Question in relation to starting out building a bi-ped

[Robo1]

Or for me, you send them to a position they can't reach and not notice for 6 minutes and burn out the motor.

To all the people starting out I would say buy this Kit.

http://www.parallax.com/Store/Education ... fault.aspx

It's the basic stamp starter kit. It will teach you embedded programming and electronics at the same time. You will end up using the stamp for lots of things. This is how I started out 4 years ago, and now I'm at cmu working as a researcher working on full sized bipeds for a living. It gives you every thing you need. Once you finished with it you can sell it and make most of your money back.

Bren

[NUB]

Thanx everyone for the support,
SK it's indeed not so uncommon for servo's to go awry, especially with these kind of "weird" things we're trying to get them to do, I guess.
But each and every one, I feel, since it is part of something we're creating from scratch is a little bit of a link in the chain, and when having to replace them that kind of stings...or is it merely the fact that I haven't used that many of 'em yet, does the feeling fade out when gaining more expertise? Anyways...
Robo1, bren, I have yet to google inverse kinematics, but I didn't quite get the intention...
As for the sample codes I would be genuinely greatful if you would be able to share some, yet I warn you I am a NUB, starting my programming course in c# this saturday so I probably won't be able to share anything in return but some more questions and silly remarks about stuff I really haven't gotten the hang of just yet...
xx2747, yes nuts and bolts would be my first best guess, or is that a 'no go' in robotics? I do have to mention I have come up with a rather lazy but very easy way to assemble a very basic testing structure using nothing but some rubber bands and plastic strips, if I find the time I'll post the ridiculously looking creature maybe this weekend...and yes, you figured correctly I ain't no US resident, instead try a little European country called Belgium. Famous for its...well if you know let me know
greetz Kris

[xx2747]

rubber bands and plastic strips... That sounds very interesting!
I too am too lazy to use nuts and bolts so I'm using servo tape (hard to pull off when applied but better than more expensive and complicated nuts and bolts.)

I think we all would like to see some pictures of your progress so far!

Thanks

Ooh, ooh! Belgium is well-known for its world-famous cuisines right?

[SK]

Or for me, you send them to a position they can't reach and not notice for 6 minutes and burn out the motor.

Now that you mention it, we did that a few times too. I distinctly remember a head servo change at the RoboCup 2008 German Open where that happened twice.

[SK]

But each and every one, I feel, since it is part of something we're creating from scratch is a little bit of a link in the chain, and when having to replace them that kind of stings...or is it merely the fact that I haven't used that many of 'em yet, does the feeling fade out when gaining more expertise? Anyways...

It depends. I don´t own a humanoid privately, but work in a relatively large team of 10+ people. Our robots are the platform that run our software, which took a lot of work to make. So when we´re at a competition and a servo breaks, there´s not much room for sentimentality, the robot has to work and that´s what matters.
But of course I can fully appreciate that this is different when creating a humanoid from scratch yourself.

Robo1, bren, I have yet to google inverse kinematics, but I didn't quite get the intention...

In a nutshell, in my own words:
Forward kinematics answers the question "Which is the pose of the endeffector (or foot, or whatever) of my kinematic chain for the given joint values?"
Conversly, Inverse kinematics answers the inverse question: "Which are the joint values by which the endeffector can reach a given pose?"

Forward Kinematics are normally relatively easy to compute (just multiplication of multiple transformation matrices), while inverse kinematics can be very hard to compute, with often no closed form solution (mainly with many DoF).