Trash Train: I put my trash cans on rails so they move automatically [video] | Hacker News Reader

Trash Train: I put my trash cans on rails so they move automatically [video]

youtube.com

316 pointsPotatoNinja87 comments

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

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.

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.

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.

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.

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.

If it's was automated correctly (I don't think his is, but it could be ) not missing a pickup day makes it worth it.

Yeah but you have to put a lot of them out at once.

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 ;)

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.

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.

Where can you find supplies for these kinds of projects? Each 6ft. section of rack would cost $81 plus shipping from McMaster-Carr.Never mind the other components, that would add up quickly. I was trying to find some spur gears for a project, but wow those are expensive. I reworked it to use pulleys, but those weren't as cheap as I'd hoped, either.

I don't believe these kind of projects are designed to be economical :). He is building this to drive traffic to his youtube channel, which in turns, pays for the next outlandish project.

They can be made economical if you have a lot of salvaged parts for cheap / free.

(As a boy, I could not have built much of the stuff I've built from 74xx ICs if I had to buy all of them new; instead, I had a bunch of PCBs from a decommissioned computer.)

The motor for the little trolley in the video is a salvaged wiper motor.

Can't answer the supply question, but it seems to me his problem could have been solved by drilling a line of holes into the sheet metal rail. Then have a sprocket stick into those holes.

It works for film in movie projectors (https://en.wikipedia.org/wiki/Film_perforations), so why not for a trash train?

It uses fewer parts, at least. But maybe cutting a bunch of holes in already-installed rail is more work than welding rack onto it.

Movie film doesn’t have to go around corners. Trains have a lot of side to side slop. That’s why cog railways use rack and pinion drive, not sprockets.

Also, track has a not-trivial amount of thermal expansion and needs to be replaceable in-situ.

What you're describing is still a rack and pinion system, but likely to wear much more aggressively. I would be happy to be proven wrong, but my gut instinct is that such a system would wear out very quickly, because due to the involved forces, the sprocket teeth would wear out the holes really quickly.

Film, I would assert, works because the teeth just need to align the film; force transfer just uses the filmstrip itself.

Gear teeth are a reasonably complex design - settled, for the most part, I think, but still non-trivial - and do their job very well. It's a lot of "meshing together heterogeneous surfaces to maximize force transfer" problem solving.

> Film, I would assert, works because the teeth just need to align the film; force transfer just uses the filmstrip itself.

I'm not a projectionist or a mechanical engineer, but I reasoned the opposite way:

* Film sprockets have a difficult job because the film has to be stopped and started 24 times a second[1]. It needs to be held still while light is shining through it, and then (to reduce flicker) it needs to get moved to next frame quickly.

* Film seems soft and easy to damage, but somehow all that seems to work anyway.

* So it's probably OK if you have a stronger material like steel and all you need to do is move at a constant speed.

> Gear teeth are a reasonably complex design

That's a really good point. Here's a page that covers film sprockets: https://www.sprocketschool.org/wiki/Sprockets

And sure enough, it covers the evolution of sprocket design, saying a manufacturer "created the VKF ('very kind to film') sprocket in order to improve presentation and reduce film wear".

So, although you could probably steal ideas from film, there is indeed some engineering to do.

Also, the trash train designer said he wanted smooth, quiet operation. He 3D-printed the pinion gear out of plastic, which achieves that. If he'd done a sprocket and hole thing, my guess is the sprocket teeth need to be strong, so it probably has to be made out of steel. So it'd be harder to make and less quiet.

I guess the main reason film does it this way is because it has to. It's flat.

I still think the film-style approach is probably feasible, but now I'm convinced it has some significant drawbacks.

---

[1] Apparently the sprockets that do this are called intermittent sprockets. Wikipedia has a good explanation and illustration here: https://en.wikipedia.org/wiki/Movie_projector#Film_gate_and_...

Note the blue arrows showing slack between the intermittent sprockets and the other sprockets that move continuously.

Ah, so both of us were in the dark about the others' expertise - I hadn't really thought about the fact that FPS is basically a start-stop frequency in analog film.

> Film seems soft and easy to damage, but somehow all that seems to work anyway.

I still think it likely that force transfer relies more on the filmstrip, but I'm also assuming that the filmstrip stays under tension (the shock absorber in the Wiki link appears to be that slash a belt tensioner for the filmstrip).

(Also, I wasn't aware that Geneva drives had actual real-world use cases. Thanks for the Wiki link!)

