The two bay Cleansweep crane has been aquired. One bay has a working board in it to run the machine. The right bay has nothing but the DB-15 and DB-25 connectors.
We will construct a board for the right bay that interfaces to a PC. This board will allow the PC to operate all functions. The PC will serve up a web interface to the game.
In addition, for Shmoocon, the left bay will be modified with an IR sensor to detect prize winning. A PC will allow a player to insert a USB key and receive a signed certificate. They will receive credit for playing, and credit for winning. Perhaps we can award free games as well.
Left bay is the USB “ticket” side. It will use a PC running Linux with automount. Parallel port pins will be tied to an IR beam at the price door, input from the coin acceptor (detect credit), and output to the coin acceptor (award credit).
Richard Meserve is going to allow use of a RS232 connected 4 line LCD display. This will be placed in the machine to display information to the user.
We have to construct a system to drive the crane. The system uses 48vdc to run standard motors. The factory board uses 7 relays to accomplish this. 2 relays (one for each direction) for each of 3 motors, plus a final relay to trigger the claw. The claw is a solenoid, and the board allows 3 adjustments of pressure. We should duplicate, fully, the portion of the circuit for triggering the claw.
There are limit switches only for “homing” the trolly, and for raise/lower the claw. This should require 4 input pins.
There is a good chance we can use 3 H-Bridge ICs to drive the motors, followed by a relay or transistor setup for the claw. The H-Bridge ICs require a pulse to control the speed of the motors. We can use a 555 timer IC to generate this pulse, and maintain a setting so that we can set the speed of the motors. This should allow us to tweak it out, and not have to pulse the controllers from Linux via parallel port (really hard to time).
We may wish to add limit switches to X and Y on the right bay, this way the software can calibrate itself as to the time it takes for the trolly to move X and Y. We might actually be able to be uberleet, and use the same inputs for the home switches for end-of-line switches, just letting the software figure it out based upon motor direction.
*Linux based * Input/Output via parallel ports (2) * craned - waits for dropfile containing instructions (from web app) * Upon startup software will test trolly. * Home trolly, all the way left and forward * If no home switch detected, quit after 7 seconds * Move it all the way back and time the travel * Move it all the way right and time the travel. * Store time value, divide this by 100 for percentages for web play * craned wait for dropfile * Dropfile received * Move X + Y values from dropfile * drop claw * trigger claw * raise claw * home position * open claw * close claw * anything detected in output bin? * leave dropfile if anything was fetched
* Player goes to webpage * Player sees webcam images of machine + toys * Player enters in X + Y percent * Dropfile is created * craned executes movement * pictures are taken and stored during move * User is told to wait, depending on # of people in line this could be a while * Return dropfile created * User is shown pictures * User is told if they won
* L298 Dual H-Bridge * http://www.hobbyengineering.com/SectionSM.html * Dual H-Bridge * Rated to 46v and 4A total current * Many availible on eBay cheap * http://www.dprg.org/projects/1998-04a/ (No pulse generator) * LMD18200T H-Bridge * More expensive ($13 on eBay) * H-Bridge * Continuous power requirements up to 55v at 3A with up to 6A surge currents * http://www.national.com/ds/LM/LMD18200.pdf * LMD18245T - Newer version of chip * Can't find cheap?