A.T.L.A.S All Terrain Land Assault Sphere (Hexapod)

Let's Make Robots Projects Showcase
61

hi kurt,
im using HS645MG’s. i know digital might be better but i cannot afford it right now.
i have been thinking of using start up sequences like just setting another set of parameters for the #Initpos# commands. and call them up at Startup/shut down.

i should have more time next week to play about with it so we will see what happends.
Im not using SSC sequences in this projects so i guess the PO commands will work just fine.

one thing i was looking at was, what command can be changed to make the leg stance wider and smaller. you might not understand why i ask or what i mean as i haven’t explained the reason but im thinking that if i was adjust this then i might achieve the leg angle needed. kinda like a command i can call up to widen and shorten the stance?

62

Hi again,

From what I have seen the stance is mainly dictated by the values of the LegPosX/Y/Z arrays. The code that I put up yesterday on the Contact switch thread, changes the stance such that the foot is square on the ground for a given logical Y (bodyposition + initial Y)… So with my current code the if I make the body go higher, and then hit the “9” key on my remote it readjusts and brings the legs in closer to the body…

Yep I have also kept with 645. Both for the cost as well as the noise!

Without worrying about sequences, yes the PO command works. Likewise you could simply keep it as a local array(s) and simply add or subtract the value in the function ServoDriver or ServoDriverStart depending on which version of the code you are using…

Kurt

63

going back to the Size of the segments, i already feel that they will be to big and might clash.
also i would have thought that this will also happen with body moves 'especially with Y rotation, and translation.

this really wouldn’t be a problem as these functions can just be ignored but i would like to find a way around this.

The option i have is to divide the segments up even more so that the whole leg (or the last two dof (tibia/femur)) adopt a “panel” of each segment.

i may have to extend the tibia sections to allow for this. once i find out the benefits of this ill post the results.

64

This part troubles me a lot too. As long as you are using the complete sphere circle only divided to each leg the sections are really wide. In my case each section is about 14 cm wide at most. But I’m using a different leg design, and still this is going to a big challenge when walking. I think relative small walking steps is one solution.

That is an interesting idea, I’m just afraid it easily get mechanically very complex. Looking forward to see how you are thinking of solving this.

65

14cm.
the front sections come in at 20cm at the widest part and 17.5cm for the middle ones. :imp:
http://i531.photobucket.com/albums/dd355/innerbreed/Image012.jpg

66

hi zenta.
i have been looking over my project and i have been thinking about a few changes, first ill explain the problems so far and then the reason for my changes.

**im expirencing a high amount of servo backlash maybe by about 2-3º per servo. on a 4dof leg that’s anywhere up to 8-12 total degrees of backlash per leg,

Battery drain, mainly due to weight (without sphere) but mostly caused by servo population,

Sphere segments are massive and too big to walk with.
**

*Well im using a 4DOF leg design this will be changed to 3dof.
doing so will help with the amount of power needed.
it will also help take off the amount of backlash im getting back,
and also the total weight will be lowered.

The sphere size will hopefully be changed. im looking into getting a smaller sphere as
this will mean the segments will be smaller and also the weight will also come down.

i think its very important to keep things simple, as you mentioned these sort of projects could become very complex and now i see why.

i am also looking into dividing the sphere up differently to what i have shown in past posts.
im looking at maybe keeping the main structure of the sphere intact and have the segments cut out of it rather that dividing the whole thing up.
*

These might be something you want to consider while building yours. good luck. :wink:

67

Hi Jonny,

I didn’t thought that the servo gear backlash was that a big problem for you since the ball/sphere isn’t going to roll, just stay in one pose. Or am I wrong?

What battery are you using?
If you are not planing to make the hex walk for 30 minutes or so you can use a relative small LiPo. For Archer I’m using a light weight (97 gram) 1700 mAh 2s 20-30C, cheap battery. This battery make Archer walk for about 25 minutes or so. I’m thinking of using the same or two of them in my next project. Even a small battery like this can deliver 34 Amps without any trouble (not for a very long period of time since the total capacity is very limited). So a small LiPo can power any big hexapod. Just for the fun of it I mounted 1 kg of payload on top of T-Hex and powered it with the small 1700 mAh LiPo. Total bot weight of 4,1 kg! :smiling_imp: Both electronics and servos are powered from the same LiPo (no BBII reset :wink: ).

