Self-locking differential using pneumatic

[mescalinum]

while I wait for the parts & sets I just bought, I decided to try to implement a self-locking differential (I made a search on the forum, and apart from sariel's "limited slip" differential, I found nothing else).

the key point of this is the slip detector, which makes use of a second differential to detect slip (in fact, rotates proportionally to the main differential's slip).

on top of that, one could implement various concepts (for example putting friction on the slip detector, would make a limited slip differential -note I've put that in italic, because it would be limited by force means, and not by amount of slip- )

what I implemented instead is a pump+actuator(pneumatic) based design, with a rubberband to disengage the lock when there is no more slip (NOTE: the pneumatic hose has to have a little leak in order to allow that).

honestly, I was hoping to make it smaller, by the way this is the very first prototype, and I don't have the small pneumatic cylinder... and I'm posting here to hear some feedback from you, maybe someone could improve over my design (cad file will follow...)

c'mon, comment!

[timslegos]

Wow that is a very clever idea! i bet if you had a mini cylinder you could make it much smaller.

tim

[dhc6twinotter]

Wow...that's pretty slick! I like it!

Would it be possible to build this without the pneumatics?

Good job!

[mescalinum]

Wow...that's pretty slick! I like it!

Would it be possible to build this without the pneumatics?

Good job!

without pneuumatics you can build a differential that doesn't lock, but offers some resistance to slip, i.e. when an axle is in condition of slipping, some resistance (clutch gear?) on the slip detector still ensures that some force goes to the other (non-slipping) axle.

I had to use pneumatics because they introduce a delay which turns to be useful, when exploiting the force of the slip detector to engage the differential lock

I hope that was clear.... that's what I can do at 1.30am

tomorrow maybe I will make a try

for sure, such a design would be simpler than the pneumatic

[captainmib]

This is really nice!

Good job on these mechanics. I think this would be usefull for trucktrials?

Did you build it for that purpose?

To bad I don't have those older differentials, they seem much more usefull.

The new ones only fit to those 20T gears and they can't be used with the red locking mechanism...

Maybe someone can explain that to me?

[technicmad]

OK I'll admit I don't understand what this is for

It looks like when a wheel starts to slip, you disable the differential so it forces both wheel to turn together. But if that's the case, why use the diff in the first place?

[Illuwatar]

To improve handling on smooth surfaces. Locking the diff is a way to get out of situations where traction is lost due to one of the wheels are off the ground. Useful while offroading or in icy conditions. Normal driving is done with the diff enabled.

[mescalinum]

OK I'll admit I don't understand what this is for

It looks like when a wheel starts to slip, you disable the differential so it forces both wheel to turn together. But if that's the case, why use the diff in the first place?

as Illuwatar pointed out, differential is beneficial for normal driving (you'd rather have problems *without* it).

but there are side-effects you can get stuck into, like for example when a wheel begins to slip because has lost contact with terrain.

the differential lock is meant to be temporary, and should disengage automatically just after the axle has stopped slipping.

(the use of pneumatic is meant to introduce some delay and avoid oscillations between the two states)

EDIT:

I produced building instructions for it (quick & dirty, there are a couple of inaccuracies but nothing fatal)

you can get it here (pdf) or in my brickshelf folder as soon as it becomes public

Edited August 23, 2011 by mescalinum

[tripletschiee]

I like your idea and the transformation into real stuff!

But I still don't understand the system "self-locking diff/automated diff-lock". How does the system distinguish between "normal" drive around a corner and slip due to different surface-conditions under the wheels? It cannot! That is also the problem with a torsen-diff.

Example: Driving up a driveway into the 7th floor of a parking-house will cause the diff to work for a pretty long way as the turn is almost "never ending". How does the locking-system know, that this "slip" (for the differential it is slip as both axles will have different turn-rates) is wanted? It does not, so it will lock up. And then you'll have the tension in the drivetrain. That means I can get rid of the differential at all (or accept that I'll have slip in some special situations).

Maybe you cann answer that...?

Regards,

Gerhard

[technicmad]

as Illuwatar pointed out, differential is beneficial for normal driving (you'd rather have problems *without* it).

but there are side-effects you can get stuck into, like for example when a wheel begins to slip because has lost contact with terrain.

the differential lock is meant to be temporary, and should disengage automatically just after the axle has stopped slipping.

(the use of pneumatic is meant to introduce some delay and avoid oscillations between the two states)

EDIT:

I produced building instructions for it (quick & dirty, there are a couple of inaccuracies but nothing fatal)

you can get it here (pdf) or in my brickshelf folder as soon as it becomes public

Very nice instructions, do you have instructions for other MOCs too? :)

[mescalinum]

I like your idea and the transformation into real stuff!

But I still don't understand the system "self-locking diff/automated diff-lock". How does the system distinguish between "normal" drive around a corner and slip due to different surface-conditions under the wheels? It cannot! That is also the problem with a torsen-diff.

Example: Driving up a driveway into the 7th floor of a parking-house will cause the diff to work for a pretty long way as the turn is almost "never ending". How does the locking-system know, that this "slip" (for the differential it is slip as both axles will have different turn-rates) is wanted? It does not, so it will lock up. And then you'll have the tension in the drivetrain. That means I can get rid of the differential at all (or accept that I'll have slip in some special situations).

