Automatic Torque-Sensing 4-speed Transmission

[Sariel]

Here's my previous transmission design equipped with a torque-sensing shifter. The result is that the transmission starts at 4th speed and shifts down lower the more resistance the output meets. It's not perfect, the shifting is jerky because of a "gap" when both transmission rings are disengaged, but it works. I'm hoping somebody can improve on it and make it shift more smoothly.

And the original transmission with instructions:

 

[Rudivdk]

That's brilliant. When you do the testdrive I don't think the jerkyness is really problematic, depending on the type of vehicle you want to run this in. It would be acceptable for a small tractorlike vehicle imho, maybe some tank even. It would probably look a bit out of place on a supercar type build though...

[hjxbf]

I would have thought it might help to increase the force around the stepper piece; it seems that the gearbox lingers undecidedly between 90 degree intervals when it struggles to shift.

Don't know if it would prohibit shifting altogether though, would probably need some testing.

[agrof]

Fantastic solution, there are some ideas out there, but I think this is the first actual one, which is robust enough, has reasonable gear ratios - and proven by test drive. A real milestone in my eyes. 

[Bluehose]

Maybe you can fill the gap with some diff and white clutch

[MangaNOID]

This is great! I love what diffs are capable of. They still baffle me of course...

Anyway I like how it is not so smooth in operation. It lets you know something mechanical is happening which is pure Technic 

[amorti]

The skipping seems to be because for a moment on each shift, the box is between each gear. This means the resistance "sensor" has nothing to sense.

Have I got it right so far?

Does it help to put two bands where you have them, and another on the other side? Might damp some of the shifter's movement... maybe.

Apart from that I'm thinking of ways to get the system's impulse to shift, to stick. Something perhaps like a 4x4 round plate with system triangles making 4x traps for the shifter? IDK, seems complicated but might be what it needs to give a detent. 

Edited May 27, 2020 by amorti

[steph77]

Hi Sariel,

I'am stunned by this proof of concept working.

I have no idea of how it can be improved, as the issue you get is linked to gearbox shifting gears themself.

However, even not 100% reliable this system seems correct to be used in mocs especially as it didn't stress or dommage parts.

Well done! 

 

 

[suffocation]

This is terrific! Thanks for sharing

[aol000xw]

This is so great, the kind of thing that younger me expected to find in technic sets.

Older and cynical me lost any hope long ago.

Anyways, thanks for sharing, great job, as always.

 

 

 

 

[Pas_de_nice]

Saw it on FB earlier. I think you can add an up-geared flywheel on the torque sensor in order to add inertia to the shifting mechanism, therefore hopefully reducing the jerkiness of the gear changing...

[Ivorrr]

19 hours ago, Rudivdk said:

depending on the type of vehicle you want to run this in. It would be acceptable for a small tractorlike vehicle imho, maybe some tank even. It would probably look a bit out of place on a supercar type build though...

I think that it can actually have quite a number of applications! Think about crawlers for example. In addition, any modern car has an option to have an automatic transmission. Supercars as well as LEGO supercars are usually large and heavy so can't go too fast so it also seems an option if you want to use it for a motorised MOC.

Thanks for sharing Paweł.

Brilliant!

[Didumos69]

This is not really new, but it is the first application in a vehicle that I come across. @Charbel did something like this and I've seen something similar with a shock absorber.

The general issue I have with these kind of gearboxes is this: Imagine you are driving a steady pace in 3rd gear. Now you open the throttle wide, with the idea to accelerate quickly. Torque will increase, which will cause RPM to increase, which will eventually cause the automatic gearbox to shift to 4th gear. That's what you would expect. However, when you open the throttle wide with this gearbox, the increased torque will make the gearbox shift down to 2nd gear .

Edited May 27, 2020 by Didumos69

[Lipko]

Maybe simply doubling the whole thing with a slight offset of the switching torques? Maybe it would only smooth in one direction...

[robert]

56 minutes ago, Didumos69 said:

This is not really new, but it is the first application in a vehicle that I come across. @Charbel did something like this and I've seen something similar with a shock absorber.

The general issue I have with these kind of gearboxes is this: Imagine you are driving a steady pace in 3rd gear. Now you open the throttle wide, with the idea to accelerate quickly. Torque will increase, which will cause RPM to increase, which will eventually cause the automatic gearbox to shift to 4th gear. That's what you would expect. However, when you open the throttle wide with this gearbox, the increased torque will make the gearbox shift down to 2nd gear .

