3

The closest description for what I'm looking for is an electronic slot machine with a hundred wheels. Each wheel has, say, 8 faces/drawings/symbols and I need to be able to create any of the 8^100 permutations in the most cost effective manner.

The wheels are mounted on a shaft (can be a geared or smooth shaft, whatever needed), and can be spun all at once stopping one at a time or (preferably) just spun and stopped one at a time (so it's not exactly a slot machine).

I've considered many possible solutions, but I'm looking for a method that's as cost-effective as possible so a motor per wheel is definitely out, a solenoid per wheel is less than ideal. I considered mounting a motor onto a belt (a la inkjet printers) moving from one to the other from behind and doing the spinning, which is a very clunky and imprecise solution.

I'm assuming this is a solved problem, but one I'm just not familiar with. Does anyone have any ideas or prior art I can look at for inspiration to solving this?

Trygve Laugstøl
  • 1,410
  • 2
  • 19
  • 28
Mahmoud Al-Qudsi
  • 138
  • 1
  • 8
  • Just so we all understand you project... All 100 wheels should be started spinning at approximately the same time and stopped one at a time? Is the position that it is stopped in random, or does that need to be controlled somehow? Put another way: do you care which symbol it is stopped on, or is any symbol OK so long as it is lined up correctly? –  Feb 16 '12 at 20:43
  • I don't care whether they all spin up at once and stopped one at a time or if I spin up the first wheel, stop it, then spin up the 2nd wheel, stop it, etc. I need to determine what symbol is displayed (not hard if there are gears attached and the microprocessor knows what the previous face was - just calculate and subtract). – Mahmoud Al-Qudsi Feb 16 '12 at 20:45
  • Here is a mechanism reminiscent of a combination lock, with all wheels driven by a single motor: http://tiltedtwister.com/timetwister.html. It would require a lot of twisting on 100 wheels to get an arbitrary position, but its an option. Cost effective, but horribly inefficient in time. – captncraig Feb 16 '12 at 20:50
  • 2
    There is a reason that modern slot machines have moved to video screens. Spinning wheels are complicated, error prone, and expensive. – captncraig Feb 16 '12 at 20:53
  • @CMP very interesting! That's one possible solution I had in mind having seen something that works akin to this before, but I don't know what this technique is called or how it's being implemented under the hood? (And I'm not making a slot machine, just the closest example). – Mahmoud Al-Qudsi Feb 16 '12 at 20:55
  • 2
    Here's a better explanation of a similar mechanism: http://www.youtube.com/watch?v=CZ8WRDVgKrk – captncraig Feb 16 '12 at 21:00
  • @CMP not to mention, easier to rig... – Faken Mar 07 '12 at 01:59

2 Answers2

2

How about using something mechanically similar to a combination lock? In theory, you could get by with just one motor and no other solenoids or other such things. Unfortunately, the time required for an update would be proportional to the square of the number of wheels operated by a single motor (since a full update of 100 wheels would require about 100 rotations clockwise, 99 counterclockwise, 98 clockwise, 97 clockwise, etc.) On the other hand, if you were to split your device into ten sections, each of which could be connected or disconnected from the main drive motor using a solenoid, and allowed any or all to be driven at once, that would cut the update time by a factor of 100, leaving it pretty reasonable.

Addendum:

One could probably control a lot of wheels with a single motor, without the O(N²) spins requirement of the "combination lock" method, by giving each wheel a differently-notched cam. There are many possible approaches one could take to doing so, depending upon what types of cams one could fabricate most readily. If you have a convenient means of cutting flat sawtooth "gears" with various patterns of missing teeth, you could mount all the wheels to a common shaft with enough friction that the wheel would turn with the shaft unless blocked by a cam. Each wheel would have a ratchet mechanism which would allow free travel in one direction, and would block travel in the other direction when a tooth was present (so any attempt to rotate the shaft would cause the wheel to slip on the shaft). Use of two cams per wheel, along with a mechanism to globally engage and disengage a ratchet on one of them could greatly improve efficiency.