Maybe you cann answer that...?

Regards,

Gerhard

good point,

in fact, considering those corner cases, you cannot make such a autonomous differential lock.

in my system, the only guarantee that it will work correctly (that is: the treshold that activates the diff-lock) is the amount of slip VS time:

* slipping (due to a real slip condition, or simply a curve) makes the pneumatic pump work, causing the diff-lock to engage at some point

* the rubberband makes opposite work, pushing the pneumatic piston back, causing the diff-lock to disengage

* the pneumatic hose has a small leak (made on purpose) that allows the air to slowly leave the pneumatic circuit

now you can conclude that: if the pump pushes air fast enough (faster than the air leaking due to pressure + rubberband force), the diff-lock engages/stays-engaged;

otherwise, it means that is leaking more air than what is "pumping in", so the diff-lock will at some point disengage/stay disengaged.

is it clear?

so, in the "real" world, in the parking house, the slip detector will slowly pump air in the circuit, and the diff-lock doesn't reach its "activation treshold".

in the case your truck gets stuck in the offroad terrain, where one wheel is lift, the truck driver begins hitting on the accelerator, and it happens quite fast that enough air goes in the circuit, causing the difflock to engage, serving its purpose of getting out of that slippery situation :)

from a technical point of view, it is all about making equilibrium between 2 forces (pump and rubberband), all dumpened by a certain factor, determined by the dimension of the leak (that is probably the crucial part to calibrate if you want perfect functioning)

and of course, this is Technic, so it's hard to get perfectly working stuff

Very nice instructions, do you have instructions for other MOCs too? :)

small MOCs, yes.

check my brichshelf home (I also have something on youtube of which I didn't make cad representation, for various reasons)

it's not exactly MOCs in the common sense.

I just like robotics, mechanisms, and small ideas.

[Alasdair Ryan]

You have some interesting ideas.

[le60head]

Hey, that's a great idea!

I love your design!

I'll go check your other stuff for sure!

[tripletschiee]

@ mescalinum:

Thanks for the explanation. I think your concept is pretty cool, as it triggers the locking mechanism not that fast.

But I am still wondering how the torsen diff is working. I build one with technic Lego, and it did not work when turning. So I could not find out, how that stuff really helps.

Regards,

Gerhard

[Illuwatar]

Maybe little bit OT but also maybe an answer on the previous question:

It seems to be used only in between the front/rear, not between wheels.

Found also this: http://individual.ut...lik/torsen.html

[Nazgarot]

Maybe little bit OT but also maybe an answer on the previous question:

It seems to be used only in between the front/rear, not between wheels.

Found also this: http://individual.ut...lik/torsen.html

It's used as diff between wheels in several cars. Some examples: Hummer H1, old Audi S1 (rally car) and one edition of the old S2 (Tuned by Porsche).

There is several other cars that uses one between wheels, but a lot more use it as a center differential.

-ED-

[mescalinum]

about torsen differential, when used between wheels, does it suffer of the problem pointed out by tripletschiee ?

I have not much understanding on how it works, but it seems the answer is yes...

but for a rally car would be reasonable

[tripletschiee]

I know how a torsen diff works. But it still can not distinguish between turn and slip. So why adding one in between the two axles?

Regards,

Gerhard

[Nazgarot]

I know how a torsen diff works. But it still can not distinguish between turn and slip. So why adding one in between the two axles?

Regards,

Gerhard

If both wheels on a axle have good grip with the ground, the idea of a Torsen diff is that the difference in speed between the wheels are the same compared to the diff housing, but opposite. One wheel is running a little faster, while the other is running a little slower. The result is that the diff doesn't lock. If one wheel slips how ever, the difference is no longer equal, as one goes a lot faster than the diff housing and the other still moves slowly compared to the diff housing. This will result in immediate locking of the diff.

This works extremely well in the few Audi models I've tried that use this solution. Especially on snow. You can have a bit of slip on the central diff do to the power diversion being about 70/30 in limited slip condition. When you lose control the car will gradually gain better grip as the diff lock up. The diff between the rear wheels is in limited slip condition as long as both or non of the wheels are slipping, and if both are slipping the central diff will gradually lock up and return full control to the driver. If one rear wheel slips (typically an non intended slip situation) the rear diff locks immediately.

The result is that if you learn that in turns, when sliding you only need to accelerate to make the car go the way its pointing (both center and rear diff locks up) you get an amazing level of control...

If you stop accelerating and clutch out how ever the car will be sliding like any other as the rear diff wont lock up (no power added to the diff housing => it rotates freely, and will always have both wheels rotating with the same but opposite difference).

This is a simplified explanation, but the effect can clearly bee seen in films of the 80 rally championship group B cars from Audi...

Here is a link to a YouTube video. If you look closely you can see how the cars shoot forward coming out of corners when the driver accelerates, no matter how wide the slide is. But you can also see how the drivers struggle when a wheel slips and the diffs lock up going into corners... You will also see how the diffs never lock up with good grip on asphalt, even in tight corners.

Probably one of the best rally cars on snow and gravel ever, but also very dangerous as you never knew if the diffs would lock up going into the turn, in the middle or during acceleration comming out of the turn as you would wish for...