Going for a smaller sphere sounds like a good plan though.

I did mention that I’m using a 28 cm globe as a sphere. To day I peeled off the globe coating/layer on two sphere parts. To the left is one original part. The middle part has the globe layer peeled off, and the right part is sanded by hand.

http://zentasrobots.com/wp-content/blogs.dir/3/files/2010/10/D-Hex00001.jpg
(The servos in the background are from Felix (RIP :cry: ) )

I did try using a sandblaster but I wasn’t pleased with the result. I’m not sure if I’m pleased with the sanded part either. One thing that amazed me was how transparent the plastic is:

http://zentasrobots.com/wp-content/blogs.dir/3/files/2010/10/D-Hex00002.jpg
(I’ve written some numbers on the back just to keep the parts in order)

You can clearly see what’s behind the sphere part. But this can be a good thing too, especially with some LED’s inside the globe. I’m not sure how the end-result are going to be at the moment. :unamused:

Another picture for comparing in size:

http://zentasrobots.com/wp-content/blogs.dir/3/files/2010/10/D-Hex00003.jpg

Like you said, the sphere parts are still very big and will clearly inhibit large walking steps. I guess there have to be some trial and error in a project like this.

68

yes that globe has surprised me. i never realised the coating would peal off. IMO i think the finish is fine.
i too wrote on the same areas of the segments. helps make sure it goes back the same way.

the battery i have is a 6.0v 3300mAh. and a 9v 170mAh rechargeable. :confused:

the hexapod has now been stripped off down to 3dof so im now just working out the best way for the hexapod to sit while sphere is closed. .

im not quite sure how you manage to fit your Hex…/s into that 28cm sphere. my hex needs a 34cm diameter sphere for a tight fit! :open_mouth:

…RIP Felix … :cry:

69

slight change. i can now use a Sphere with a diameter of 25cm. :laughing:
just need to find one. :wink:

Image008.jpg
Image008.jpg1600×1200 371 KB

Here shows a more compact chassis and coxa design.
Image010.jpg
Image010.jpg1600×1200 482 KB

Image014.jpg
Image014.jpg1600×1200 470 KB

70

I really admire you guys’ creativity!

And I enjoy reading about your progress.

Alan KM6VV

71

thank you Alan for that comment! :slight_smile:
Thats very kind of you.

i have added the Boards. this time i am placing them on the top. the reason for this is that because i have made the frame smaller the tars sit closer in, making the base of the Sphere smaller, it makes sense to move the electronics. i want the robot to sit as low down in the sphere as i can get it. with the exception of the battery, that will remain on the bottom.

again some more poor quality pics… lol

Image004.jpg
Image004.jpg1600×1200 479 KB

Image003.jpg
Image003.jpg800×600 113 KB

Image002.jpg
Image002.jpg1600×1200 463 KB

I think the secret is to have the hexapod crouched up as tight as possible inside the sphere, so once initiated the hex will spread the sphere segments out as far as possible, allowing for better movement.

The only problem is finding the right sphere.

With the hex being slightly bigger than a Foot/Soccer ball by 3-4cm i feel it might be easier to find the right one.
ill keep looking. :wink:

72

Looking very promising this time Jonny!

Maybe a sphere (globe) like this :mapsworldwide.com/sku_33169.htm would fit?

Are those the extra long Phoenix like femur parts?
Sounds like a good idea to keep it simple and light with few DOF’s. The opposite of what I’m working :unamused: , I might have to do some trial and error too before it gets perfect.

Was it a problem to stack the two boards, since you mounted the BB2 standing?

73

yes I think your right about the globe idea. seeing that iv called it ATLAS it would make sense. :wink:
Those Femur parts are the new 4" (phoenix) ones. (Thanks Jim)
I didnt have a problem with mounting the boards it was just I like the idea of having the serial port at the top. Also it made it much easier to connect everything as it makes things more accessible.

I think this is the most compact “Standard size” hexapod iv seen. Mounting the femurs on top of the coxa has made things a whole lot better.
I’m thinking of changing the tibia’s with something that is more curved so it makes better contact with the inner surface of the sphere.

