Question in relation to starting out building a bi-ped

[NUB]

Hi everyone,

As I am completely new to the fascinating world of robotics, and I would really love to build my very own bi-ped as mentioned above...
I have a few questions (in fact these will evolve in millions more no doubt)

but here I go:

1) I've found a bunch of servo's online mg996r, they are metal gear, digital servo's with a good load of torque (16 kg/cm at 6V) but are they suitabe for robotics?

2) Is the Duemilanove ATMEGA328 a suitable controller for my purpose? and if so how many servo's will it be able to control

That's it for now

Hope to hear from someone soon

Thanx in advance
Kris

[Robo1]

Hi

Building a biped from scratch is good fun. I would say the servos would do the job. The control should be up to it two.

I would recommend just getting a couple of serovs and build a 2 to 3 link chain and then try and control it. You can then see the make torque and what weight they can support. You can then also try out the control and see what it can do.

Then move on to building simple legs and don worry about the upper body.

I normally buy to many piece and end up with lots of spare. that why i take it slowly and do some experiements with a couple of components before making it bigger.

Bren

[limor]

The Arduino on its own is not right for controlling multiple servos.
you need many individual pwm output ports and separation between the battery output and the regulated voltage that's on the arduino board.

You can use this board :
http://robosavvy.com/store/product_info ... cts_id/551
( Inex - 16 Servo Controller Board with USB and TTL interface )
which can power 16 servos.
Either the Arduino or the PC can serve as a brain that orchestrates the servos communicating with this controller board (usb or uart).


If you have the budget check out this other option which gives you a full blown miniature PC that can also power your servos.
http://robosavvy.com/store/product_info ... cts_id/521
http://robosavvy.com/forum/viewforum.php?f=17
http://robosavvy.com/site/index.php?opt ... Itemid=135

[NUB]

Hi Robo1 and limor,

Thanx for the support, I'll try the inex for now, it looks like an affordable board, the other option is of course more attractive to me, yet as I'm new and more importantly only have a small budget I'll have to stick with your first tip, limor.
I've googled the roboard too, just to torture myself what I wondered about is this... is this kind of board the only way of building a selfcontaining robot?
I'm talking about a robot that can function without a pc connection, just a remote and some programs?
Not that I'm up to that already, starting out small with just some servo's and a pc-connected board is all I think I can handle for the time being.

Hope to be able to get my plans worked out by way of your tips

Greetz
Kris

PS Does anyone know of a robotics enthousiast site or community in Belgium? "I know, I know! Where the f**k is belgium? Oh wait, I think I know, isn't that the capitol of Brussels?

[limor]

"just a remote and some programs" - connecting a PS2 to the arduino can be an interesting project (you can get the madcatz ps2 wireless controller for almost free at ebay or amazon).

[NUB]

Hi there, back again, stuborn me, huh...

Well I did it, I just ordered the lynxmotion SSC-32 serial servo controller...anyone have any experience with that?
Specs looked good, price even better... only paid the european equivalent of about 40 bucks for it.
Now on to the servo's, I'm about to get myself two packs of 4 MG996R servo's.

Thanx for having patience with me, will try to check in very regularly since I have more and more questions...

kind regards
Kris

[NUB]

I've been reading about the muscle wires, some electrical wiring that has a contracting ability when charged or heated and re-assumes it's primary shape when unconstrained when the external power/heatsuply stops... would seem like an ideal way to create muscled joints for arms and torso... has anyone got any experience with these sma? What is the reason they haven't found their way into hobby robotics yet? Are they hard to control of is it because they are really heavy consumers?
just wondering...
project underway still waiting for the goodies...

[SK]

There were some experiments with an SMA-based (specifically Nitinol) robot at my university some years ago. With observations from this robot and the generally very sparse use of SMA by the robotics community my conclusion is that this stuff just isn´t ready for primetime yet. Sounds great in theory, but there are all sorts of problems regarding latency, low robustness, hysteresis etc.
I certainly wouldn´t recommend building such a robot to a beginner. It would probably only be usable for very basic research, which doesn´t necessarily translate to fun (On the other hand, if you´d get that robot working really well and solve those problems, you could earn big bucks from that )

[NUB]

i'll definitely steer clear from it when trying to have as much fun as i could possibly have figuring out how to build and program a servo-driven bi-ped.
but if i do ever get there i'll let you know. And yes, specifically Nitinol seemed like it could do the trick! Was just wondering since the prices seemed far less expensive than i'd have guessed. then again i wouldn't have a clue what kind of hardware would be needed, hopefully as my experience slowly grows I'll learn a thing or two and maybe one day be able to get something cookin' as for now, still waiting for the regular servo goodies

greetz
kris

[xx2747]

Hello from the US,

I am working on (almost) the same project as yours. I am building a custom made biped with cheap chinese servos and a SSC-32. The servos are not MG995s but something similar to that. As of the controller board, I will be using the SSC-32 but right now I'm going to use the one from my RoboNova for now because I want to see if my design works before I spend $60 on the board.(I have a very tight budget so...

But that's awesome to know that someone else is doing the (sorta) same project.

Also, how are you making the frames? Are they going to be hand-made or ordered?

Good Luck

[NUB]

Hi xx2747, thanx for the reply! Sounds indeed very similar to what I had in mind...your plans.
the brackets and stuff are supposed to be hand made, I was thinking about some sort easily crafted plastic, alluminium or maybe even just plain old silicon boards...
Right now I'm actually really pissed since I blew up one of my servo's because I used a lipo with a higher voltage!!! That little mistake will probably set my project back a week or three, till the new servo's arrive in the mail
but that will give me the chance to look into the programming part of the project, I am exactly what the name says... a newby so have to start looking into c# and all of it's options...
Nice to know someone else is out there with the same hopes and troubles!

hope you make it with your project!
Keep me informed, ok?

Greetz
Kris

[Robo1]

Hi NUB

Blowing up servos. Happens to the best of us. I blow up a DX117 and was not happy, cost me lots to get a new one.

If I was you and want to practice programming and the servo control. Put 2/3 servos together and try and control them using inverse kinematics. It's not to hard and will be a good thing to learn. Try and get them to move around in a circle keeping the same hight.

If your interested I could always send you some example code.

Bren

[xx2747]

Hi Kris,

I'm sorry to hear about the servo.
So far I haven't thought of the programming for my robot. I do not know any programming languages (not even BASIC) so I'll use the SEQ software by Lynxmotion.

To keep the costs down, I made my brackets by hand. I used aluminum.
Also, how do you plan to attach the frames to the servos? nuts and bolts?

As I understand it you don't live in the US?

Thanks

[xx2747]

Robo1:

I thought DX-117s were like unbreakable... Don't they cost like $200 a piece?

[SK]

Even though the Robotis servos are very tough and certainly among the best out there, they are not unbreakable. You just have to abuse them enough (like with RoboCup bipeds doing very stressfull goalie movements) and some will break eventually.