Let's Make Robots!

Gipsy, steps to quadrupod

nothing but plan to walk at least ;)

As seen in my blog post about Sticky, first attempt to build a quadrupod ended with a jump from the desk.
Hopefully, I learned a lot and I was ready to launch the next experiment.

My wife just received this very nice material from a friend and there’s enough for me too :)

I think the best english translation is paper foam board, but you will see it later.

This time, I choosed to get a step further and build a 3DOF quadrupod.
As usual, as I don’t have 3D software knowledge, I draw my ideas.
Very quickly, thanks to the foam board, I was ready to try real parts:

from thoughts to real

Easy to cut, strong enough, first leg parts were quickly ready:

one leg kit

I had to get back servos and servo horns from the poor Sticky.
Bye Sticky, you were very useful:

good bye sticky

Using the hot glue gun once more, I started to assemble the legs:

new femurs horn glued

When a leg is ready, I thought it my be good to check the 90° position of the servos:

align servo

Quite a lot of work, but that was fun:

4 legs

I quickly cut a platform and here is the first shoot of Gipsy:

structure ready

And there, I learned another lesson: glue is good, but you have to screw horns on the servo.
Until now, I just pushed the horn, but here comes the first true physical constraints.
So, I carefully use my cutter to get back all horns, screw them and glue back.
No true difference to see, but hey, I’ve done it! :)

same structure but horns are screwed now

Then, I started to plug all 12 servos to my DFRobot Romeo board.
Which bring me to my next lesson: always put a proper code on a board before switch on :)

Never plug the board before having initial code

No harm, but I thought I was stupid because I didn’t initialized all the different 12 ports used.
To be honest, I believe there’s also an electrical issue, as I was only USB plugged.
As I didn’t meant to move all servo at the same time for the moment, I thought USB was enough.
Obviously, this is not the case.

Next step will be to determine if this is an electrical issue.
I’ll add an external power source as in Gary

Also, I started to study the way I’ll code the movements. It looks harder to move a 4 legs bot than one with 6.
It’s a matter of balance and you need 3 legs on the floor to keep a stable state.
But I start learning anyway, even if at the end I code an hexapod ;)

 

I'll try to keep updates on LMR but if you want more about me and my experiments, I have a small blog on heroku.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I decided to transfer what was a blog post to a true robot page as I intend to push further my investigations.

I suspect that I didn't use my DFRobot board properly: I didn't used the right power connector, there is a separate one for servo power and there are 4 jumpers to disable to use pins 4 to 6 I think.

I hope to have time to test this this evening.

Just ask Bajdi. :) He made some custom boards and made use of some UBECs(?) (Battery Elimination Circuits) to supply enough current to drive the servos.

I really appreciate the time you take to share the details good and bad with us.

Bajdi is on my list of good info on power management :)

I just lack of time to properly learn everything I need, that's the goal on top of having fun ;)

For the moment, it looks like 5AA will allow me some experiments. So I can start bugging Badji about UBEC (and learn what it means :) )

If I don't post a robot page, I put a bolg note. Every robotics experiment will be tracked here as I owe a lot to LMR.

Are you still working on the power? Did you begin to program it? You can do it even while it's still tethered to a power cable!

I case you need more inspiration, look at the code templates for NUKE:

https://code.google.com/p/arbotix/source/browse/trunk/pypose/tools/models/lizard3/template.ik

https://code.google.com/p/arbotix/source/browse/trunk/pypose/tools/models/core/template.ik

I received all missing component and I'm ready to try different combination of regulator and batteries.

Even better, thanks to your work on uKubik, I worked again on my code.

The way you used the Serial.read didn't worked in my case: don't know why but it doesn't really matter.

The most important is tat I worked on it again and have a way to address each servo from a serial connection.

I already had a look at PyPose when I was dreaming on Trossen Robotics crawlers. I think I have to keep each step small in order to still progress.

So here's what I envision:

1) write simple python code to connect to serial and move one or two servos

2) write a basic GUI to control extended script wrote in 1)

3) add BT to avoid cable?

Lovely critter ;)