74

I have been spending the last few night thinking about this project and also about how im going to get the right sphere. funding is tight so i have been looking at other ideas.

I have been looking at making my own sphere…ish design for this.

I’v looked at trying to find the sinusoidal projection of the sphere (Maps), Origami, etc.
http://i531.photobucket.com/albums/dd355/innerbreed/ft_thumb02.jpg
The last few nights i have been looking at Sphere Math!
i have learned so much about the mathematics of a sphere. i didn’t realise so much Math was involved in this subject.

anyway, i have been looking at how to calculate the total size of each segment, or “Lune” as typically known, using Math.
The idea is to make 6 identical Lune’s that will go together to form the Sphere.

I was trying to Reverse engineer the sphere:
If a sphere is divided into 6 equal segments from pole to pole, How do i work out the dimensions of each segment if laid out flat.

The only thing i had to start off with was the Diameter of the sphere. This number was taken from the Distance from opposite legs on the hex.

The goal here was to calculate the lune’s dimensions, curves etc.

Opposite leg distance: (Sphere diameter) D= 250mm
circumference: (DxPi=C) C=785mm
Sphere segments: =6
Segment width: (C/6=W) =130
Segment hight: (C/2=H) =392.5mm (arc)
360/3.14/6=20 (H2 =200mm)

to calculate the diameter of the top and bottom sections
H-H2=192

diameter of top and bottom sections = 192/2=96.25
to make these sections we need to make two sections with a 96.25mm Radius

using the above Math i made these Lune’s:
Iv used the Old Yellow disk as a stand to hold the robot. its no longer part of the project.

Image020.jpg
Image020.jpg1600×1200 387 KB

Image021.jpg
Image021.jpg1600×1200 418 KB

Image022.jpg
Image022.jpg1600×1200 445 KB

I’v used 3mm Black Sintra for this. boil in water for 5 seconds and it gives you about 5 seconds of moulding time.
after making them i realized i couldnt curve the edges in two directions at the same time. thats why you can see the gaps.

anyway. back to the drawing board.

I re-evaluate the Math a realized i made a small mistake. anyway here is the new Lune’s.

Here you can also see the Cap that will sit on the top.
I have also adopted the same bottom section as made in the first set of Lune’s.
http://i531.photobucket.com/albums/dd355/innerbreed/Image024.jpg

Im going to cut off the grey section so the base of the robot sits flatter. it also makes the legs sit correctly in the “Sphere”
http://i531.photobucket.com/albums/dd355/innerbreed/Image025.jpg

Once i get the money i will order the Sintra sheets and cut the new sections.

Yes this wont be a completely Spherical Sphere. more like a Spheroid with flat sides. Might look more robust too. :wink:

75

I think the early cartographers had the same problem! Looks like you’ve worked your way around it, 'tho.

I wonder, if you had a suitable sized sphere, could you use heat to work Plexiglas? Use your cut-out lune segments, tack (tape?) them down to the sphere, and apply heat. You might even be able to cut them out rough, form, and then trace the proper cutting lines. Or is that kinda what you’re doing?

Alan KM6VV

76

Thats pretty awesome man. So are the spherical plates gonna extend past the base, or will it sit on the base with the legs above it when closed up?

77

@alan:
well im not forming the sphere segments over anything, im using measurements and math to determin the form and shape of each part.

@dennis:
thanks. That bottom plate is from the older design from a larger sphere i wap using. In the pictures i have only used it as i stand for the hexapod. From the center of the chassis is a tube that fits nicely to that disk. So its there only while i build it. Once i have added the sections i will removes the disk and then remake a new, smaller base.

78

Looking to get new bits by next month. :slight_smile:

79

i have decided to share some hexapod images with you. this is the latest assembly with the electronic housing installed.
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12509.jpg
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12505.jpg
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12508.jpg
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12506.jpg
http://i531.photobucket.com/albums/dd355/innerbreed/SDC12510.jpg

its a bit dusty as iv not done anything with it for two weeks.
i will be receiving the Sintra® sheets this week so i hope to get most of the lune’s made by the weekend.

80

Looks naked, but very spider-like!

Alan KM6VV