> He 3D-printed the pinion gear out of plastic, which achieves that.

Yeah- this part actually surprised me a bit. Plastic gears are great for low-torque applications, but my understanding is that with high-torque/high-force applications, you start running into shear stress material limitations. This is one of those "just try it out and see how it holds up" questions.

I'll also say that when I read your suggestion, I was really confused because the first sprocket that came to mind was a bike sprocket... :)

(FWIW, I'm not a MechE, just a software person who spent enough of their high school years in a shop.)

I think he made his design such that he could reuse tools/skills he already had. From what he tells in the video, he seems quite experienced with welding, whereas he doesn't seem to have a good way to ensure precise dimensions or distances: There are lots of thoughts in the video how to ensure that things still work if they are wobbly.

So maybe he favored a design that let him weld components together vs a design that would require high precision when drilling the holes and making the sprocket.

Of course, when designing a commercial, mass-produced version, you'd approach this differently.

I don't know what racks he uses but one meter of rack doesn't cost very much on Aliexpress.

Hos is the shipping?

You could probably use aluminum stock from eg Lowe's as rails. I built a bunch of solar panel brackets from that last year and remember being surprised by how cheap it was.

McMaster-carr is great but it's almost always crazy overpriced.

I've built scale track out of 25x6mm flat bar steel, plenty strong enough for people to ride on. I wouldn't call the time spent 'fun', but even after buying equipment you'd break even over the mcmaster track very quickly.

With mechanical stuff, if you're not being paid for it, the best approach is to design around cheap, easy to find parts.

You need cheap rack? Find a way to secure bicycle chain to a flat surface and use a sprocket for the drive.

Use mcmaster to figure out what you need, and then many of these drive component type things like rack, gearboxes, bearings, etc. are made to standards. Once you know what you need, say a 50:1 nema 23 gearbox just search that on eBay or ali express. Even if there isn't a listing for exactly what you need, just message a factory/supplier thatbhas something similar posted and they can likely get it to you.

McMaster-Carr is for "I need a very specific thing right now at any cost", not for projects like this. Use it as a guide to discover what is available, but only order from there if every day your project is held up is costing you real money.

If you're near a metal fabrication shop (stairs, metal fences, signs, etc.) then they should be able to source it for you if they don't have it in stock and also easily bend it for you. Total labor hours charged for this would likely be under four hours.

When I take out the trash, I put on a pair of rollerblades and skate it over (I live in an apartment complex and it’s across a parking lot)

It seriously feels like a real life cheat code when I’m doing this

Reminds me of an apartment complex I used to live at that had a really long walk from where my place was to the dumpster. I'd put anything that was significantly awkward to carry in my car for the trip. I knew a few people who moved all their trash there in their car, but I enjoy walking and would prefer not to risk getting smelly food waste spilled in my car.

This is what happens when you ask an engineer to put the bins out.

A software person would have done the next couple of pickups by hand while working on the automation, but then never have to do it again once it has been entered as a cron task.

The hardware is someone else's problem, like this guy

why didn't he just use a cable to pull it? no need for gears, no traction issues, the motor etc can stay stationary. trains have their own power because you can't stretch a cable across the country, but this is so tiny it'd be trivial.

Yeah, early precursors to railroads (cable railways used for mines) did just this :-D , and there were even passenger railways that used cable haulage as the form of motive power!

Here's a really cool example: https://en.wikipedia.org/wiki/London_and_Blackwall_Railway#C...

And there still are in San Francisco.

And they can go around corners: http://www.cablecarmuseum.org/archive/Anat/Anat.html

The rails don't go in a straight line.

Neither do cable cars?

Look at all the complexity needed to get a cable car to follow curves. Top comment was implying that using a cable would eliminate the issues with gears or traction. That’s not true.

Not only the complexity… the power. The San Francisco system uses 95% of its power just overcoming cable friction.

It has to go around a corner, that I think explains why he didn't just use a cable.

It's a lot harder to go around a corner without some kind of guidance for the wheels.

Since it seems like the main concern is moving the bins without messing up the yard, if it were me I'd do the rail and platforms for the trash cans that he did and then just push or pull the train manually. If I wanted to get fancy I'd angle the rail track down slightly so that when bins are full they are easier to move, but that runs the risk of interfering with the landscaping.

Others have mentioned both of these things, but not together so...

