This is a place to put info about the NoiseBridge DollarEater donation machine.
big mondo sensor on the larger of the two sensor boards is a visible light phototransistor: http://www.datasheetarchive.com/Indexer/Datasheet-03/DSA0050960.html
from Jtfoote 04:33, 9 December 2010 (UTC)
I got two validators, they live on my hack shelf (highest and southmost on the eastmost shelves). They are "MEI" pn 251015176. Not much about them on the goog; I'm guessing pinouts and repair manuals are trade secret for obvious reasons. At $10 surplus, I'm not expecting them to be anywhere near prime. Worth it for the motors and sensors alone.
I figured out that power comes in on the blue/green two wire header. Seemed to "work" OK with 12V on green and ground on blue. Needs about 1 amp when the motors are running.
I couldn't get it to do anything more than suck in 1/2 way then spit out the bill despite trying several kinds of bills and many combinations of dip switches. I once got the rear motor to work in "coupon mode" (see small toggle switch on the side) but not repeatably.
My best guess is that the bill sensor doesn't understand new bills. It's also likely it needs some enable or handshake signals from whatever it was plugged into, but there were too many variables to reverse engineer. (Seems sensible that that a powered-up acceptor should default to not accepting if the main unit was off, disconnected, or crashed.) I think our best bet is to reverse-engineer the sensors and interface them and the motors to an Arduino.
If we can get some kind of voltage out of the main optical sensor, we could average it to get a rough yes/no go to stop blank paper spoofing. But I'm not going to worry too much about that.