PDA

View Full Version : First Quad Mech, yay!



vladpp
07-07-2009, 09:02 AM
Hello all,

Allow me to introduce myself. My name is Vlad, i live in Melbourne, Australia. I've always been fascinated with robots and computers ever since i was fairly young. I never really had the means to build a robot, so i got into cgi and game design. Last term in one of our uni classes the lecturer keep worshiping the arduino as great i/o interface for physical games. A megaduino and 12 servos later and i got half a robot done. Ill add some photos tomorrow.

In the mean time here's a short movie me and a friend did. YouTube - Evolution

And the 3d model i used to cut out the templates for the body, witch ill also use to figure out the IK.
||
||
||
\ /

Adrenalynn
07-07-2009, 09:07 AM
Hi there, and welcome to the forum!

Nicely animated! I suspect you'll find when you go to implement that model in the real world that its long legs supporting the mass off-axis like that will crush your servos.

Have a look around the forum for IssyDunYet and Bhekka - the two first real-world quad mechs that competed in Mech Warfare this year at Robogames.

Again - welcome to the forum!

darkback2
07-07-2009, 11:12 AM
Welcome to the forum.

Adren...you forgot squidword (http://forums.trossenrobotics.com/showthread.php?t=3100) and charlie (http://forums.trossenrobotics.com/showthread.php?t=2481)...

Vladpp I think part of what your looking at, and I may be completely wrong are bots like phoenix (http://www.trossenrobotics.com/phoenix-hexapods.aspx)and IC hex (http://www.trossenrobotics.com/micromagic-systems-hexapods.aspx)...They use the servo horns as structural elements like your template. This is ok for them because they are relatively light hexapods. There is a lot of weight associated with mech warfare, and the weight is distributed over fewer legs.

Part of it would be supporting the back of the leg so that its not torquing around the servo horn. Another consideration is that the way your servos currently sit they will always have to be working to hold up your robot. Try to find a way to reorient them so that they are more balanced might help with the whole 15 minutes of battery life issue.

Here are links to 2 tutorials on building mechs...probably not what you wanted, but I hope you find all of this helpful.

building a mech using DB's bracket system (http://forums.trossenrobotics.com/tutorials/how-to-diy-128/building-a-mech-using-dbs-bracket-system-3099/)
Build your own issy (http://forums.trossenrobotics.com/tutorials/build-articles-130/build-your-own-issydunnyet-3257/)

as for your movie...sweet!

DB

Adrenalynn
07-07-2009, 11:14 AM
:o:o:o Right you are, DB. Mea Culpa! Sorry!

WGhost9
07-07-2009, 11:40 AM
Very nice robot plans; welcome to the forum :)

darkback2
07-07-2009, 12:18 PM
Here is a mech that is more along the lines of the pic that you posted... (http://mech-warfare.com/scythe.aspx)

DB

lnxfergy
07-07-2009, 12:31 PM
Here is a mech that is more along the lines of the pic that you posted... (http://mech-warfare.com/scythe.aspx)

DB

Sure, but has anybody seen that walking while carrying the requisite mech gear?

-Fergs

Adrenalynn
07-07-2009, 03:57 PM
But even that one has a better chance because this joint is lacking:

vladpp
07-08-2009, 07:35 AM
First let me clarify that i don't intend on this being able to do anything else besides teach me the basics of building and controlling the robot. Destruction and world domination are optional :veryhappy:.

It appears that phoenix weighs 1.644... kg without the batteries. My robot's parts (servos, battery, controller and a 30cm^2 piece of wood prior to cutting) weighed 1.510. The wood has already been cut so i might as well test it and see how/if it performs.

The templates on the metal sheet are for connecting the shoulder servos together.I didn't include it in the design because it was a last minute addition and it was faster to draw it by hand.

Adrenalynn: when you say joint: are your referring to the physical connection between the two servos or the connecting the leg directly to the servo shaft?

Adrenalynn
07-08-2009, 11:09 AM
I'm talking about having that servo at all.

The weight from the body is pressing down on that point of rotation (joint). Those servos are having to fight to hold the body aloft, and they're not all that helpful since without them you can just swing the foot up and out, then move the leg forward, then swing the foot back in to take the weight.

ScuD
07-08-2009, 11:20 AM
The first bot I made illustrates Lynn's point:

http://www.arrickrobotics.com/robomenu/arachno.jpg

It's not too clear, but keep in mind this thing was built in 2000...

Anyway, notice there's only one servo to move the leg upward, it actuates the tibia. The femur is stationary in the vertical plane, but can rotate on the horizontal plane to move the leg back/forth.

As an added bonus, this design has almost zero torque on the servo's when standing upright, so I got away with measily 30 Ncm servo's to get the thing walking.

Needless to say though, there's not too many options as to it's walking gaits.

Adrenalynn
07-08-2009, 11:25 AM
That's an ubercool design!

ScuD
07-08-2009, 11:45 AM
Thanks! too bad I didn't know anything about brazing back then yet, those are plain coat-hanger wires soldered together, with brass screw-terminals as joints. I even built some shock absorbers :d Still have it gathering dust in an old shoebox, but most of the joins have broken off..

Adrenalynn
07-08-2009, 02:33 PM
I think what's so cool about it is exactly its construction method. Very approachable - probably one of the most approachable quad walkers I've seen.

ScuD
07-08-2009, 02:55 PM
True, anyone with either a high-powered soldering iron or miniature torch could build one of these.
They're just not as mechanically forgiving as other solutions, but I guess you could turn this into a week program for kids wanting to get into electronics/robotics.
Cheap, easy, and not meant to last.

Hmm.. maybe I should draw out those templates again.. I have too little time..

vladpp
07-10-2009, 06:06 AM
Ive decided to ditch the shoulder servos at this stage of development. I`m not sure if/how the thing is going to turn, but for now experimenting with the walking gate should be challenging enough. Not sure if its clear in the picture, but the servos aren't actually connected.

ScuD
07-10-2009, 06:13 AM
Interesting design, looks a bit like a praying mantis :)

You can achieve rotation by torqueing the body, as a tank would do, i.e. move one tread forward and the other back. You'll have to use small steps though, as there will inherently be drag on at least one of the legs (unless you can get it to balance on two feet)

Adrenalynn
07-10-2009, 11:24 AM
You decided to keep the servos that will be under the most torque and stress, and get rid of the ones that wouldn't be. :) And they were also the ones allowing you to have a real walking gait, weren't they?

robologist
07-10-2009, 04:26 PM
Scud, I agree with Lynn, that is a really cool quad design.

ScuD
07-12-2009, 03:32 PM
I've decided to make the design public.
A tutorial is in the works (draft in progress).