Theo Jansen legs - possible applications

tj-biped_svg.svg (44270Bytes)
tj-hexapod-sidelegs_svg.svg (18893Bytes)
tj-multipods-frontrearlegs.svg_.svg (24647Bytes)
tj-sidelegs-arrangement.svg_.svg (84003Bytes)

Here is an open invote to all of you to join yet another Shoutbox debate. Oddbot and I are both thinking of ways to mount a vehicle, or creature, on Theo Jansen legs. (use the search box)

Most creatures living on the 'net have two sets of three legs. Each set on a side. In the front and in the back. Many creatures I saw even double up on the legs: six in front, six in the rear. All connected to the same crankshaft in the center of the creature.

See the original Transporter or Animaris Rhinoceros

At best a TJ-leg touches ground 50% of the crank's revolution (or 180 degrees of the cycle). The foot is air borne the other half of the time. Probably a bit longer. So for a set of legs to replace a wheel, it needs at least two legs. A stable vehicle needs three points of contact. So either give the creature four sets (Transporter) or plenty of legs on opposite sides of the vehicle.

Oddbot poses the question of steering the vehicle. Tank style steering (stopping one thread, powering the other) seems somewhat inelegant for this beautiful piece of art.

I plan to build a "shoe box with legs flush against the sides". But to be honest, steerability has not entered my design plans yet. Mostly because I plan to make it (eventually) a towed vehicle. Maybe I should drop Mr Jansen a line. My version probably will be much lighter to pull...

All this for a trailer?!?!?!?!?!?

What?!?!?!?!? (see title)

 

too much hype?
Just because it’s not self-powered, does not make it less impressive. Besides, my plans evolve faster then my legs. Did you not see my TING video?

how many negatives makes a good argument?
I tried to correct the double (?) negative in the previous comment, but may brain got twisted. Just to rephrase: very cool even without power.

Rik takes svg lessons

 


tj-biped.png


The point of this silly drawing is to present a common nomenclature: words we can all use.

In dark brown: the drive train. The $ axle is the shaft that is driven by a motor. The € axle is fixed to the chassis. The leg parts rotate around it. This is the vehicle's weight bearing point.

In this example, the starboard crank has a 180 degree phase difference from the port side crank. The $ axle is divided, but could easily be one and the same.

When looking at the green (starboard) side, a clockwise rotation of the shaft will also make the foot rotate clockwise. When point M (where the rods connect with the crank) is in any position between noon and six (afternoon), the foot touches the ground. During the night (six until midnight) the foot is lifted.

This creature will not walk. It will have it's tummy scraping the dirt while hopping from one foot to another.

 

One more before bed

Man, these lessons are time consuming and brain warping. But I’m getting closer!

This is a picture I have had in my mind since dont-know-when. Does it make any sense?

legs in front of the bot (or in rear)

When a number of legs must replace a wheel, it’s important to coordinate the timing of the steps.

This first example has three feet on the left (port side) touching ground, then three in the air, then three on the ground again and finally three more on the ground. I only drew the cranks for the first six legs. You can imagine the remaining ones.

You can see how the stages of the legs are “phasing” through the constellation. Compare Animaris Ondula.

 

This example has the "wave" rolling inwards from the outsides. Note: this could have one long crank shaft.

 

This last one could get away (...) with only six legs. By better spreading the touching and the swinging feet across the width of the vehicle. Sould it still need another set in the rear of the vehicle? The foot placement seems pretty well distributed back to front as well.

I suppose it would tip over. In the indicated position, all three touching feet are aligned. It would fall on the right side of its face...

 

Pretty Pictures

Ok I have to ask for the record since the guys with the prettiest diagrams are here.

What software do you use for your drawings ?

Rik?

OddBot?

Interesting article Rik, although I’ll have to re-read the discussion a couple times (leg bone connected to the knee bone …)

inkscape

Inkscape.org because ponoko.com supports it and because it is free (and gratis too).

I think Oddbot uses some presentation app that he even uses for his electo schematics. can speak for himself.

[root@home]# yum --install
[root@home]# yum --install opensource !

Visio
I used Visio for my SpeakJet diagrams - the first and only ones I’ve made so far. It’s a bit cumbersome, but would get easier as you tailor the library of pre-defined symbols.

Central Pivot

Rik,

Here is an alternate idea for steering a jansen walker:

Insead of having two legs per drive wheel facing opposite directions, put them both facing the same direction on each side of the drive wheel. Then mount them to the front and back of the robot instead of the sides. It would look like a crab walking sideways.

For steering, add a pivot point in the center, like some big trucks have.

For powered movement, the back legs could be unpowered, and be "dragged" along.

