Inversed Timing Belt transmission
ฝัง
- เผยแพร่เมื่อ 11 มี.ค. 2019
- I had an idea for a compact gearing mechanism the other day. It is light, compact and has no backlash (that I've noticed). The next step is to see how it behaves under load. The inner pulley is 20 teeth, the outer ring is 120 teeth, the belt is 150 teeth.
I have only found this mentioned in one place, in the conference paper "ePaddle mechanism: Towards the development of a versatile amphibious locomotion mechanism". Please let me know if this is more well-known through another name.
UPDATE 1: Turns out there is an expired patent for a very similar design that seems to eliminate belt detachment at the cost of less contact area: patents.google.com/patent/US5...
UPDATE 2: Hehe, after around 2 years this went somewhat viral. I've uploaded a second video showing the first run of this prototype to my channel. - วิทยาศาสตร์และเทคโนโลยี
Wow, 2 years later this video went viral! Thanks for all of the comments and feedback! I've uploaded a video of the first run of the prototype here: th-cam.com/video/IKbWTLlZe88/w-d-xo.html.
As many of you have guessed, the contraption works really well for light loads but if I force the outer gear to a complete stop (which requires quite a bit of torque), the belt detaches from the outer gear (between two bearings closest to the pulley, NOT in the open area opposite of the pulley). I think a belt with larger teeth would reduce the risk of this happening considerably but teeth clearance might be an issue. Another idea I had back when I was working on this is to turn the design into a large bearing by putting cylindrical rollers all around the back side of the belt instead of using multiple bearings. This would probably reduce belt detachment considerably while keeping the large contact area between the belt and the gears.
The closest prior work I've found is this patent: patents.google.com/patent/US5954611. Their design seems to eliminate belt detachment at the cost of less contact area.
Here is a bit more info for those interested: gustavolsson.com/projects/inversed-timing-belt-idea/
The TH-cam algorithm finally caught up lol
@@rogueroger0
.
Yes, the belt detaches: you have to push it all the path against the outer crown, with many smaller ball bearings, or maybe with a 3D printed path of recirculating small Delrin balls. Can be an cheap and efficient alternative to harmonic drive ! Congratulations !
Yeah lol
Perhaps this is a naive comment, but, unless specific timing is involved, why wouldn’t you use gear to gear drive?
Is it strong? - No
Is it efficient? - No
Is it cool? - YES
How do you know it's not strong? Have you tried?
@@levanphuong5379 Just a joke man. Don't be serious.
@@levanphuong5379 эта херня на палочке и сбоку бантик. Не нужно быть гением, чтоб это понять...
@@user-dg2qj2cn2l no guy, I don't criticize you, because I wanna know how effect it is and think you tried to test it before!
@@levanphuong5379 Ah yeah. I'm not mad about you. Just in case haha sorry for sharp tone.
Point that I said it's not strong is, compared to metal gears belts are definately weak on high loads so thats why.
Looks like better sim wheel force feedback gearing to me
To me it looks like a transmission that can be used in every wheel for a car or bike
Bad day when the belt stretches.
Use chain.
could say the same for 90% of modern engines which use timing belts.
@@gramursowanfaborden5820 ain't modern engines use chains?
@@blackeyedpanda9884 timing belts have been more common than chains for about 30 years, sometimes you still see them on HGVs and such, and some cam timing is still done with gears, but that is very rare. soon it's likely valve timing will be done electronically, but even if it is, it'll likely be a belt driving the cams.
@@blackeyedpanda9884 Belts are cheaper and quieter than chain drives. Many modern engines still do use chains, but most use belts unfortunately. Older engines primarily used chains.
Most heavy duty engines use gears, and tend to be long lived.
Some airplane engines and some older heavy military engines used bevel gears and driveshafts to drive the overhead cams, which I think is cool.
How has this channel avoided my gaze for so long? [sub'd now]
Sim-racing FFB wheel idea?
I would like to see this done with large width timing belts, very clever way of transmitting power! I'd bet its very quiet at speed
Meh timing belts are loud because of the teeth.
@@ipadize I would have to listen for a comparison, but I doubt it could be louder than 3d printed spur gears. There are few gear systems that make little noise, maybe except for those produced using magnetic locking. Rubber is just a natural sound dampener.
@@jonathonhebert7042 yes the timing belt version will be much quieter, if the plastic gear is straight. If the plastic gear is angled then it is a little quieter.
