Comment by Someone1234
4 years ago
6 bins is a lot. He does give an explanation in the video: They only get trash pickup monthly, and they have several recycling bins.
He switched to Rack and Pinion to overcome toque/wear/contact issues which seems like it should be reliable. I like his mechanical end-stops, but if he is serious about wanting a commercial version, I wonder if there's a way to count "steps" as the gear moves over the rack (e.g. hall effect gear tooth sensor[0]), and just slow/stop based on a predefined end-point? The same sensor would also likely function as a fault detector (e.g. stop engine burnout if the train gets stuck), and you won't have plants/debris potentially setting off the mechanical end-stops (plus reducing parts and points of failure).
[0] https://sensorso.com/rack-gear-count.html?filter=inventory:2...
i know these youtube maker projects aren't really meant to be practical, but the fact that his trash pickup is only monthly just makes this absolutely hilarious to me. all this work to avoid putting the bins out once a month
This system is used every time you need to access the bins (i.e. every other day), since the final location is not as accessible. Trash goes from your kitchen to the bins every other day. Collection happens monthly.
This is not true. He says in the video and the comments that it is used once per month.
Even more than that - he seemingly still needs to take each bin off the train and put it on the curb and then put it back.
Their audience takes out trash at least twice a week. If it wasn't physically impossible for this to work where I live, I would be sourcing parts right now.
With the missed opportunity cost being so high, I too would want to ensure my bins went out automatically no matter what. Hooking the control unit up to the city’s trash schedule would be the next logical step.
Usually the different types of bins are not emptied on the same day, so for paper, plastic, organic and other it's four times every month ;)
Relevant https://xkcd.com/1205/
So five hours invested in the project would break even for him after five years. Probably not worth it from the time saved perspective, but the experience was likely worth it. Plus the youtube video will likely bring in some income, and he'll likely miss less monthly trash pickups.
1 reply →
Yeah but you have to put a lot of them out at once.
I was tasked with designing a "dunker" years ago. It dipped a basket into a basin containing a solvent. It only has to move about 3 feet, but I found that since the endpoints were hard physical constraints, a hall sensor tripping a circuit to either reverse, or stop the motor was far more elegant than trying to code it with stepper motors and way cheaper.
The whole thing was supposed to use a fancy PLC, but when I understood the use case, I ended up making it completely using analog using relays, hall sensors, and timers. It was even able to buzz at you if it got stuck. All relay logic. Mostly because I had a big leftover box of PCBs with 8 relays apiece. The clicking sounded super cool.
I checked in a while ago to see if it was still in use. They got rid of it for a fancy PLC system that cost 1000x as much but is connected to the cloud and stuff.
The nice thing about doing it mechanically is that it doesn't go out of calibration.
You can use an encoder on the motor and count the rotations to slow down.
Limit switches can be used to maintain calibration, in case the encoder misses some pulses over time.