1. The track is curved.

2. One of the design goals was to disrupt the surroundings as little as possible.

This second point is really important, and reveals the utter brilliance of this design. He uses the existing driveway as a second rail, but only for part of the travel. Once the train leaves the driveway there is an actual second rail. I have never seen this before. It is sheer genius IMHO.

Couldn't he just route the cable via a pulley (or series of pulleys) to navigate the curves? Admittedly would fall foul of requirement #2.

> Couldn't he just route the cable via a pulley (or series of pulleys) to navigate the curves?

Yes, that is possible. But...

> Admittedly would fall foul of requirement #2.

You got it.

There are all kinds of other problems as well. The train will make a very sharp turn every time it hits a pulley. It's hard to attach something to a cable in such a way that it can reliably go over a pulley. Yada yada yada.

A cable isn't as fun as a fully-functional train.

Despite some of the assertions in this page, no, he does not use this daily. He uses it once per month. No, this is not automated. He has no intention of automating it. Yes, he must manually remove the bins from the train to move them out to the curb.

I did like the track idea, but I cannot help but feel that anything more than a manual push is over-engineered for this use-case.

As a creative project, it was neat, though.

If TechShop were still alive, I'd be tempted to build that. I have to move cans about 150 feet. I miss doing building projects, but don't do enough to have a real shop.

The rack and pinion thing was overkill. The steel wheel was never going to have enough traction, but big rubber wheels probably would have. The vertical guide rail, though, is nicely unobtrusive.

All-wheel drive is an option. Put rubber tires under each trash can, with a worm gear and a flexible drive shaft to the next car. Like a Shay narrow-gauge logging locomotive. Maybe integrate the power car with one of the cargo cars and eliminate the "locomotive".

The big problem here would be compatibility with the local Recology robotic garbage trucks.[1] The cans have to be spaced about 2 feet apart so the big gripper can grab them. (Otherwise, the driver has to get out of the truck, and they hate that.) The robot isn't that good at positioning the cans back exactly where it found them, although it does try. So getting them back on the train would be tough.

[1] https://youtu.be/AK3uPcWCNLo?t=150

I’d have liked to see him succeed with the steel on steel option. I think he just needed multiple drive wheels to distribute the normal force more, and a bin holder over the locomotive. The route seemed rather flat. He might have also been handicapped by not having it entirely on rail, trains have low rolling resistance because of steel on steel but in some spots, he had worse of both worlds, driving on steel and rolling on pavement.

This is really cool! Like something out of Wallace and Gromit. I love projects like this.

I've been wanting to add a remote controlled motor to our trash and recycling bins so that I can drive them up our driveway (which is up a steep hill.) I think this would be a very expensive project though, the motor and batteries must cost quite a lot.

Now, get trash collectors to put bins back on the cars. Maybe it is easier where he is.

Going out to do that yourself is a pretty small chore.

Putting empty bins back on cars is much easier than hauling them full from behind house, so that system is a win already. How to convince trash collectors - maybe put a small sign ("Put back bins on platforms for some fun") and when platforms detect that all bins are put back and empty - play some fanfares and make train go back? A little fun for collectors never hurts.

The speed at which the trash collectors I have seen work means that they would likely not care a bit about putting the cans back in the plates, much less be entertained watching a train take them away.

You still have collectors? In my, not especially large town, we’ve had the robot hydraulic arm things for almost a decade.

I can remember 20 years ago they not only coll cred by hand but for not that much extra per month they’d even walk back and pull them down.

Fully automating the process in some way would at least spare you from your neighbors loading up the recycling bin you've left by the road with dog waste bags.

>Putting empty bins back on cars is much easier

Not according to the collection employees that pick up in my neighborhood. Not once have the bins ever been put back in any manner other than "just throw the damn thing where ever".

Indeed. Here I'm lucky if the bin is no further than my next door neighbour.

I meant ".. easier for owner than hauling full bins from backyard". It depends on locality, in my town it's not a problem, collectors even put them back inside fence for some older people. If you can find some incentive for collectors for putting bins back on platforms, maybe they would put them back.

The way I understand it, this system does not take the bins all the way to the curb where they get picked up. It just takes the bins close to the curb, you still have to move them off the moving platforms manually, and on again after they have been emptied.

Usually, the different recycling bins do not get emptied on the same day, so it would not make sense to put all of them out for pickup. The purpose of this system (I think) is just to make it easier to get the bins from beside the house (where they're usually stored) to the front of the house (from where they can be put out on the curb more easily).

