The main problem of "slip" that you talked about might be solved by a differential as you said, but there might be another way. If what you need is the ability to slip (as opposed to distribution of torque), why not try a pulley belt system? I know this is not ideal, but I have used this method before when slip was desirable and it worked fairly well. By selecting different sizes of belts, you can "tune" the amount of torque required to cause the slip.

Also, what if instead of using gears in the triangle, you used three belts, each running from the meta-axle to one of the three true axles. Now the rotation of the metawheel and of the three smaller wheels are in the same direction instead of the reversal that you get with gears. If the lead wheel encounters and obstacle that it cannot overcome, it stalls and if the torque is right, the metawheel will rotate.

Third idea is in reference to the Bill Bradley link. In the message, he describes building a transmissionsystem that would allow you to enguage the metawheel and the truewheels seperately by means of a free spinning gear. I just wanted to let you know (in case you didn't already) that the SuperCar transmission is built around exactly this. The set includes 4 free spinning gears with 2 "clutch sleeves" that slide over an axle extender to selectively activate the different gears.

-------------

Selected Text from Karim Nasar's Email
on October 15, 1995

Me again. I just had another idea. I know for a fact that I don't have enough pieces to do this, so if you try it out, let me know how it works.

Take the triangle pieces you start with and on each "arm", attach a 1x4 technic plate (only 2 dots & with hole on each end) like so:

        /o/                                  /o/
   ____/o/                              ____/o/
   o o +(                               o o +(
       \o\               (O\                (O\
        \o\          add  \o\   <--->        \o\
         \O)               \o\                \o\
                            \O)                \O)

Now in the true end of the triangle arm, you put a "floating" 8-t gear. The extention holds the true wheels just like before. This way, you reverse the rotation so that the true wheels are turning in the same direction as the metawheel.

The problem is that this will require 6 1x4 plates for EACH wheel assembly, and I only have something like 6 in my whole collection.

-------------

Selected Text from Karim Nassar's Email
on October 16

I got home last night and tried the ideas. Here is what happened:

Previously Karim said this, which is what he tested:

> Also, what if instead of using gears in the triangle, you used three
> belts, each running from the meta-axle to one of the three true axles.
> Now the rotation of the metawheel and of the three smaller wheels are in
> the same direction instead of the reversal that you get with gears. If
> the lead wheel encounters and obstacle that it cannot overcome, it stalls
> and if the torque is right, the metawheel will rotate.

I built this, and it seemed to work. THe only problem is that I only have one white belt, which is the necessary size. I improvised by doubling a grey one and ignoring the third wheel (only two touch the ground anyway). I replaced the 8tooth gear on each true axle with a bushing and replaced the 24 t gear with a small wheel hub (the hub that is about 2 dots in diameter and 1 dot wide. Each true axle is connected separately to the central meta-axle by a belt, so the belts are in the same pattern as the tri-armed hub.

I didn't motorize it, but instead turned the axle. It seemed to work. It rolled fine on a flat surface, and when the wheels encountered a bump, the whole metawheel was able to turn without reversing the rotation of the axle.

The second idea I mailed you will not work =8^(

I realized after I mailed you that adding the 1x4 plates will get in the way of the second axle. The only way to do this would be to drill out the "+" at the center of the inner triangle so that the axle would turn freely.