Pan/Tilt joint

[cyberjeff]

The way you do the robot walk is to walk stiffly with no twist in the spine.

Robots have no spine, no ability to twist and shift weight. I think that is a design flaw. So I've been looking at pan/tilt to at least give my robot quadruped project some ability to steer and shift weight to balance.

I've almost ordered a few but I can't tell the good from the bad. What I would like is, I believe, mostly modest. Plus and minus 45 degrees on pan and tilt (more OK, but not needed). The servos can be fairly slow (geared down) because they normally won't be moving far but I think they may need to withstand 4 kg/cm of torque. I think digital and metal gear, and relatively cheap! I believe I will eventually need 3  sets.

Who has a suggestion?

I live with cats, and it is starting to look like I am putting together a robot cat! At least it won't leave dead batteries in the litter box...

[mklrobo]

  Hello!
I would offer an opinion, where permissible.
You can follow 2 directions; Yogi beara said, " If you come to a fork in the road, take it". 
If you are designing for  a hobby robot, the stressors in the robot will be specific, and thus,
you will have fun investigating the dynamics of that.
If you are designing a robot for functionality, then I would follow simplicity, reducing
design time, cost, and repair/programming issues.
Have fun, either way you go.    I have  not given the spine much thought, because I have
chosen to follow a path that did not require the use.(?) However,  maybe there is an advantage
to the addition.
insofaras a suggestion, the study of Statics, physics should give an insight to the Free body diagram
you will need to determine loads, then torques, then motors to drive, then amps.
The Joys of Robotics!   

[cyberjeff]

Hello!
I would offer an opinion, where permissible.
You can follow 2 directions; Yogi beara said, " If you come to a fork in the road, take it". 
If you are designing for  a hobby robot, the stressors in the robot will be specific, and thus,
you will have fun investigating the dynamics of that.

I just got in the first servos, MG996R, and it  has altered some of my construction ideas. I've only toyed with microservos before and these are 55g and comparatively look like beasts. I had a design that took the bearing load off the servo, but I think I will simplify and connect the "leg" directly.

What usually breaks in a servo? What kind of loads damage them?

If you are designing a robot for functionality, then I would follow simplicity, reducing
design time, cost, and repair/programming issues.

I'm more into the amusement part.

I've calculated, to the best of my ability the torque requirements and scaled to what I think is within range. Torque is not horsepower and I can only guess at that.

I'm trying to make this as light as possible, about a pound, power source will be tethered as I think it would double or so the weight if it was onboard. The chassis and legs are mostly made out of thin "beams" of cypress which is light and strong and has some spring and give to it.

Have fun, either way you go.    I have  not given the spine much thought, because I have
chosen to follow a path that did not require the use.(?) However,  maybe there is an advantage
to the addition.

I've been watching cats move and a cat (or other  animal) can not sit or turn without bending it's spine. Most legged robots that steer require a rotator servo at the top of the leg. I think those can be done away with if the spine curves.

insofaras a suggestion, the study of Statics, physics should give an insight to the Free body diagram
you will need to determine loads, then torques, then motors to drive, then amps.
The Joys of Robotics!   

10 servos. 4 of the 55g, the others are smaller. I would think this would run fine on a 7.2 V NiMH pack of 3 or 4 AH  for the better part of an hour, I want to "switch off" the power to the servos when the catbot is resting. All the servos are rated at up to 7.2 V and I think that is what I should run at.

I have little interest in doing something that has already been done, it's just not my nature. That doesn't mean that I won't use off the shelf parts where I can.

Which Axon are you using?  I've got a Due on order, but the Axons with the built in communications look a great idea.

Thanks for the advice.

[mklrobo]

  Hello
I have bought the Axon II. I have worked through most of the problems you may encounter with them.
and have posted the resolutions in the Misc section in this forum; Analyzing the Axon: Coding, Construction, and Contraptions.
When you buy the Axon, they are not like the other MCU's on sale. I am finding it a challenge to program
one, after I have overcome the first obstacles. The up side is, once you download the AVR 6 software,
you can program any Atmel MCU series, (theoretically). The Axon is the "portal" to a larger world,
because most of the Atmel processors are in everything! I found a Atmel mcu in a digital Watch!
Good Luck!!!

[cyberjeff]

Hello
I have bought the Axon II. I have worked through most of the problems you may encounter with them.
and have posted the resolutions in the Misc section in this forum; Analyzing the Axon: Coding, Construction, and Contraptions.
When you buy the Axon, they are not like the other MCU's on sale. I am finding it a challenge to program
one, after I have overcome the first obstacles. The up side is, once you download the AVR 6 software,
you can program any Atmel MCU series, (theoretically). The Axon is the "portal" to a larger world,
because most of the Atmel processors are in everything! I found a Atmel mcu in a digital Watch!
Good Luck!!!

OK. I'm working my way through that,  thanks.

I'm  thinking that the Axon must support C++ as well as plain C, although I could not find that out. I have a hard enough time with C++, not having objects and classes makes controlling multiple servos tedious work.

I just got my Arduino Due so I will be occupied with that for a while. Note that all the Arduinos, including the Due run on Atmel chips. The Due is much faster and has a good bit more                                                                                                                            memory.

Exhausted, more later.

[Gertlex]

I'm  thinking that the Axon must support C++ as well as plain C, although I could not find that out. I have a hard enough time with C++, not having objects and classes makes controlling multiple servos tedious work.

You can program the Axon in C++.  The Webbotlib Libraries/Suite are the best supported approach for that.