Me too! I moved onto other projects after this prototype but hopefully someone will try it :) Again, see the patent in the video description for an alternate design that eliminates belt detachment (I think) by running the belt straight between the bearings.
@@gustav.olsson two words for you, wet belt. encase it in cooled oil. see how it gets on.
The youtube algorithm decided that it is this week this video gets recommended to all of us. I thank the youtube gods for this.
Mi a baj vele? Szerintem hasznos videó és még nem láttam.
@@6pici6 nincs baj vele, sőt tetszik, hogy ennyi idő után egyszer csak felkapott lett, ezért írtam...
Please don't refer to YT employees as Gods... they already think they are, don't reinforce it
BTW: There was a patent related to this design that has expired: patents.google.com/patent/US5954611
Thanks! That's exactly what I was trying to find. Sheet 6/14 shows a very similar design.
@@gustav.olsson Lot lower contact area
genius and counter intuitive I would never think of this idea
That's one elegant design, well done
To me a more common arrangement that you asked about would just be a planetary gear set. Although two benefits that I can see would be less noise and that the input shaft would not have to be concentric with the center of the outer ring, so this could help with packaging depending on the potential use.
Why dont u create a follow up video on this to measure some parameters like backlash torque etc?
I would like to, but I moved on to other projects after this quick test.. The design needs tweaking to make it robust.
@@gustav.olsson im curious about one thing, if rotated from the middle shaft, wouldn’t the belt just go in a straight line between the bearings instead of following the outer gears curve? Cause I’m thinking the belt will go shortest path once it gets a load on it?
Brilliant ! Well concieved ! Thank you for sharing this build ! Keep it coming !
Did you test this transmission under load? This could be extra solution for electric mountainboards :)
I was going to say euc
Yes I did some tests, but the design needs tweaking in order to avoid the belt detaching under heavy load, unfortunately. I've written some thoughts in replies to other comments!
@@gustav.olsson did you post the files anywhere by chance? I would like to print it out to try it on a ball mill
@@piercer4882 No I didn't.. I might spend some time finding the STL files and publishing them, but since the prototype needs more work I rather upload a better version in the future
@@gustav.olsson That would be very welcome
OMG such a cool design!! You are so smart
this achieves the same thing as a planetary gearset, its cool, but gears are way stronger than belts bros. still cool though
Gears also require precision alignment, weigh more, and make more noise; and they won't yield in a lockup. In an application where these factors matter, a belt might be a better pick.
@@stickyfox ultimately it depends on the situation. While you make some valid points, my argument still stands. Gears are stronger than belts. Planetary gears reside in an automatic transmission, if we replaced it with a belt, that thing would wear out 10 times as fast as those gears. But like i said, perhaps theres a different use case for this
yeah, but car maker still use belts for timing, and high end motorcycles for transmissions
Strength gears win
For vibration dampened applications this would be superior
@@deaththekid3318 'But like i said, perhaps theres a different use case for this' you didn't though. stickyfox did, and that was the point.
Amazing guy, your system is very wonderful!
clever! there must also be a way to vary the valve timing based on RPM, maybe with a centrifugally responsive central shaft or something like that. could one of those first retainer bearings lever in and out effectively shortening or lengthening the belt? that'd be more controllable.
I just thought, maybe you could add a flange to keep the belt held stable in the center, just the right diameter to hide the belt.
If you mean stable in the up/down direction in the video, I already did :) There is a problem where the belt goes to the pulley however, the belt rubs against the flange and if I clamp the cover on too hard it introduces too much friction. I think that problem would be reduced by having more space between the belt and the flange though.
@@gustav.olsson You could include some magnets into the PLA, which will be attracted by the ball bearings (e.g. ceramic bearing) and guide the gear. But, being used with the motor, this will produce eddy current and maybe counteract on the stator coils. I will do some thingering.
Brilliant. I had an idea to do this with a triangle of bearings to make a strain wave gear while getting the advantage of the drive ratio. Your build looks very nice.
Using a belt for the flexible part in a strain wave gear? There are videos of people doing just that here on TH-cam actually!
@@gustav.olsson I'm one of them! I meant that if you made the belt wrap around three bearings, contacting the inside of the outer gear ring, you can make a better strain wave gear than some of the existing belt-based designs because they usually use an arm with bearings on to press the belt against the outer ring gear. With your method, it does not suffer from the lever effects of the arm having to span the full diameter (which makes eg nema17 motors stall). According to this arrangement, the larger the outer ring gear, the better the gearing ratio available. Congratulations!