Where I live they use large automatic arm that just puts the bin back down where it was before.

That wouldn't reach the end of his track. Probably he rolls out the bins they mean to collect that day, and puts them back himself.

I have a background building 10x10 feet fabric cutting robots. This was a trip to the past I really enjoyed. I took many of the same steps 20 years ago, robotics is my little forgotten passion.

It's only lacking a little horn so the user can make a small train sound!

> It's only lacking a little horn so the user can make a small train sound!

One of our Japanese or German readers better have a bespoke native word that describes this type of feature/purpose combo or I'm going to be disappointed.

Yeah, it needs a "Makrofon" (very loud, high pressure signal horn) which it then can use as a "Achtungssignal" (signal demanding the attention of nearby people) to announce its departure! :D

With the power of compound words, you can combine it to "Achtungssignalmakrofon" (a Makrofon which is used to give Achtungssignale), add the information that it's small, "Achtungssignalminiaturmakrofon" (a small Achtungssignalmakrofon)! :P

You could even add the trash train information:

"Miniaturmüllzugachtungssignalminiaturmakrofon".

(Small but loud high-pressure horn signal demanding the attention of nearby people on a small trash train)

Worst startup name ever

My neighborhood has wheeled cans that you tip back and push toward the curb. I have been thinking about converting mine into a self-balancing trash can. Even if it had to move very slowly to avoid falling over, that's OK; it could take hours if it needed to.

Getting into the self-balancing state with a standard wheeled trashcan would require a large initial torque, more than enough to tip over the trashcan. There's no "moving slowly" when that's involved... either the control algorithm and motor works correctly, or there's trash all over the driveway.

If you built a completely custom trash can with the wheels in the center and a very small amount of ground clearance when balanced, you could drastically reduce the initial torque required, but at that point, concerns about needing to move slowly would probably be a lot less anyways since this is being fully engineered at that point, not just an attempt at a quick hack.

I neglected to mention the helium balloon.

And if you’re going custom why not just stick 4 wheels on it - no power needed.

Alternatively, some of these wheeled cans have a bar in the front for the hydraulic arms, and you may could have a robot that aligns to that bar(if it exists--mine does) and then lifts it up via that, which seems to use much less force in my experiments[yes, I've been playing around with automated trashcan systems too lol].

I have no trash pickup where I live so I bought a used, homemade utility trailer to haul stuff to the dump. A big plus is that I can fill multiple trash cans (I like the old aluminum ones best, as they are so light) and only need to haul once or twice a month.

However, having a utility trailer has also come in handy in many other ways - from hauling small boats, to moving substantial amounts of dirt, to transporting many bushels of apples that I scrounge from miles around.

I'd love an electric F-150 but for the foreseeable future I'm happy being the guy with a trailer as long as, and wider than, the Subaru hauling it.

I'm in Australia, we call them utes.

I've had a couple utes, but I'm much happier with my long wheelbase VW T5 Transporter (auto, turbo diesel) and a small trailer with removable hurdles and gate.

I'm eyeballing a property on a slight slope and thought about cutting a trench alongside the driveway to act as a guide in which to lower a garbage trolley on a cable. One concern is maintenance and debris getting in the track. Anyone seen/done anything like that?

This is a very cool project. But it's so early in the morning for me that I read the title and took it to mean someone was putting their trash bins on the railroad tracks and was horrified. Sorry... brain isn't all the way turned on yet this morning.

That's a fun little project but it'll never work in the real world. After the collectors empty the bins, they never set them down gently in the original spot (almost always it ends up a few feet away, sometimes knocked over on a side).

Maybe it's different where the author lives.

Has anyone hacked a "smart" garbage can boot that can self-right the bin and find its way around like an autonomous luggage? (Bonus points if it can somehow unlock the bin for you and detect and ward off bears and other animals).

This is super nice. Awesome job

Easy to criticize, I know. But what about wheeled carts with wheels suitable for uneven terrain and a center rail to guide them?

wouldn't make for an interesting youtube video now would it? i mean, if you thought of it...eh?

Cool! I'm a two-bin guy and trash monthly. But I do recycling separately (take it to center)

> and we added some thyme

So now the train runs on... thyme? That's an xkcd reference if I have ever seen one.

I don't really like Ruby, but sounds cool.

This is a different framework

Reply on news.ycombinator.com