Ideas For A Sensor

[philba]

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by rstofer:
...
As chips block the levels, I am not certain I will be able to see chips at lower levels, assuming some specific oblique angle. It would be acceptable to move the camera vertically and perhaps even change the tilt angle although, once established, it shouldn't be necessary.<hr></blockquote><p>Unless its a cost issue, two cameras on opposite sides should be able to see all locations.<p> <blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>
It would also be acceptable to emphasize the cell boundaries with black stripes such that analyzing the image is made easier.<hr></blockquote><p>Yeah, that should help, especially if you can ensure contrast from the background. a black-white-black stripe would remove background as a factor.

[Chris Smith]

RFID can be focused or de-tuned. Same goes for the antanae pick up. <p>You can read through a top layers down deeper, from the side, or from the bottom, all through a small array of transmitter/ pick up coils that are focused. <p>Distance is also adjustable. RFID is not just a contact “pick up” it can penetrate certain materials quite deep. <p>Also if you have access to all of the sides of all the items at hand, even if it takes two or more sides to accomplish, a laser scanner [s] will count the pieces, and soft ware can do the rest.<p>And Encode the stepper arm to tell you everything it has already done with RFID as well?<p>[ July 11, 2005: Message edited by: Chris Smith ]</p>

[Newz2000]

This was hinted at before, but quickly abandoned, and I'm not sure why, maybe because of apearance sake.<p>If you took a single strand from a braded wire, and cris-crossed it across the "cell" of a board, then took a second strand and criss-crossed it the oppoisite way, being careful to insulate the intersections, you could create a normal switch that is closed when a bit of metal is placed on the "cell." <p>You could powder-coat or enamel one side of a metal disk and it would give you a nice look that could be picked up with nothing but a magnet.<p>The metal conductors could then be covered with a black line of tape or something similar at the boundry between each cell and then run the length of the board. Something like this:
<p>The wires would be much thinner than depicted in the picture, almost unnoticable, I would guess. The spacing of the wires would be entirely dependent on the size of the playing piece, and you would likely need to cross back and forth only a few times.<p>It would be a painstaking process, but not as time-consuming as restoring a classic car. <p>[ July 11, 2005: Message edited by: Matt Nuzum ]</p>

[philba]

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by Matt Nuzum:
...
The wires would be much thinner than depicted in the picture, almost unnoticable, I would guess. The spacing of the wires would be entirely dependent on the size of the playing piece, and you would likely need to cross back and forth only a few times.<p>It would be a painstaking process, but not as time-consuming as restoring a classic car. <p>[ July 11, 2005: Message edited by: Matt Nuzum ]<hr></blockquote><p>If embedded wires were acceptable (I believe he said there weren't) then I would recommend using a phototransistor. one wire per row (x and y) and scan matrixed. a 1206 PT is pretty small. cheap and easy to interface. <p>I agree - fabrication is going to be an issue in all these wired schemes.

[Newz2000]

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by philba:
If embedded wires were acceptable (I believe he said there weren't) ...<hr></blockquote><p>Ah, didn't catch that.<p>I just wanted to note, I thought of an even better way of doing it than what I listed above. I started thinking that the difference in height of the wires (because some have insulators at the intersections) might make it hard to have a good contact by a flat object laying atop them.<p>Instead, simply making a helix from two side-by-side-but-not-quite-touching wires would be better. Not only would it be simpler to lay down, the wires would not intersect each other, meaning that there is no concern for uneven wire-thickness caused by small insulators.<p>Of course, this may be a moot point if embedded wires are a concern.

[bodgy]

Actually talking about embedded wires (insert going off at a tangent here), the special piezo wire that is used in traffic road speed sensor/traffic counter equipment would do the job.<p>It is hard to get, but from surplus places surprisingly cheap.<p>And continuing on this line of 'intelligent' wire, there is also the 'muscle' wire available from some robotic specialists might do the trick as well.<p>You can tell I've read the book '1001 ways to skin a dead cat' <p>Colin

[rstofer]

I think embedding anything or using any sensor that would limit the visibility through the layers would be a problem.<p>I was thinking about the scanner approach. Perhaps a 2 axis scanner that would go over the top of each layer in turn and sense the colored chips - perhaps in rows of four. If I could just get it to sound like my HP Scanjet that would be cool!

[philba]

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by bodgy:
Actually talking about embedded wires (insert going off at a tangent here), the special piezo wire that is used in traffic road speed sensor/traffic counter equipment would do the job.<p>It is hard to get, but from surplus places surprisingly cheap.
...
<hr></blockquote><p>Interesting, I thought the traffic sensors used a magnetic pickup. Google gave a bunch of interesting links on the piezo wire. ya learn something new... <p>I think those wires need a little more force than a chip would provide. <p>Got a pointer to a surplus house that has this stuff? I dont recall seeing it at All or Goldmine. I'd like to play with some...

[rshayes]

You might consider a magnetic approach. A wire around a square excited with an AC current would act as a transformer coupling to a second wire around the square. A piece of metal, such as copper or brass, would change the degree of coupling. By arranging set of coils along rows and other coils along columns, it should be possible to read out individual squares in an array.<p>The wires can be places in grooves between the cells and then covered with black plastic. This would outline the squares as well as hiding the wires.