When you can't decide if you're going for tracks or wheels.
Nice demo model!
Perfect idea which I really like. How did you solved tensioning? Any fancy mechanism inside or are you running the axis of the inner ring gear offset from the axis of the motor through tolerances in mounting holes of those 4 bolds? OR are only 3 bearings used to specify the ring position and the 4th one is tensioning?
This old prototype only had the ability to tension the belt using one of the bearing holders (the one with the hole). When the cover was attached locking the mechanism in place you could move that bearing in a circle, effectively tensioning the belt between the pulley and the first bearing.
Привет этот механизм можно использовать как редуктор, например на станках чпу?
Since you can't push on a belt, the only force is transmitted on the first bearing and the 2-3 belt teeth that are compressed. The only tension is provided by the outer ring compressing the belt on the 2 bearings.
It's like using a gear made out of rubber and expecting to have torque.
This is a very neat alternative to 3D printed planetary gears! Thanks for the inspiration
I believe it could be with some more development! This prototype worked okay, but I think it could be greatly improved if the belt travels between the bearings directly instead of in the outer gear. This would hopefully eliminate belt detachment while reducing belt contact area, but that could be solved with a more sturdy belt..
Awesome 👍How to calculate the ratio for this setup? Using the belt teeth or outer ring teeth?
i love the design
best decision for hubless wheel) great work👍
Cool design!
I have a use for this, definitely
Awesome 👍How to calculate the ratio for this setup?
Nice for a vertical wind generator!
Looks awesome! I'm having a hard time understanding why does the belt stay put along the inner concave surface of the ring gear when it could just easily detach?
Someone else asked the same question, see below!
The teeth keep the belt from ever slipping so that loop is basically a fixed length that will always sit in the teeth due to centrifugal force and the stiffness of the belt.
Seems like the belt at the 9 o clock position could be problematic at some point where the belt comes off of the outer ring gear w/o any support. Would be interesting to see in actual use.
Yes it seems
thats clever, i might try that with teeth machined on outside as well for rack and pinion
Less durable version of planetary gears?
The increase in views is INSANE, it was only a few thousand, then went viral to now 867k -
Damm i like that!! it would make for an excellent remote-controlled excavator slew ring. Do you think it would work well with a 3d printed tpu belt?
Thanks! Not sure about the 3d printed belt, my guess is that it performs better with fiber reinforced belts, but who knows :)
Can the outside teeth be made of a belt glued to the outer ring? This method is not the same with a planetary gear because it can have way higher ratios.
Beautiful
Humanity is going next level on technology.. 3D printing and computers are just helping.
I didn't ask for this to appear. Yet here I am.
Wonder what youtube saw in me.
Oh right, fascination in mechanical design.
I can see optimization I would install, but it looks in essence you just invented a track driven external gear train. Neat.
Maybe consider different belt types, it looks as though the push belt designs as used by CVT designs might offer better transfer of power and stability against the outer casing.
Of course they are all complicated, rather than a nice elegant solution like the one displayed. Still it looks like a very quiet operating belt design, and should be applauded.
Also greetings from a stray corner of the internet.
nice work
very cool! but without a way to tension it, I am not sure it will last!
Where i can use this.
How I should use it ?!
& why I need it ?!
The belt can't be tightend, then the belt will no longer have contact with the wals, except for the points where the bearings is located.
CONGRATULATION VERY CLEVER AND INNOVATOR
Is the hollow-cored bearing the tensioner?
Correct!
So a planetary gear without the gears. Interesting but is it practical ?
What's its use, or advantages...
I can't wrap my head around this. I will need to print something to play with. Does the belt never detach from the outer ring?
Congratulations on this simple, yet seemingly very functional design.
The 2 bearings on the right bare all the forces from the transition of the belt from the pinion to the ring. It also holds the belt securely to the ring gears so that the belt may apply force without jumping teeth. Once the belt passes that first bearing, it no longer transfers force to the ring so it sits in the teeth in a static state until it meets back up with the pinion gear. The other 2 bearings on the left side or just supporting the belt and ring gear in a center rotating axis.
@@MicroAppleStudios thanks!
