Medium-Duty CVT Transmission

[2GodBDGlory]

After seeing quite a bit of activity on the CVT front from Sariel lately, I got inspired to try my hand at some designs again. I started trying to make something similar to Sheepo did years ago, but stalled there. In the meantime, I did some research, and saw something that inspired me. It was in the [GBC] Let's build ball mechanical flowmeters post, but the post was very in-depth, and I quickly got lost. Thus, I am unsure what the original use of this mechanism (posted second from the bottom at that thread) was, but it inspired me. After a bit of experimentation, I came up with this design, which has a large spinning dish meshing with a tire. The tire slides along an axle to "mesh" with the dish at various diameters. I did some fine-tuning to get the friction to the right point, and where it is, with drive from an XL motor, there is no slippage in the fast gears, and some slippage in low gears, but only with lots of resistance on the output. This could be avoided by gearing for less torque on the input, but considerable torque is needed to work the transmission (pointing to low efficiency). One of the nice features of the design is that if the wheel is slid far enough to the low side, it will pass over the center of the disc and create a range of reverse gears. It is also important to note that the gearbox cannot be inverted, that is, you cannot put a motor on the output and treat the former input as the output, because the tire will quickly pull itself off of its wheel if this is done. Shifting of the tire can be done in multiple ways, but I used a large linear actuator. An XL one would take up a lot more space, but would allow for some faster reverse gears. It is also recommended to find an older, worn-down wheel, because worn ones will slide along axles readily, while still transmitting drive.

P.S. If you are trying to build this design, you should know that I used Technic bricks with the studs facing up to raise the disc slightly, increasing friction. This is sadly not visible in my images.

 

[Pattspatt]

On 6/27/2020 at 5:26 PM, 2GodBDGlory said:

some slippage in low gears, but only with lots of resistance on the output.

Very interesting! Looks like pretty good performance for such an unusual method. I wonder, does the slippage in low “gear“ have anything to do with a higher differential speed across the width of the tire? I wonder whether a thinner tire would help. 

[2GodBDGlory]

4 hours ago, Pattspatt said:

Very interesting! Looks like pretty good performance for such an unusual method. I wonder, does the slippage in low “gear“ have anything to do with a higher differential speed across the width of the tire? I wonder whether a thinner tire would help. 

I agree, the ratio spread over the tire is a bit of a problem, but I could not achieve enough friction with the other tires I tried, and there is at least more surface area with this tire, hence more grip. I suppose its width also contributes to the rubber streaks that showed up on the disc. Another possible explanation for the slippage there is simply that there is more torque available to put towards slipping in that gear.

[Void_S]

10 hours ago, 2GodBDGlory said:

I agree, the ratio spread over the tire is a bit of a problem, but I could not achieve enough friction with the other tires I tried, and there is at least more surface area with this tire, hence more grip. I suppose its width also contributes to the rubber streaks that showed up on the disc. Another possible explanation for the slippage there is simply that there is more torque available to put towards slipping in that gear.

I would double that these tires really are the "grippiest" ones in the given size.
Beside this, your CVT is historically known as "friction drive" (https://en.wikipedia.org/wiki/Friction_drive) as the first and simplest CVT type.

[amorti]

Reminds me of this guy:

Can friction be increased with these ?

https://www.bricklink.com/v2/catalog/catalogitem.page?P=98283&idColor=155

Or this, with grilles on it? Obviously you'd have to put something on the back to get an axle hole too.

https://rebrickable.com/parts/11213/plate-round-6-x-6-with-hole/

[2GodBDGlory]

The first suggestion could be worth trying, but I am unsure abho

12 hours ago, amorti said:

Reminds me of this guy:

Can friction be increased with these ?

https://www.bricklink.com/v2/catalog/catalogitem.page?P=98283&idColor=155

Or this, with grilles on it? Obviously you'd have to put something on the back to get an axle hole too.

https://rebrickable.com/parts/11213/plate-round-6-x-6-with-hole/

The first suggestion could be worth trying, but as for the second one, I actually started out with that part (albeit with normal tiles on top), but I found the diameter too small to get enough ratio spread, and it seemed less controllable. I am now starting to think about testing this gearbox in a large car model, so any help I can get will be appreciated! I will test with textured bricks soon. My design is a similar one to that bike setup, but mine adds friction to obtain a continuosly variable ratio spread, while the other design appears to have a number of fixed speeds. It is a similar principle though, and one that is interesting to see used practically.

Thanks for the feedback!

15 hours ago, Void_S said:

I would double that these tires really are the "grippiest" ones in the given size.
Beside this, your CVT is historically known as "friction drive" (https://en.wikipedia.org/wiki/Friction_drive) as the first and simplest CVT type.

Cool to see the history of that design, as I had no idea it had been used before. I guess no ideas that simple are original these days!

Edited June 30, 2020 by 2GodBDGlory