Automatic Diferential Lock (Fully Mechanical)

[2GodBDGlory]

This is a design I spent a while developing. Open differentials are obviously poor off-road, and limited-slip differentials are fairly challenging to make work reliably. Manually locking differentials are common and reliable, but I thought it would be interesting to go the extra step and create a differential that automatically senses when a slip situation is occurring, and then locks the differential to stop it.

This is what I came up with:

Yes, it is huge, even when you realize that the whole assembly with the 7x11 frame is not actually part of the lock. Additionally, there is considerable friction, so between these two factors, it is not at all a good candidate for use in MOCs. However, I think it is interesting just as an engineering exercise. The basic working principle is as follows: There is a differential in the center, with the lock, and the two outputs geared to the two wheels. Then, there is a set of two speed-sensing differentials, one on the right, and one on the left. These differentials have an input directly from the drive motor, and an input from one of the half-shafts between the center differential and a wheel. If the half-shaft is spinning at a normal speed, the differential's remaining output rotates, say, clockwise, but if the half-shaft is spinning at double its normal speed, as it would when the wheel is slipping, the differential's remaining output starts to rotate slowly counterclockwise, which then locks the differential. This system is mirrored on each side, to allow it to lock regardless of which wheel is slipping. To continue my previous example, when the output is rotating clockwise, a sort of rotation sensor (using the red 16T gears in the image) prevents the output from doing anything, because if it tried to, it would just stall the drive motor, but when the direction is reversed, the lock is engaged. Unfortunately, although the mechanism was ultimately reliable, it took its time before locking! (I was trying to incorporate it into a model, and to make it lock faster, though still too slowly, I replaced the vertical 8:24 gearing on each side with a 16:16 one), and when reversing, the differential would lock immediately, and then start slipping the 24T clutch gear.

 

As stated earlier, this is NOT a good candidate for inclusion in a model. I was trying, but it was just too inefficient, and provided no real advantage over a manually controlled lock. Still, though, it was a very interesting exercise.

There is a video, too

 

Edited September 25, 2020 by 2GodBDGlory

[Hogwartus]

Very interesting mechanical solution! However, as far as I can see it cannot unlock the differential once locked, am I right?

[JintaiZ]

Simply fantastic!

[2GodBDGlory]

7 hours ago, Hogwartus said:

Very interesting mechanical solution! However, as far as I can see it cannot unlock the differential once locked, am I right?

No, it cannot. My original plan was to hook it up to the steering system such that when you steered, it would unlock the differential, and then reevaluate whether it should be locked or not.

I should have mentioned that in the original post...

[Pnumetac]

Pretty nice design! I'm building something similar to this right now. Here's mine: unlike yours the speed difference between both outputs is compared using a differential. When both run at the same speed the output speed of the differential is 0. A rubber band is connected to that output in a way, so that there needs to be a speed difference big enough to stretch the rubber. The locking mechanism is directly connected to the output and thanks to the new orange rotary clutches it doesn't matter in witch direction the output turns.

What do you think about this?

I found it to be the most reliable and compact design i could think of, considering that you can use a pneumatic cylinder instead of a rubber band in order to change the locking-behavior.

Keep building this amazing stuff, your build inspired me a lot!

 

[MinusAndy]

here’s my take on this idea. I made it much more compact by using 12t 1/2 bevel gears to make the transfer between the two diffs.

this system can be used to make a locking or limited slip system. The red diff outputs to the wheels as normal, by having the inputs to the grey diff going in opposite directions it means that the output from the grey diff crown wheel is stationary when the speeds of the wheels are equal but any difference in speed manifests in the grey crown wheel turning one way or the other and that output can be used to either operate a clutch gear or just have friction applied to it so the diff is semi open.

basically it produces a single measurable output of the difference between wheel speeds which can then be used in loads of ways.

[Jeroen Ottens]

I like it! Both the original post + the build upon each other's ideas thart follows. I see more optimizations possible in the size of the assy by flipping one of Andy's differentials and then moving it one stud sideways. You then can use just 3 16T gear to go from the drive axle to the sensing axle.

[Pnumetac]

20 minutes ago, MinusAndy said:

here’s my take on this idea. I made it much more compact by using 12t 1/2 bevel gears to make the transfer between the two diffs.

this system can be used to make a locking or limited slip system. The red diff outputs to the wheels as normal, by having the inputs to the grey diff going in opposite directions it means that the output from the grey diff crown wheel is stationary when the speeds of the wheels are equal but any difference in speed manifests in the grey crown wheel turning one way or the other and that output can be used to either operate a clutch gear or just have friction applied to it so the diff is semi open.

basically it produces a single measurable output of the difference between wheel speeds which can then be used in loads of ways.

Hi Andy,

That's pretty much it! Mine was less compact, so thanks for this idea! 

I'm going to try it using a pneumatic cylinder instead of a rubber band tomorrow, let's see if that works the way I want it to: By controling the air pressure you could change the locking-point.

[MinusAndy]

That’s a cool idea! I’ll try and do an LDD of the thing I built as I had to strip it for parts.

[2GodBDGlory]

4 hours ago, MinusAndy said:

here’s my take on this idea. I made it much more compact by using 12t 1/2 bevel gears to make the transfer between the two diffs.

this system can be used to make a locking or limited slip system. The red diff outputs to the wheels as normal, by having the inputs to the grey diff going in opposite directions it means that the output from the grey diff crown wheel is stationary when the speeds of the wheels are equal but any difference in speed manifests in the grey crown wheel turning one way or the other and that output can be used to either operate a clutch gear or just have friction applied to it so the diff is semi open.

basically it produces a single measurable output of the difference between wheel speeds which can then be used in loads of ways.

That is a far simpler mechanism than mine was, and is probably much better to use in practice, but it does have some functional differences from mine. My design was set up so that the difference in speed between the two wheels had to exceed a fixed constant in order for anything to happen. The advantage with this is that a vehicle with my system (provided it could somehow overcome the fearsome friction involved!) could run around in circles all day without ever engaging the differential lock, because that constant was never exceeded. Only if one wheel is actually slipping, causing there to be a major difference in speed between the wheels, will the lock engage. My understanding is that your design could eventually engage the lock in routine maneuvering if the vehicle continued to rotate the same direction for too long (I don't know about @Pnumetac's design, though).

I would say that my design is overcomplicated in pursuit of a theoretical ideal, while yours is much more practicable, if a bit less sophisticated.

Good work, though!

10 hours ago, Pnumetac said:

Pretty nice design! I'm building something similar to this right now. Here's mine: unlike yours the speed difference between both outputs is compared using a differential. When both run at the same speed the output speed of the differential is 0. A rubber band is connected to that output in a way, so that there needs to be a speed difference big enough to stretch the rubber. The locking mechanism is directly connected to the output and thanks to the new orange rotary clutches it doesn't matter in witch direction the output turns.

Italics mine!

Your design sounds interesting, particularly the part I italicized! I'd be interested in seeing how this looks, especially if it would provide a more practical way to only engage the lock in a true slip situation.

[By the way, seeing that you are a new user, if you want to include images in a post, you will have to host them on a different site, and then embed the image's link in your post]