@@MicroAppleStudios being that the case, wouldn't it be better to select a belt size and roller positions that would keep the entire belt under tension? You would then have 4 effective points of transmission instead of only the two adjacent to the motor gear.
It would still be justified by the simplicity of the mechanism, but you wouldn't be open to an abnormal belt configuration.
@@halfeld That's a great idea, but it wouldnt provide much if any strength and performance benefits do to the weakest link being the limited surface area of the center pinion gear. The center would see the most force due to effective contact area and torque ratio from gear reduction.
@@halfeld take a gyroscope or a spinning top and stop it by squeezing on it's widest point. Do the same thing again but ty to stop it by squeezing the skinniest point. That's the best way I can describe the forces related between the pinion and the ring in this example.
Which timing belt used ?
Example gt2 ,htd3m,5m
How does the belt remain arced, concentrically mated along the outer ring and not pulled under tension around the ball bearings creating a square belt path?
The teeth meshes with the outer gear, forcing the belt to a constant length in each section between the bearings. The belt is somewhat stiff and has matching teeth for each section and this keeps it in position. It could theoretically detach and bend inwards but that would require some external force to happen. In addition, the inward bent state results in more kinks, which might make it a less desirable state to move into...
Great job
Funkar den bra? Sitter bakre remmen fast hela tiden eller lossnar från väggen? Borde det inte vara bättre att låta den pressas mellan lagerna istället för runt hela inner ringen?
Ja, det tycker jag. Det största problemet är att remmen kan lossna mellan lagret som är närmast den drivande pulleyn och lagret som är ett steg längre bort under load. Den har aldrig lossnat i den kurvade delen som saknar lager så jag tror inte det är ett stort problem.
Congratulations. Beautiful alternative proposal for planetary gearboxes.
I have a question. For some time I have been looking for a manufacturer that sells belts for high rotation speeds (> = 60000 RPM) and powers equal to or greater than 10 HP. Do you know any manufacturer? Thanks.
I think I can help, if U sure it's needed
This guy went and made a planetary belt
Does this function in a similar way that a planetary gearbox does? Would the output ring rotate if the middle gear was held stationary and the carrier with pulleys was moved?
Yes, it would.
But I think the efficiency would be worse.
@@myxnox why would it be worse?
If you close the gear with the cover, the rigidity will be sufficient.
really cool man!
Nice idea!!!
What a cool idea!!!!
Что помешает ремню выскочить под нагрузкой? он же не прижат толком к внешней обойме слева, да и между роликами есть слабина. .
How to calculate the ratio for this setup?
Can you link to the timing belt you used?
using these and stacking them with different sized outer gears could be turned into a working gearbox
It is for me very difficult to understand how the engineering do this system to should work its amazing
its nice but, isnt there a lot of moving parts?
Hallo reduction ratio?
very interesting
I wonder why the belt doesn't fall out where there are no bearings? (far left) Why does it follow the curve instead of going straight across mmm
I have the same question.
There looks to be a stationary ring behind the ring gear
It's because the overall length of the belt cannot change. The leftmost bearings are fixing those points on the belt to the outer ring, keeping the belt between them the same length at all times. It's too much belt to be a straight line, and it sure as heck won't bend to the right
Correct. But unfortunately the belt is not very secure in that state. A better solution would be to make it more like a large bearing and have many cylindrical rollers all along the backside of the belt to force it in place, but that would complicate things. The patent I mention in the video description make the belt go straight between the bearings eliminating belt detachment. This is another solution that might work well and doesn't complicate things too much.
@@gustav.olsson yeah i thought that was weird, cause no significant tension or force keeping it in the teeth . thanks for clearing things up, nice device!
How do you control belt tension?
This old prototype only had the ability to tension the belt using one of the bearing holders (the one with the hole). When the cover was attached locking the mechanism in place you could move that bearing in a circle, effectively tensioning the belt between the pulley and the first bearing. That would not tension the belt between the other bearings though...
Can it be considered as planetary reductor with belt instead of gears?
I view it as a normal belt and pulley system with two pulleys but where the larger pulley is turned inside-out
Maybe it’s just my OCD but it feels like there is a bearing missing?!
good idea
nice, do u would release the stls?
I can check if I still have them, the source design file was modified for my next (not completed) prototype, so that would not work at the moment I'm afraid...
Amazing!
This could easily be stacked with different ratios along a shaft, with clutches to change which one transmits power. boom transmission
That is what the planetary gears do in automatic transmissions and they are far tougher than a rubber belt.
