?x4 quadruped design. (IRON WOLF)

i have been thinking about building a quadruped following the straight legged design.
here is a photo shopped vision i have done of it…
and yes its two biped scouts head-on.
http://i531.photobucket.com/albums/dd355/innerbreed/quadrupeddesign.jpg

before i go ahead and order the parts i wish to get peoples view on it.
the design will change slightly but you get the idea.

will hs645mg servos be enough? (a normal scout uses HS-5645 but im thinking high torque was needed to control the balance in the scout as its swaying on one leg at a time.)

is there basic atom code for the scout that i can work on?

the design will change slightly but any more input?

6DOF? That’s ambitious! Anything past 3DOF, and you get ambiguities is the IK solutions. And won’t you want to run modified Phoenix code, instead of BRAT code? Or am I missing something?

Might need more then '645 servos, lotta weight!

Maybe put the PCB on top?

Add head/tail!!

Looks cool, 'tho!

Alan KM6VV

4.962lb without battery’s.

im not sure about the coding. i would have though biped code but then again it wont need to sway side to side. phoenix code might be best. im not very good at writing code from scratch. if i do order the parts (if i do it will be in a few months.) then ill start modding phoenix code. might need some help there.

the pcb will most prob be in the chassis.

a tail and head would then be added once iv got the main unit built and working.

will cost about £550 for two scout kits and 12 more hs645mg’s

thanks.

Do you really need to use the Scout legs? You can’t really take advantage of the exaggerated leg moves that work on a biped, can you?

But with those big feet, you might be able to move one leg at a time without the fear of tipping over!

What kind of leg action are you looking for? Scout (and my Loki) use exaggerated leg movements to keep their COG inside their base, which is much narrower then that of a 'quad.

Alan KM6VV

I’d prove the software can be done before plunking down that much cash on hardware.

I think it would look cooler if you flipped the orientation of the legs.

Wow!

One thing is very sure, you wont have any trouble with balance with those large feet. :wink:
Some might think its a tiny little bit OVERKILL with 6 DOF…
5 DOF IK should be very possible, but the 6. DOF (horisontal coxa/hip) gives to many IK solutions, so some sort of algorithm with combined FK and IK are necessary.

I’ve also been thinking of something similar but with reduced size of the feet and “only” 5 DOF per leg.

I guess the logical place for the electronics is a more centered place.

645 MG should manage it, I think. EDIT: I hope… :confused:

once/if i go ahead and build this im sure i will encounter many new design ideas. 6dof is overkill and to save on weight, cost and programming it may well be reduced to 5dof.

got plenty of room for electronics in there. may even lipo this baby up too, hay zenta. :wink:

especial if im adding other dof for a head/tail. but this wouldnt come till later.

im sure ill have plenty of questions later when it comes to programming. :wink:

happy new year.
my wife is ill in bed so iv had to open a crate of beer all by myself. it doesnt get any sadder than that. :cry:

thanks for your comments.

Neat, it kind of reminds me of a horse.

5x4 design. and electronics moved to top.
http://i531.photobucket.com/albums/dd355/innerbreed/quadrupeddesign2.jpg

will bring the cost down too. most important! :wink:

Removing the tars joint was a bit surprising way I must say… :open_mouth:
I’m afraid the IK won’t be any easier by removing that joint.

Would you mind explaining what you had in mind when you choosed this solution?

You also probably need to think about how you are going to make it turn.

the reason for this i thought was that it would be easier to program it first with 3dof (High lighted) servos, and then move on to the others once i get there.
http://i531.photobucket.com/albums/dd355/innerbreed/Captu99-1.jpg

Do you have any other recommendations?
please remember these are only concept designs but any ideas i will take on board.
combined effort is always best.
like i said once i get the kit i sure there will be lots of other alternatives ill come up with. :slight_smile:

I suggest you go to a home improvement store and get a 2’x2’ sheet of 5.2mm luan plywood (~$3 in the US) and some small bolts/nuts to make a simple mockup of your design. Cut the wood pieces to the length/size of each leg/bot part and lightly bolt together so you can move the parts yet they will still hold position. An afternoon’s work doing this will probably save you a lot of future work and agravation and possibly $$$. I always try to identify a project’s most critical and difficult aspect and solve that prior to investing significant time and $$$ in the project. I see potential issues that might not be easily solved.

The big challenge with your latest leg configuration is keeping the foot parallell to the ground, because the 3 DOF IK you mention are just taking care of the position of the end of the foot. Not what angle the foot have in relative to the ground, thats the whole point of the tars joint (keeping the foot in parallell to the ground).

A wooden mock-up? That could be useful. I normally just use an old set of S2003 servos and the brackets. These old and low toque servos can’t support any weight, but when run on low voltage, will hold their position, and can be moved as desired.

But you did give me an idea. I just might machine some plastic servo stand-ins. In fact, having a pot in this mock-up (real servos), would be quite useful. Read the pots, get the servo angle!

Thanks!

Alan KM6VV

a simple and quick cut and paste.

Design #3 5x4
http://i531.photobucket.com/albums/dd355/innerbreed/quadrupeddesign3-1.jpg

Design #4 4x4
http://i531.photobucket.com/albums/dd355/innerbreed/quadrupeddesign4-1.jpg

Hi,

Sorry, I missed your last post.
Ok, so you are dropping the feet?
I believe the 4x4 DOF is a bit more realistic, the IK are still a challenge when it comes to the hip though.

yeah i think its best i work with 5x4 or 4x4. just running though the possibility’s and available designs.
might take a while in till i get this up and running. still got a hexapod project going at the moment so i wont be giving up on that just yet.

the other thing im thinking is that the more i chop away at the limbs the more less interesting it gets. of coarse it will be a lot harder, the more dof i give it.

the coxa servo will change a lot as well, so it will give it more rotational º.

i just don’t want it to be a copy of Felix! you understand.
i could order the parts now but im going to see what other ideas i think of. :wink:

I think the 6DOF leg is do-able, provided you have 6 pre-defined value to get single solution. Since you have the feet, by virtue, you already have 3 standalone value (x,y,z), rotation of feet in y-axis & z-axis. The additional DOF at coxa (the 6th DOF) will be determined by the direction of the feet.

The additional DOF at coxa would play more significant role when the leg has foot, which explain the infinite solution for point contact leg design (due to lack of that variable).

Despite that, I can’t imagine the IK equations for 6DOF, you will need a faster BAP or other microcontroller solution!

im taking everything on board, and the advise is really good. mixed views with both good and bad points.

If the first design was to be built then the rotation of the coxa would be determined by the direction of the feet, and i was thinking that i could use a Y servo connector for this, or just run them as different variables.

I might do ahead and order one scout kit, (being cheaper than buying the SES part separately)
and then from this work on a stable gait that will mimic the movements i wish to achieve for the quad.
“but only for the front legs”

doing this i get the change to play with a biped. kinda.