Tuesday, April 24, 2001

Hexapod Study


Before you flame me for owning six turntables and 12 large pneumatic cylinders I must say this; A year ago when I had just crawled out of my dark ages cave and started searching the net for AFOL's I was amazed when someone had more then one turntable and I thought I'd never own one myself!

Then, in August 2000 I bought my first 8480 Space Shuttle and I had my first turntable, after that I rather quickly bought another 8480 (on sale, I could have walked over dead bodies to keep it in my arms after I had picked it up) and two 8462 Tow Trucks. Now I had four turntables and six large pneumatic cylinders

I then ordered some large cylinders from LEGO Service while they were still shipping spare parts..

In Mars 2001 I bought me another two 8480 Space Shuttles when I was lucky enough to find them in a toy store in Sweden.

What I want to say is; I'm crazy enough about LEGO to spend a lot of money on this hobby, hence the parts. If you want to feel envy, feel it for my luck in finding those 8480 and having the money to spend at the time!

:) This was written mostly as a joke to a friend of mine who claimed that he hated me when he saw this study IRL :)


After the Hercules I felt that a robot might be in order before my next serious project. I thought about the robots I've done so far and decided I liked the walkers best. I had already done a Biped and a Octapod so why not do a Hexapod?

Thinking some more I decided I wanted to do a pretty large platform that could carry my radio control equipment and maybe some cool weapon or something.

To build large and simple I thought that the leg should have three joints, a "knee" and a "hip" and also the turning function in a hip. I used MLCad to test a idea I had. One pneumatic cylinder for the "knee", one for the "hip" and a large turntable for the rotation in the "hip". It proved to be doable in MLCad so I was thrilled when I found out I own 12 large pneumatic cylinders, at first I counted to eleven but I had forgot one used in my Hercules at the moment.

Yes, I plan my next project while building the current. That's probably why I always get sidetracked and loose interest in the current project, ah well...

The reason I wanted a "knee-joint" was that I wanted the robot to be able to do nice turns and also walk like a crab, sideways by not using the turntables and also forward by not using the knee-joints but the turntables and hip-joints.

I made a nice control scheme for how to make all this work with my five channel radio.

After I've built the project as I had planned I realized that the legs were far to weak to give the performance I wanted so I skipped the knee-joint and made that part stiff and took the cylinder to make a dual cylinder in the hip-joint.

Now the legs were strong enough but I got serious concerns for my turntables, they tended to separate when I drove around the platform, and that was before adding any weight. Afraid to ruin 6/7ths of my turntable collection I decided that this project was a design study and that the study proved my theory wrong in real life. Damn!

I might fix this turntable issue by placing the turntable underneath the platform so that the platform stands on the legs instead of hanging in them, more on this further down among the pictures when you have a chance of understanding what I mean. It would have been a total rebuild of the drivetrain to make this at the stage I was in so I decided to lay this to the past and perhaps come back to the idea when I feel motivated enough. I build for fun and so should you :)

The numbers represent the name I gave the legs. The circles are the turntables, the colored rectangles the legs and the color markings show witch legs move in unison.

For example, when the robot walks forward it first lifts the blue legs, 1, 3 and 5 from the ground. Then it turns all turntables. The blue marked turntables turn in one direction, putting the leg forward and the red ditto turn backward thus pushing the robot forward since the red legs are on the ground.

The schematic picture to the left tries to describe this state. The arrow is the direction of travel.

Then the robot puts down the blue legs 1, 3 and 5 and lifts the red legs instead, 2, 4 and 6. Then the motors drive the turntables so that the red legs moves forward and the blue legs backward (now on the ground) so that they pushes the robot forward. And so on...

When the robot turns, the robot don't move all red legs forward and blue backwards and vice versa. Instead one of the colors on one side moves back while the others in the same color moves forward, giving the robot a form of tank steering.

Like this (I'll try to explain):

First the robot lift all it's blue legs (assuming all six legs were on the ground), then leg 1, 3, 4 and 6 move forward while leg 2 and 5 move backward. This makes leg 2 pushing it's side forward while leg 4 and 6 pushes their side backwards, thus the robot has to turn (unless something brakes).

Now it should look like the pic below. The arrow is the direction of travel.

Then the robot puts down the blue legs again and lifts the red legs instead. Now the leg movement reverses, leg 1 and 3 pushes their side forward while leg 5 pushes it's side backwards.