For unpowered movement the natural force of pulling the front section would force the machine to pivot in the center.

http://www.4volt.com/lj/other/PivotIdea.png

Articulated vehicles

I am considering the option of making a creature with a “wasp waist”. And I am very much aware of the option to put the leg assemblies on front and rear of the vehicle.

Note: I use both terms “vehicle” end “creature” as if it were the same thing. I reserve the word “robot” for everyone else’s stuff. Stuff that has brains, rather than “exists in brains” only.

But, as your diagram clearly illustrates, it leaves very little room for the actual creature. Where would I put the battery, motor or even payload?

I like the idea very much and I’m sure Oddbot is also considering something along these lines. Thanks for joining this discussion. Don’t be modest. Jump in. Please.

And feel free to provide bigger pictures. I love’m.

Space

Your right the main drawback is a lack of space to mount or carry things, it seems like most designs have the legs to the side, or don’t have any payload area.

With the pivot design you could maximize the payload area by overlapping your base pate with the rear module’s swivel area, but there’s still not a ton of space.

You could also put the payload above the legs, with more more smaller legs. That would leave you with supporting the platform from the pivot and front shoulder mount. The back legs would kind of trail around under the back of the platform and there would also be considerable downward strain on the pivot joint.

For my design I’ll be putting the legs to the side, it seems like anything but that makes the design complicated fast. Also, my bot will probably be pretty small, so simplicity will be wise in that case. But I do see your point with the elegance of the solution, it does leave something to be desired. I think it might be the lesser of two evils.

As a side note, some day I would love to repourpose a wheelchair with big flat jansen legs off to the side. Imgine seeing someone riding around on that next to a segway, that would be pretty cool.

Step by step, timing each step.

This is my current test setup as seen in my videos. Times three.

tj-sidelegs-pt10.png

Note that I am too lazy to draw each leg in its proper 120° phase difference. One side plate of my test chassis measures 21x30 cm (as does a sheet of A4 paper). That means that this contraption measures about 90 cm or three feet.

That will not do. Hopefully I will be able to condens the arrangement somewhat.

tj-sidelegs-pt10-condensed.png

But when this leg starts to move, it needs a lot of "elbow room". Maybe, I should connect two legs to one crank after all. Something like this.

tj-sidelegs-pt10-alternate.png

This drawing does not do the design much justice. The third leg has no twin and should be placed further to the left. But it is only meant to give a rough impression. Too bad InkScape does not do physical simulations. Or at least, I don't know how it might.

 

All three cranks and legs need to maintain an accurate phase difference of 120°. How will I achieve this? I am asking your help here folks. What do I need to shop for? I have two main alternatives. Gears (which I might fabricate myself with my jigsaw) or a timing belt (which I would have to buy).

tj-sidelegs-pt10-gears.png

I omitted the legs, much easier to draw and to see. Imagine the gears on in the inside of the chassis. They are driven by the red gear. The two smaller gears could of course be bigger. The five brown gears could be all the same size. Just as log as they all turn in the same direction at the same speed.

 

Here's the same drawing with a timing belt. Kinda.

tj-sidelegs-pt10-timingbelt.png

I have zero experience with these belts. But they seem like the strongest and most accurate solution to use. Would a setup like this work at all? Does the middle pulley have enough grip on the belt? Does size matter? Much?

 

What do I shop for? Automotive belts seem so heavy and rigid to me. And the teeth are so coarse. Aren't they? How about this one.

(rapidonline)

It's 9.8 mm wide and can be as long as 840 mm. That means a max. axle-axle distance of 300 mm using two pulleys with 25 teeth and a diameter of 40 mm. That means that the third pulley needs to be driven by a second belt. Hmmmm. Not very KISSable.

Belts

This is the place I always go to look for belts and gears, they have just about every kind you can imagine, with belts up to 72in+.

https://sdp-si.com/eStore/

The site leaves a little to be desired, but if you have the patience you can find about anything.

Also, if you decide to go the gear route and there not too big, I don’t ming cutting a few on my laser for you:

http://4volt.com/Blog/archive/0001/01/01/laser-cut-gears-and-inkscape.aspx

About your belt setup, the

About your belt setup, the middle pulley is likely to slip a tooth or several. Usually the makers specify that the belt needs to engage a certain radius of the cog, in order tonot slip. So it might be better to have 2 cogs coupls in the middle, with a belt from one to the front and another belt from the middle to the back.

Stock drive is a good source, Small Parts probably has a few, and perhaps WM Berg too.

thanks for the links

they will keep me busy for while

Thank you too Jeremy and Russell.

I did it!

I ordered a shipload of parts. From technobots I ordered timing pulleys and belts and shafts and collers and bearings.

And to replace the one 9.6 V gearless motor I ordered two geared 12 V motors from robot-italy.

And a Sharp thingey something 120. I’m gonna make me a proper start here bot, methinks. My Picaxe has been jobless for too many months.