuit and socket connection resistances (noise) still allows the correct token value to be interpreted. I considered connecting the resistances in parallel but 9x 1kΩ resistors in parallel with an indeterminate amount of noise would be difficult to distinguish from 10x 1kΩ resistors in parallel with slightly less noise.

As an added sophistication if using resistive tokens, including an extra pin that is either open circuit or short circuited to one leg of the resistor would enable you to add positive and negative numbers. Adding three pins and a rotary switch would make provision for each token to add, subtract, be a multiple, a denominator, a power or a root .

How about a 3 x 3 array of strain sensors that measure the weight of the items placed in each grid cell?

You could use color sensors....once the color is established the property could be added and the chip disposed...sort of like a mechanical bank with many slots, but different....

https://www.adafruit.com/products/1334

Would it be OK to have two sockets on each "square" and two pins on each token?

Each token would then also need two sockets on the top to accept further tokens.

Each value of token has a resistor between the two pins.

Each additional token adds another resistor in parallel.

As long as the resistors have reasonably low tolerance all parallel combinations would have a unique resistance.

A small microcontroller could easily scan round all the squares with a resistance meter and check the measured result against a table of given "sums".

Light filters could work.

If you place transparent disks on top of each other that each filter out a color of light.

As for the micro-controller to read whichever sensor you use, you might look into arduino. some of their cards can read sensor ouputs and send out information to a display. The programming is not impossible to learn on your own or you could have someone build it for you (not me though...still learning).

Drew K

Very interesting challenge. For the disks, you might choose from a variety of LED's including infrared or possibly incorporate a level switch to prevent cheating. On the squares you could use inexpensive photo-conductive cells in series (within the square) and determine the amount of light being emitted based on the total resistance.

You wouldn't have to pulse them, but that might be another option making it difficult to cheat. Electronics are much faster than the human eye so 60 Hz and up are indistinguishable. Light sensors and sources are absolutely the least expensive way to go. The level of sophistication depends on you.

Another popular circuit type that comes to mind is the phase locked loop by which simple communication is very easily accomplished. That could be transmitted by radio wave or light wave as well. I suppose sound is out of the question but it would be very easy to incorporate periodic transmission of serial data by any medium over short distances. The list goes on and on if you know where to look.

You might want to look for a booklet or website on electronic kits. You will get a lot of ideas there as well. Good luck!

What if the sensors aren't under the grid, but rather are on the corners that pick up an x,y coordinate and derive the location of the tokens?

Maybe you could adapt imbedded magnetic bar code technology with the reader(s) located in each of the grid squares then program the bar code on each item with whatever weighted value needed.

If you lay the grid out in and octal or hexadecimal page frame array and connect each sensor to a PLC/CPU input circuit then a standard weighted number system could be used for developing any/all standard math movements/functions and the resultant can be displayed on the HMI screen.

Whatever bar code value is read by the square from the item placed in the square would be input into the CPU program for manipulation and processing of the output.

The only issue I see would be in that only one item could be placed in each square at any one time.

Oh simplify the whole thing with the tokens.

Imbed LASCR's in clear tokens which will pulse the number you wish to add.

Your 3 x 3 grid can then activate them as they're piled on.

Simple, right.....?

Simple Digital logic should do it......even old fashioned TTL logic gates or similar.

You need to expand more as to how many different "items" you are talking about....you would then get better advice and help here.

Interesting one. I would forget a board sensor and go instead for intelligent game chips. Computing power has now become extraordinarily cheap, see here:

http://eu.banggood.com/Wholesale-Warehouse-ATmega328P-Nano-V3-Controller-Board-Compatible-Arduino-wp-Uk-940937.html?currency=GBP

You then can have game chips, each of which is programmed with a unique predetermined value, a chip identity, which it will place on output pins when it receives a request signal. The game chips can be stacked in 2 ways. One is to use stacking headers, but these may be too fragile for little game players. A more robust alternative would be magnetic power connectors.

The main game processor then has a relatively easy task, polling each chip in turn. If it receives a response (because the game chip has been placed on a stack and a connection has been made) it adds in the numerical value associated with that chip identity. The process is updated many times a second to allow for both positioning and removal of the game chips.