I hope you understand this, it took me a while to grasp it my self while building, how the legs should move I mean :) I'm not sure if I could explain this in Swedish either LOL!

This is the first version with the knee-joint that I skipped in the second half-time.

Did I just use a sport metaphor? I did, didn't I? Yuck..

Anyway, in this pic leg 1, 3 and 5 are raised so the robot stands on legs 2, 4 and 6. Leg 1 and 3 is stretched out using the knee-joint.

Another angle, same status.

Same status, yet another angle.

Detailed view of the leg.

When the robot moves, either forward/reverse or turns, the corner legs always move in another direction then the center leg. Thus the drive axle had to spin in a different direction at the center then out in the corners, this is how I made that work. I first thought about several gears and chain links and stuff, then I realized this would work with less friction. It felt great to fell the brain kick in like that twhapping my stupid first thought. I wish it could happen more often :)

This is the second version of this hexapod as far as I built it. Now the knee-joints are long gone and the knees are stiff. I had to skip crab walk and smooth turning when I skipped them. I might have put in some springs in the knees to allow smooth turning never the less. What a thought. Gotta try that on my next hexapod!

One set of legs lifted from the ground (I don't know what numbers but it does not matter :) ).

Close up of the "hip". The leg is mounted on top of the turntable and the chassis in the bottom. This means that the robot "hangs" in it's legs instead of standing on them. Pretty stupid construction of me there. But at least I know how to make it next time :)

This is what the turntable looked like after a short drive. When I noticed this I chickened out and decided to make this an ex. project.

Placing the chassis on top of the turntable and the leg underneath it would probably solve this. It meant to much work however so I never tried it.

Top view. Why do my tummy feel like eating spaghetti now?

Close up of the drive train. I had a similar setup in the other end of the robot although reversed. The two motors drive the same axle and the other drive axle is slaved via the gearbox setup. I had a gearbox in order to allow the steering and forward walking. See schematics in the beginning of this page. When it walks forward the two drive axles has to move in different directions and when it turns they have to turn in the same direction, hence the gearing.

Looks cool eh? :)

It has a nice ground clearance and it seemed to be able to do a lot of terrain driving, damn I would have loved to take this one out in the bushes. Stupid turntables, or is it stupid designer? Nah!

The bottom is n├ądd! Sorry, Swedish joke. This is the chassis any way.

Not much to see, I tried to keep it flat and nice for off-road action.

Well, it would not have been much action any way, it drove painfully slow...

Zooming in underneath this beast the secret of the gearbox is visible. The two z24 gears are there to reverse the direction of the output shaft. I used a odd number of gears on one side of the gear lock mechanism so it had to be a even number on the other side. Power comes in on the center axle from the motor, the two axles in the top and bottom of this image are the drive axles for the turntables. More on the gearbox on page two.

Continuing where we took off, here's another shot of the gears. I like gears :)

The difference from the previous shot it that here some of the gears to the left is visible that were hidden behind a beam. Also, the two z24 don't show off as much.

When I made the gearbox I made it as another part while the platform with all it's legs and such had to wait. I found it more practical to make a smaller version with the parts that were relevant and work from there. This way I could explore the limits w/o having to get tangled in a lot of pneumatic hoses 8) Uhm, this reggae music is making me wishy-washy, what happened to the metal that was pumping from my speakers a minute ago? Bah!

Another view of the drive train. Notice that I mounted the engine upside down on the yellow gearbox. I fixed the gearbox to the beams using 3/4 pins, the dark gray type you know..

Before I came up with the version above this is what I made. It uses 17 gears and one chain instead of 11 gears. Also, this version has two gear pieces that has to move in place in order to shift gear. Not good, so I decided to give it another go and then the version above was born. It's good to redo things when building.

Another shot of the first (not used) gearbox.

And yet another. Why do I even bother to tell ya' folks I failed the first time and then show you so much photos of that failure? It's beyond me and far in on the intelligence of a hen! (Swedish saying translated word by word, always fun). I need to change this music!

Caught in the middle of a step..

Another view of said step. The center legs are parallel so this baby is walking, not turning.

A cool feature was when I raised all six legs and let the robot lay down on it's belly. It looked nice when I "told" it to get back up again :) It always raised one end first. Probably due to friction/efficiency of the pneumatics.

No comments: