TOP SECRET robot #2

TOP SECRET
robot update notes

Welcome again to bent machine labs.

"bent machine labs: Desire is our one true engine." (TM.)

TOP SECRET
May 2000 update

This month's update is going to be brief, because I spent too much time writing two other related pages about walking and getting started with PICs.
Since I lack a video camera, I have had to settle for taking long exposure pics of the robot. Most of the pics on this page are 4 to 8 second long exposures of the robot walking or turning

I've got a lot of the preliminary software written and despite a few bugs here and there, it's working pretty well. The robot can walk forward, back up, and turn.
I've also changed the gait a bit, making the body move forward a bit each step, so it doesn't sound like it's dancing a samba anymore.
For those of you who are curious, I've posted part of the current software so you can take a look at the walking and turning routines, as well as the basic program flow. This upload is missing the classified routines and as such will not tokenize as is.
Speaking of the classified software, I have been slowly building it & the routines I have so far work pretty well.
I added some feet to the legs to make it walk better.

How do you like the pictures?

Looks like I might make some adjustments to the electronics, and also to the body itself.
Specifically, the servo control setup has got me down. Slow motion with this setup is jerky, and I realized that this is because the servos hit their target positions before the Stamp can load the next coordinates into the SSCII.
Which means that the micro that controls the speed will have to issue the servo pulses too.

So I have decided to build & program my own servo control system using PICs. My wish list:
8 bit resolution on servo positions.
4 bit global speed setting.
Each controller will move up to 6 servos. The outputs for the servos will be in two groups, and each group will be assumed to be the servos of a single leg.
Movement of all the servos on a leg will be coordinated so that they all arrive at their target positions simultaneously.

Since it doesn't make much sense to continue programming until the servo controller situation is resolved, I've just been playing with the bot every once in a while. Considering different designs and such.
The excitement of 3 degrees of freedom per leg is a maddening temptation.
The servo batteries hold up for a couple hours, as expected, but I might upgrade to a 2500 mAh pack if I add more servos. I've also been looking into ways to clean up the servo power system so I can power the electronics off the same batteries.

Oh! and for obvious reasons, the robot now has a name: Petal.

more fun to come!

back / next

home

Everything copyright, 2000 Dave Benz

	This month's update is going to be brief, because I spent too much time writing two other related pages about walking and getting started with PICs. Since I lack a video camera, I have had to settle for taking long exposure pics of the robot. Most of the pics on this page are 4 to 8 second long exposures of the robot walking or turning
	I've got a lot of the preliminary software written and despite a few bugs here and there, it's working pretty well. The robot can walk forward, back up, and turn. I've also changed the gait a bit, making the body move forward a bit each step, so it doesn't sound like it's dancing a samba anymore. For those of you who are curious, I've posted part of the current software so you can take a look at the walking and turning routines, as well as the basic program flow. This upload is missing the classified routines and as such will not tokenize as is. Speaking of the classified software, I have been slowly building it & the routines I have so far work pretty well. I added some feet to the legs to make it walk better.
	How do you like the pictures? Looks like I might make some adjustments to the electronics, and also to the body itself. Specifically, the servo control setup has got me down. Slow motion with this setup is jerky, and I realized that this is because the servos hit their target positions before the Stamp can load the next coordinates into the SSCII. Which means that the micro that controls the speed will have to issue the servo pulses too.
	So I have decided to build & program my own servo control system using PICs. My wish list: 8 bit resolution on servo positions. 4 bit global speed setting. Each controller will move up to 6 servos. The outputs for the servos will be in two groups, and each group will be assumed to be the servos of a single leg. Movement of all the servos on a leg will be coordinated so that they all arrive at their target positions simultaneously.
	Since it doesn't make much sense to continue programming until the servo controller situation is resolved, I've just been playing with the bot every once in a while. Considering different designs and such. The excitement of 3 degrees of freedom per leg is a maddening temptation. The servo batteries hold up for a couple hours, as expected, but I might upgrade to a 2500 mAh pack if I add more servos. I've also been looking into ways to clean up the servo power system so I can power the electronics off the same batteries. Oh! and for obvious reasons, the robot now has a name: Petal.
more fun to come!