This is an old project I put together from old junk and parts I had lying around. The motivation for this stems from my involvement in our local elementary school's bottle drive. We collect cans and bottles from the local community to raise money for our school. We sort the cans and bottles and then count them into bags. Beer cans are 288 per bag, pop cans 288 per bag, plastic water bottles are 150 per bag and small juice boxes are 360 per bag. Sometimes we can collect almost 10000 pop cans, beer cans, juice boxes and water bottles -- that is a lot of counting!
I've always hated the counting, so when I found a treadmill lying on the side of the road near our mailbox, I knew what I could make. A can counter! (As and aside, my neighbourhood isn't really so trashy, I think someone thought that maybe someone would want a free treadmill).
 |
| Free treadmill! |
The can counter idea is simple - the cans are accelerated along the belt and then shot through a beam. When a can breaks the beam, we know to count one can. The beam is created by an LED shining on a photoresistor; an Arduino takes continuous analogue readings of the resistance. When the resistance changes we assume that a can has broken the beam.
It was all breadboarded and then put in a fancy box. The trick was to build physical guides for the cans to get them to line up in single file and to get the belt driving at a reasonable speed.
It actually works pretty well as you can see in the video. Total budget was about $5 - I had to buy a belt to connect the motor. Everything else I had lying around.
The video shows that the cans can be counted as fast as you feed them. Much faster than by hand. And it is quite reliable -- certainly counting cans manually, people get distracted and miscount.
Some problems with the counter which I couldn't be bothered to resolve.
1. The motor is an old furnace fan motor, maybe from the 60's. I might have been 1/3h.p. once but now is about 1/100 h.p. So it can barely drive the treadmill and tends to stall. I guess I could buy a motor, but that would increase my budget! I think it puts some big emf spikes into the circuit too, which tends to mess up the LCD display sometimes. Oh well, the price was right on the motor (free).
2. If the cans are really, really close together, they might count as one instead of two. But I think some creative coding could probably figure that out. Or maybe a mechanical solution to introduce a gap to separate close cans.
3. It can't count plastic bottles at all. The beam break system doesn't work with plastic; it sees right through it, or counts the bottom, then the label, then the lid -- three bottles where there is only one! Tightening the code to count only variations that last for a certain time might work. Or add another beam from the side and compare values or a different colour that sees plastic better or.... I don't know, it would require more research than I care to do for a free, low budget project.
4. The hopper to feed the bottles onto the belt never really worked. But then again, the budget was low ($5) so I didn't use anything other than scraps of plywood or cardboard for the hopper. Again, a bit of money and time and it could have a pretty effective hopper. I envision a top feed hopper with a powered paddle at the bottom to agitate the cans and feed them onto the belt!
