[MOC] Super-Diff

[Junpei]

I did this a while ago and it's been sitting in my parts bin ever since, until DrJB brought it up and I had remembered to post it.

It's a modification to the oldest style diff; the 1/2 stud tubes at the ends always bothered me, so why not just cut them off?

It ended up like this:

Some huge advantages over the new bevel style diff:

• Higher reduction ratio while having the same diameter
  
• Fits 4 bevel gears (EDIT! I guess it's fine without it.. )
  
• Has limited slip- bevel gears only turn when a certain torque is exceeded
  

I would love to see these installed in a 9398..

TLH

Edited February 17, 2014 by TwentyLeggedHen

[Zerobricks]

Is the 4th gear braced at all, or is just for fun there? I dont see any real advatages comapred to the new one except the higher gear ration.

[Junpei]

Nope, the 4 gear isn't braced; however, I think it does still spread the torque over 4 bevel gears as opposed to three.

TLH

[Lakop]

I think it's great if we can re-use and mix and match the older parts with the new. I don't like the idea of my studded and older parts being redundant.

H

[Mestari]

Wouldn't the fourth gear just fall out while driving?

[jesse66058]

This is perfect! I also was thinking about a new, stronger diff made of original lego parts, but you beat me way better!

@ Mestari

No, it's stuck. You can try it your own and see what happens. Althought it can be quite difficult to install.

Edit: Took a closer look, but are those axles cut off? Looks from here like 3.5 studs long. Maybe you can install a lego 4l bar to stiff things up if you use original axles.

Edited February 17, 2014 by jesse66058

[Hrafn]

Wouldn't the fourth gear just fall out while driving?

Nope. I don't have the oldest style diffs so I can't try this, but I always put 4 bevel gears in my 4L diffs to strengthen them. In order for the 4th gear to fall out, the two gears it is meshed with would have to rotate in the same direction, but they can't because they're both meshed with another gear.

The advantage to having 4 gears instead of 3 is that it balances the forces on the two gears that are connected to the half-axles.

[Blakbird]

The advantage to having 4 gears instead of 3 is that it balances the forces on the two gears that are connected to the half-axles.

This topic has come up before. Adding the 4th gear to the diff makes no structural difference whatsoever. In order for it to strengthen the differential, the 4th gear needs to be guided on a pin in the differential housing (like the 3rd gear). The 4th gear carries zero load.

[Junpei]

Oops, sorry about that, Blakbird!

@jesse: nope, the axles aren't cut; one is 3L and one is 4L. You can use two 3L axles as well, but if there is a fourth gear, which is now proved useless, there is no place to put that tab in the center to stop the 3L axles from sliding inwards.

TLH

[Hrafn]

This topic has come up before. Adding the 4th gear to the diff makes no structural difference whatsoever. In order for it to strengthen the differential, the 4th gear needs to be guided on a pin in the differential housing (like the 3rd gear). The 4th gear carries zero load.

Interesting. Do you have a link to the older discussion, or force diagrams? Thanks!

[Hrafn]

• Has limited slip- bevel gears only turn when a certain torque is exceeded
  

Could you explain how that works?

[aol000xw]

At least with this design the gears can't simply slide out of the axles like it can easily happen with the new diff.

[Junpei]

Could you explain how that works?

That was actually a bit unintentional; however, it actually works very well! The modification was very simple - I simply sanded off the tubes on the ends, and it left a little bit of plastic in the tube I couldn't get out, which causes some friction on the axle. It's not as strong as a 24 tooth clutch gear; it actually feels just right for a limited slip diff, I think kind of like Ralf's LSD.

TLH

[jesse66058]

I have bad news. When i started making my own i wanted to have it fully lego system compatible, but it's not.

I sanded it down to excactly 1 stud:

But the slack has improved between the bevel gear and differential which makes it after all useless. You can still use it when you sand it uneven down, but you have to make custom axles like x,8 studs + you can't use direct input of cv-joints anymore.

Maybe another design could work which is not bigger than the insides of the open center frame.

[allanp]

Anyone care to try to make this?

[JM1971]

Anyone care to try to make this?

Sweet design, could use a dremil to cut the + in a normal hub?

[aol000xw]

If anyone prints that in nylon send me a quote please

[allanp]

Sweet design, could use a dremil to cut the + in a normal hub?

