I did a little research and came up with a design for the Star Wars chicken walkers (ATST) and that turned out more or less okay; it still needs a few modifications. The design is based mainly on this image, with a few modifications.
Seeing as that went okay, I decided to move on to designing and building an ATAT. I guess the questions I have are, how do you connect mroe than 8 motors to the controller, if you can at all? Is there an easy way to do the hip joints as show in this image? The hips rotate, and the horizontal bar attached to it also rotates, lifting the legs up and down.
If I can’t get more than 8 motors on one controller without Y cables (which isn’t an option, seeing as all motors need to be independent) would it be possible to use more than 1 controller? This is most likely going to end up autonomous with an ultrasonic sensor on the front.
And that leads me to another issue. I’m stumped on how to get this thing to turn around.
You can’t run more than 8 Vex Motors independently from a single Vex Microcontroller without some external electronics. There are a variety of servo controllers that you can hook to the serial port, or you can network two (or more) Microcontrollers through the serial port.
The cheapest way I know to do it is using a PWM->PPM cable to drive a signal splitter. This adds five motor ports (it adds six, but uses one for the connection) per signal splitter. In theory, you can add up to four signal splitters for a max of 28 motors.
Here is the thread that discusses it. At the end it also shows how to mod a VEXplorer receiver to be used instead of the signal splitter. The library is also posted to the Code area of the forum.
The Cables are not a Standard Item. You will need to make your Own.
For using Just two Controllers, you can connect the TX Line of Controller #1 to the RX Line of Controller #2, and the TX Line of Controller #2 to the RX Line of Controller #1. If the Two Controllers do not have a Common Battery, you will need a Communication Ground between the Two Controllers as well…
This design will allow you to still access the Programming Serial Port on either Vex Controller…
You will need to create a Protocol between the Two Controllers to indicate what position the Motor or Servo needs to move to. A Protocol that allows Feedback would be preferable (AKA Bidirectional)
You did not mention which Programming Language, and which Vex Parts you have…
One idea to try would be to use the Chain and Sprocket and have the Sprocket on the Appendage, locked to the appendage with the “lock plate”, and connect to the Robot with the “bearing flats” (See Vex Inventors Guide, Section D, Page 2-3) This would give you an appendage on a joint, that could be moved with a Servo or Motor. Since the Appendage can not move in a 360 Degree Movement, a Servo might be all that is needed.
Well, how does it turn? Can it just be reversed?? Can One side be operated at a slower rate the the other??
I ended up using a chain-and-sprocket part for one stage of the gearing for the secondary hip joint. Thanks for the recommendation!
As for turning, I did a little research. The original 15.5m tall ATAT has a turning radius of about… 3km. Ideally, the assault transports would not need to turn around. But yeah, I’m thinking of testing it out once I get the motors synchronized through code.
As for parts, I have virtually unlimited parts and a moderate budget for ordering more. However, I’m not allowed to cut or machine most of the metal that hasn’t been cut already, so that limits my options a bit.
I am tempted to try that as well… But I have a few other Vex Projects ahead of it…
IIRC from “Star Wars V, The Empire Strikes Back”, they only show the ATAT moving Forward!! 3km would definitely keep the “Movement in Retrograde” to a minimum.
I would think that you could not just stop one side or the other, but you could let one side Drag.
As Always, measure Twice, and Cut Once, if really necessary…
I have been an active member of this forum for almost 3 years, and I have not heard of anyone attempting this…
First off, ATATs are really made to be dropped on one end, completely demolish everything in its path, then be airlifted once it is not needed anymore. So turning shouldn’t be much of an issue - to attack the sides, just turn the neck. Beyond that, it’s supplies of speeders and troops should be sufficient for any other assault than frontal. Plus they’re in formation so that one can see the other from the rear, so really turning isn’t a concern whatsoever.
Second off, they animated it based off of elephants. Look at how they walk, and below are videos of the ATAT walk and a Lego Robotics version of it. Hopefully instead of using 88,000 motors you can manage some system with 4 (really, thats all I see you need. You would only need to link the joints at the leg or even link two motors for each leg. Should be able to manage with less than eight).
To use it you simply connect it to power and the serial port of the vex microcontroller. Then all you have to do is establish serial communication and send it values like this (255,5,157), the first bit always has to be 255, the second bit is the motor number, the third bit is the value you want to set the motor to! It was made to drive servos but it will work for motors as well just make sure to set it back to the neutral pwm signal of the motor.
One alone adds 8 ports but if you want to go mad then you can daisy chain these for a maximum of 128 motors!!!
To use it you simply connect it to power and the serial port of the vex microcontroller. Then all you have to do is establish serial communication and send it values like this (255,5,157), the first bit always has to be 255, the second bit is the motor number, the third bit is the value you want to set the motor to! It was made to drive servos but it will work for motors as well just make sure to set it back to the neutral pwm signal of the motor.
One alone adds 8 ports but if you want to go mad then you can daisy chain these for a maximum of 128 motors!!!
Thanks again, guys! Sorry about not being able to post pictures; I had a head-on collision with a major design flaw that I’m still working on; the gears to drive the primary hip joints aren’t wanting to get close enough to mesh properly, and there’s too much slack in a 3-stage chain drive to have that work. Redesigning…
You should be able to avoid the serial port programming with just using two identical crystals (two microcontrollers, one receiver in each, use two receiver crystals and one remote+transmitter crystal). You could use Channels 1-4 for the legs and 5-6 for whatever else you have.
If I remember correctly, unless the two 84-tooth gears on the leg joints are being used for some purpose I don’t know, you could get rid of one and mount the motor directly on the 12-tooth gear shaft - this would eliminate the chain (which felt a little flabby when I stopped by) and possibly a source of imprecision in your autonomous mode.
Feel free to ask me any questions…I work a block away from you.
John Mueller
The main reason for the chain was because I got lazy and didn’t want to find gears that meshed at that distance. shame
Anyway. Would it be possible to meet with you so that we can talk over the design in a little more detail? That would be awesome. I’m hoping to get my legs-so-far attached to the chassis by lunch.
Edit: Legs are attached, and attached to this post is the image of the ATAT. I still haven’t put on all the logic and control components, but those can wait; I’m taking a break.
Do they have Separate Receivers on Different Channels, Separate Receivers on the Same Channel, or do you have a Serial Protocol to communicate between them??