The downshifting is called "kickdown", isn't it? 

[MangaNOID]

5 hours ago, robert said:

The downshifting is called "kickdown", isn't it? 

yes its pretty much how all old automatic gearboxes worked. They shifted down when acceleration was applied depending on the revs and how hard the acceleration input was, then the gears would change up as required as that 'load' got 'easier'

In a sense Sariels thingy here is a great representation of that outcome as the drive input torque remains the same on this of course. 

[dhc6twinotter]

Very, very cool!  Differentials are my favorite Lego part, and I love seeing the different applications they are used in.  Super cool.

[Didumos69]

10 hours ago, robert said:

The downshifting is called "kickdown", isn't it? 

 

4 hours ago, MangaNOID said:

yes its pretty much how all old automatic gearboxes worked. They shifted down when acceleration was applied depending on the revs and how hard the acceleration input was, then the gears would change up as required as that 'load' got 'easier'

In a sense Sariels thingy here is a great representation of that outcome as the drive input torque remains the same on this of course. 

Didn't know about the kickdown. Makes sense. Thanks!

[Charbel]

11 hours ago, Didumos69 said:

This is not really new, but it is the first application in a vehicle that I come across. @Charbel did something like this and I've seen something similar with a shock absorber.

The general issue I have with these kind of gearboxes is this: Imagine you are driving a steady pace in 3rd gear. Now you open the throttle wide, with the idea to accelerate quickly. Torque will increase, which will cause RPM to increase, which will eventually cause the automatic gearbox to shift to 4th gear. That's what you would expect. However, when you open the throttle wide with this gearbox, the increased torque will make the gearbox shift down to 2nd gear .

Well I think Sariel was doing this way before me : http://sariel.pl/2009/04/2-speed-automatic-gearbox/

Concerning this particular design, I like it but I'm a bit concerned about that rubber band, it looks to be about to break each time it reaches 4th. Maybe using a gear rack and extending the rubber linearly could help.

[2GodBDGlory]

That gearbox looks pretty effective.

I recall trying to build a similar one, but I kept having the spider gears in the differential skip under torque, and I had a much less sophisticated tensioner.

Perhaps adding some kind of auto-clutch like in Sheepo's Land Rover Defender could help smooth out the shifts?

I think that I will post my planetary 2-speed auto here soon, while our interest in autos is peaked.

14 hours ago, Charbel said:

Well I think Sariel was doing this way before me : http://sariel.pl/2009/04/2-speed-automatic-gearbox/

Concerning this particular design, I like it but I'm a bit concerned about that rubber band, it looks to be about to break each time it reaches 4th. Maybe using a gear rack and extending the rubber linearly could help.

Another option would be to add some gearing between the shifter and the rubber-band-beam, to decrease the travel of the rubber band. This would, however, require more powerful (or just more) rubber bands.

[Aerolight]

A few year ago I made a similar, albeit only 2 speed gearbox. And encountered the same problem, after a few failed attempts my solution was to temporarily in gauge both speeds at once when shifting. As the force increased the lower speed would engage and only when it was fully engaged would the high speed disengage, it still took some finagling but eventually it would shift smooth every time (the main aid was slack in the spring that meant the gearbox required more force to shift up than down. In order to allow both gears to be engaged it needed a clutch ring per gear instead of one engaging with both, but if the ordering of this gearbox had speeds 1 and 3 on one clutch and 2 and 4 on the other it would not need extra clutches (like a duel clutch (which I guess this technically is)). Indexing the shifting would be annoying... but cool to watch neutral brake duel clutch shifting.  

[Ankoku]

I like the fact that it basically does what you always wanted, if you are basically doing a LEGO vehicle for funnsies. e.g. whenever i drive a LEGO vehicle around using the remote control, you always wish it could do exactly what this does, but without all the NXT/EV3 stuff which is usually required to perform such a task. The pause doesn't really bother me as it is trying to sort itself out. Gives you a chance to appreciate what it is doing.

Bit like 9398, but where you can have both a bit of speed and strength as required.

Great work!

[aeh5040]

I've been similarly frustrated with these for a long time.  The earliest example I'm aware of is this one from 2001.

http://www.last-outpost.com/~malakai/lego/

[Gray Gear]

@Sariel did you try something like this? I would try it myself, but I have other projects to work on, and all my diffs are used up 

By offsetting the wave selectors one or two gear teeth you can reduce the space between the gear changes. It was just an Idea, and it works nice. But i dont know if it would work well in your application. Ill just leave it here and take my leave