Sorry no, this hub would need to be larger to accept these new 20t bevel gears with boss.

[jesse66058]

If anyone prints that in nylon send me a quote please

Count me in to!

Maybe if we ask Efferman or Alasdair Ryan gently they want to help......

[efferman]

i can try to make a diff like this and send you the files for 3d printing

[Lipko]

Interesting. Do you have a link to the older discussion, or force diagrams? Thanks!

Somehow I find it quite difficult to explain simply, but I try:

Try to put the fourth gear (no third gear!) on the side of the differential housing without a pin. It just falls out even at negligible torques.

This is no different when there is a third wheel (the fourth wheel doesn't "know" that there is a third wheel since it's not in contact with it), so the fourth wheel would still fall out under any load.

The only reason it doesn't fall out is the fact that the third gear on the pin bares the all the load.

Or:

Build the four gear setup and put load on the whole differential (don't let either wheels turn). All the gears will be held in place (no slack) but the fourth wheel will still have some slack. A good sign that it doesn't bear any load, so does nothing at all.

Edited February 18, 2014 by Lipko

[Zerobricks]

Anyone care to try to make this?

What if you use normal gears and a 4L bar?

Edited February 18, 2014 by Blakbird
: Removed large quoted image.

[Blakbird]

Interesting. Do you have a link to the older discussion, or force diagrams? Thanks!

I don't have the link handy, but I explained all the forces in detail at the time (I analyze gear systems for a living). Just think of it like this: the spider gear transfers load between the two output sides of the diff. Imagine that the input was locked and you turn one of the outputs. The other output should spin the opposite direction right? If there is no guide pin for the spider gear, the spider gear just gets ejected from the diff. You can try this easily enough by installing the spider gear in the unpinned side of an old diff. No gear can transfer load between teeth without a bearing support on the axis, it just doesn't balance.

On the other hand, the diff proposed by allanp has two spider gears and actually does have increased strength because both are guided.

Edit: Found it!

Here is an old diagram I did to show the basic forces on a gear. This particular load case shows a torque introduced at the axle being reacted by a single tooth. You can see the additional bearing reactions at the axle (B_x and B_y). This is why it is so important to have good lateral support for your gear designs.

A differential spider gear does not transmit any torque (T=0) because it is free to spin on its axle. A simplified free body diagram (ignoring radial forces) looks like this. You can see that the force at the central bearing is twice the tangential force at the tooth. If there were no bearing, then no load can be transmitted through the gear (F must be zero).

In the LEGO example provided earlier in the thread, the only reason the second spider gear does not fall out is that the first spider gear is keeping the output gears indexed.

For bonus points, can anyone guess why LEGO didn't just make two spider gear guide pins in their original differential? There is a good reason!

[DrJB]

I agree with BlakBird ... adding a 4th gear will do absolutely nothing. Such 4th gear will always spin (and transmit zero torque) no matter what. If you really want it to help with the torque, you need to add a pin for it to ride on ... but then that makes assembly very tricky (though not impossible). A simple force/torque balance can explain this.

As to the other point that, the 14 teeth gear offer higher/lower reduction than the 12 teeth gear ... this is also incorrect. There is NO gear reduction taking place inside a differential cage, and it does not matter how many teeth you use (as long as all are the same). The teeth if such gears are sized purely based on the maximum torque one needs to transmit to the left/right tires ... nothing else.

What if you use normal gears and a 4L bar?

This is something I'd love to 3D print ...

[allanp]

@Zblj

You could use a 3M axle with 12t gears and it would still work very well. By using two drive gears instead of one the torque threshold is doubled. By adding the bosses to the driven bevel gears you have increased their strength as well as having a length of 1M (module or Lego unit) of axle held in the gear as opposed to only half an M of axle being held. In fact in the current diff the half M wide bevel gear is on the very end of the axle which has a radius on the end which reduces the half M axle holding to about only a quarter, then you also have the indent on the back of the bevel gear which reduces the holding to even less than that. So yeah, whilst my original version with the 20t bevels would be strongest, by using 12t bevels with thicker mid sections (two with + hole, two with O hole), and have it mated to a 36t double bevel gear for the outer crown wheel, then you will have a far better diff than the current one.

Edit: If you wish to 3D print any technic part I would suggest you make holes smaller than the actual size then accurately drill them to size using a 4.8mm drill bit.

Edited February 18, 2014 by allanp