Hmmm with some fine tuning, I bet this could achieve zero backlash?
super cool, +1 subscriber; gear ratio ?
Thanks, the outer gear has 120 teeth and the pulley has 20 teeth, which makes it a 6:1 reduction
I'm thinking that would make a great 4th axis with no backlash for a cnc milling machine
If there is enough space inside the machine I would go with a regular belt-pulley drive though, it's more stable as the belt is pulled into the gears instead of the other way around as in this design. I think this is only really useful for compact designs!
harmonic drive... theyve been around for 60 years
- What is it for ?
- Nothing.
Not efficient in my opinion, it's a great idea, but with that amount of theeth, the rubber is going to get chewed up pretty soon. Instead, you can try to outsource a bicycle chain, couple of small and one big gear socket. Maybe if you cut the outside ring, and stick it inside the plastic you've made and swap the parasite and tensioning pulleys with geared ones you will get an actual heavy duty inverted transmission. However! This will introduce the need of some kind of lubricant, soooo maybe, you can do a closed see-through trans, which by mechanization works just like on a real car would? Maybe add a see-through plastic box, which can be dismantled for servicing and add a little funnel on the side to fill with oil. Plus a drain bolt for oil change?
Do you mind to share the source / stl files for your design?
The model was made in Fusion 360 but the tricky part was finding the measurements analytically. I ended up making a program that found them numerically. I might polish and release it one day but I don't have much time currently. The next step is to create a new prototype with a heavy duty belt. I could upload the STL from the video, would that be helpful to you?
The STL would be a nice start. Looking forward to try it out...
You could use a double sided teeth timing belt and use gearwheels on the inside too.
This would increase your torque I think.
Not sure how that would work? The bearings are just free-wheeling and the base plate is locked to the motor driving the pulley..
@@gustav.olsson you're right, the gearwheels would need to be driven by the center gear as well to have any driving effect on the belt. If the four gearwheels were being rotated around the center as input, driving the belt and thereby driving the center shaft, then it would have effect.
How is the belt sticking to the gear? Why isn't is going tangential from bearing to bearing?
The belt is longer than the tangential distance and since the belt is stiff it expands into the outer gear
@@gustav.olsson Oh right, thank you!
With resistance and a decent rpm it would slip on the bearings
Very impressive - certainly thinking out of the box. Brilliant! 🇬🇧
I’m not really good at this. Can someone explain what the benefit of this is?
It's a compact way to increase the torque of a motor while still keeping the motor centered and reducing what's known as backlash (compared to normal gears). It's only a prototype and needs more work. All mechanical gear systems have different trade-offs and this one is good at some things while bad at others.
did you test the system under load ?
Yes, it worked okay under moderate loads but if I forced it to a halt the belt would detach from the outer gear at one point between 2 bearings. I think a better design is to let the belt go straight between the bearings like in the patent I mention in the video description. This would hopefully eliminate belt detachment and the reduced contact area could be remedied by using a sturdier belt..
I don't understand how this is a transmission? How does the speed vary? Can someone explain?
It is the ratio between the number of teeth on the center gear to the number of teeth on the ring gear. The number of teeth on the belt is immaterial. In itself, it is not a "transmission" in the common sense, but could be part of one as per Zed a.
English is not my native language, what would you call this type of gearing?
chain & sprockets instead of a belt?
There is an illustration of this (but where the chain is not against the outer gear) in the patent I link to in the video description
You are not shooting the bullet , you are catching it!!!!!!!!!!!
how does the belt stay on tho?
It detaches from the outer gear at one position between 2 bearings under heavy load.. See the patent I mention in the description for a version that seems to eliminate this!
@@gustav.olsson aahhh okay, thanks for the fast response amazing stuff man
Please can i have stl file ?
It's a cool bit of original thinking, but these belts are designed to be used under tension to keep the teeth pulled into the depressions in the wheels, in this case tension tries to pull the teeth out of engagement with the outer ring.
You're correct! If size is not a factor, a normal belt and pulley system is to be preferred. This design might be useful when size does matter though!
I think it is possible to improve the design and eliminate the drawback you describe however. Since the force that is pulling the teeth out is small compared to the force rotating the outer gear, I think a set of small cylinders/pins along the back side of the belt might be enough to avoid belt detachment even at high torques.
lower efficiency, heavier, belt tension problems
Nice