[bodgy]

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by philba:
<p>Got a pointer to a surplus house that has this stuff? I dont recall seeing it at All or Goldmine. I'd like to play with some...<hr></blockquote><p>I can't recall the name of the place, it was in the UK.<p>However these people Peratech will sell you 15cm for £5.00 or 1m for £25.00.<p>Not as cheap as the surplus reel that I recall.<p>with the help of another mags board, google search for vibetek, Omal don't have prices, but this was the company who at one time had a special purchase price.<p>[ July 12, 2005: Message edited by: bodgy ]<p>[ July 13, 2005: Message edited by: bodgy ]</p>

[Chris Smith]

Geeks flex hacker muscles at Defcon
http://edition.cnn.com/2005/TECH/08/02/ ... index.html<p>A group of 20-somethings from Southern California climbed onto the hotel roof to show that RFID tags could be read from as far as 69 feet (21 meters). That's important because the tags have been proposed for such things as U.S. passports, and critics have raised fears that kidnappers could use RFID readers to pick traveling U.S. citizens out of a crowd.
RFID companies had said the signals didn't reach more than 20 feet (six meters), said John Hering, one of the founders of Flexilis, the company that conducted the experiment.

[ncstateboy81]

Alright, I need some sort of newbie disclaimer here. But what about using clear tubing and shooting a small stream/puff of air through. Then, use pressure/flow sensors to detect where chips have been placed. Depending on the accuracy some sort of flute like device could be fashioned.

[VIRAND]

SIMPLIFIED MACHINE VISION<p>1.Find a position where a "camera" can see all the chips, so that no chip hides another.<p>2.Camera=array of 64 photocells positioned inside
a box with a lens, each placed where an image of
a chip appears.<p>http://holodeck.virand.com

[MrAl]

Hello there,<p>If you use stephens idea and you make the game pieces steel it will
increase the magnetic coupling even better.<p>There is a simpler approach, however, and lends itself to cheap
mass production...<p>To make each one of the playing boards, you can use a copper
surfaced pc board (widely available). The etch pattern will be
as follows:
1. Where each playing piece can be placed, create a large circle.
Divide the circle in half (so there are two half-circles) by
etching out a very thin line crossing the circle. This forms
two islands of copper in the shape of a circle (square works also).
2. Create two leads, one connecting to each half circle. Bring
these leads out to the side of the board. Each circle will have
their own set of leads.
3. Paint the copper side of the board with enamel, each pad position
outlined as a square.<p>
The Electronics:
The electronics pulses each 'pad' by pulsing one lead of the pad,
and on the other lead of the pad the electronics senses pulse
amplitude. A high amplitude measured means there is a piece
on that pad, a low amplitude means none.<p>Now let's assume you've created one board, connected it to electronics,
and placed a metal playing piece on one of the circles...<p>The electronics pulses all 16 leads of one side of each pad, and when
it pulses one particular pad the output amplitude is much higher than
when it does the other pads. This is the pad with the piece.
If other pieces are now placed on the board, their amplitude goes up
as well.<p>The reason this works is due to "capacitive coupling".
With the thin line between each half circle each pad forms a very small
value capacitor, which doesnt pass much ac current. When the piece
is placed on top of the two half circles, it doesnt short them out
for dc current, but it causes the capacitance of that pad to rise
quite a bit, which causes much more ac current to flow.<p>Note that bringing each lead out separately probably isnt necessary,
as you might be able to arrange the pads in a scan matrix, or else
tie all of one half of each circle together and scan all 16 cells.
This would mean 17 leads for each board.
If a scan matrix would work, it would mean 8 leads per board.<p>If you use a double sided pc board you could connect the electronics
leads to the bottom of each board.<p>In any case, capacitive coupling will be the cheapest i think because
of how well it lends itself to automated construction and it doesnt
require any special sensors. The whole thing comes out to being just
etched pc boards, paint, and maybe four connectors (optional).<p>
Take care,
Al

[Bernius1]

I like all the ideas, but the playing of the game portends the solution.
Ex: the game 'Othello' uses discs, black top / white bottom. Painted magnets (Bk=N , Wh=South ), when detected, will tell you which SIDE of the disc faces up, which may matter. Also, if the board is lucite, SMT LED's & photo-x-istors embedded in the squares is a novel, albeit costly , approach. Or, as a mapping/matrix approach, put one LED in one corner of each level, and detectors (photodiode, etc.) around the perimeter, so that the lines are diagonal, and the logical combination of signals will tell which pieces are present. (Maybe two LED's /level ?) But then it can be reduced to board logic (74SNxx & 40xx !!). For the dispenser, what about a peddlers' coin dispenser? Or those spring-loaded coin holders in the newer cars? Gotta be some in the junk yard. Or the arm that does the placement can have the optical/ magnetic detector on it, so as it passes over the pieces it sees them.
But after all this work, can the arm pass you a filled shot glass ??