supercat
  • 45,939
  • 2
  • 84
  • 143
  • Is that the idea behind the link CMP posted? http://tiltedtwister.com/timetwister.html The "inkjet belt" method is more practical, though. – Mahmoud Al-Qudsi Feb 16 '12 at 21:03
  • Yeah, that's the same idea I had. the "inkjet belt" idea of moving a motor to operate the wheels seems complicated. It would be mechanically simpler to have all wheels on a common shaft, able to turn with a moderate amount of friction, and have an indexing peg sticking out of each wheel. Have the lateral traveler simply block the peg of whatever wheel it's sitting in front of. – supercat Feb 16 '12 at 22:06
2

This is a difficult problem, and I fear that there is not going to be a solution that satisfies your "inexpensive" requirement. I can tell you that IF there is a cheap solution, and I doubt that there is, then that solution would likely be a mechanical solution and not an electrical solution. Since this is an electrical web site, that's beyond our scope. Given that, here is my solution with a heavy dose of electronics...

Your requirements are: A sensor on every wheel to detect position and something to start/stop the wheel from spinning. A requirement that you didn't mention is that the overall complexity needs to be kept simple. A simple system will be much more reliable than a complex system. Also, mechanical wear and tear will be important if you want this thing to last a while.

So... I immediately throw out any idea of using motors, gears, or solenoids. The mechanical design required to hook this all up will be complex and prone to failure.

Instead, make each wheel into a motor. Start by placing the wheel on a smooth shaft. Use ball-bearings on the wheels so they run smoothly. The shaft itself does not rotate. The wheel needs to be made of something non-conductive and non-magnetic, like plastic.

The wheel has mounted on it 4 or 8 permanent magnets. I'm not sure if it is 4 or 8, as I didn't go through the design to that level of detail. I'm guessing 4 if you can align them so the North pole faces the South pole of the next magnet, and the poles are evenly spaced around the wheel.

Mounted near the wheel, on a bar that goes across all wheels, are more permanent magnets. One magnet per wheel. This magnet shouldn't be too strong. The main purpose of this magnet is to force the wheels, when stopped, to align properly so each symbol is perfectly lined up with the next. Also, if the wheels are not perfectly balanced then this magnet keeps the wheel from rotating so the heaviest end is always down.

Also mounted near the wheel, mounted on a PCB, are two electro-magnets per wheel. These form the coils of our motor. Drive electronics, probably an H-Bridge, are also mounted on the PCB.

An IR LED and Photodiode pair, also on a PCB and one per wheel, is used to detect the position of the wheel. Somewhere there is a single black mark on the wheel, and the sensor picks that up to identify "symbol #1".

Each PCB will also have a single microcontroller to drive the "motors" and sense the position. The Microcontroller code is challenging to write, but not impossible. Due to the size, I would guess that each PCB only supports 10 wheels. So for 100 wheels you would need 10 PCB's. This means that some form of communication between PCB's is required, but that is relatively easy.

The electro-magnets will probably need to be hand built, but everything else is just a PCB.

The whole unit would be fairly easy to assemble, and since only the ball-bearings are "moving", everything should be quite reliable.

  • Great idea, but I forsee it getting expensive fast. How fast do you think such a mechanism could be made to rotate? – captncraig Feb 16 '12 at 21:28
  • @CMP Yes, it will be expensive-ish but stuff like this is expensive. Compared to a solution with a standard solenoid/motor per wheel this should be fairly cheap, and much easier to build. I have no idea how fast this could rotate, since it depends on the size/weight of the wheel and the current flowing through the coils, both of which are undefined at the moment. But it should be able to rotate as fast as the designer wants. –  Feb 16 '12 at 21:36
  • How small can an electromagnet be? I didn't mention it, but my wheels are only 25mm in diameter. – Mahmoud Al-Qudsi Feb 17 '12 at 01:06
  • @MahmoudAl-Qudsi The largest you want the electomagnet is approxamately the distance from the center of one symbol to the center of the next. The smallest is up to you, but the smaller it is the less power you can put into the "motor". You will probably have to experiment a bit to figure out what size, current, voltage, and gauge of wire you need. –  Feb 17 '12 at 03:57