NeuntoterIt's a lie, this isn't really step 1.. I have iterated several servo layouts (in fact you can see glue marks from past failures in one of the pictures). I started thinking about this years before home 3D printers were even a reality and have failed cutting reliably 2 sheets of ABS, dozens of chop sticks, and even wood on numerous occasions. I honestly gave up on this for about a year due to seeing no way to fulfill my desires of symmetry. Forward to November 2016 and welcoming a Ultimaker 3 into my home and all the sudden the prospect of completing this isn't doomed.
So far I have lied and now I must admit I cheated on the project already, too. However in my defense I wanted to test my new printer, my Solidworks skills are limited , and so I came across a project on thingverse(quite by accident) that was exactly the style and shape I wanted and designed to fit my servos. So I began printing the lower leg portion. The printer passed my test with 6 prints, 0 failures, and only one little corner that didn't stick right on one leg.
I started with the lower leg portion because I am of the belief that it is the most important part of any walking robot. Well the feet more so than leg but on a hexapod it is basically the same thing.
The legs are attached to their servo using 2*(2Mx10mm Bolts & Nuts)
In truth step 1 should probably be drawing out some designs and perhaps doing some math etc etc etc... and I have made some crude cave man like drawings; more suited to use of crayon than pencil, calipers, protractors, straight edges, and other fancy implements. However I wasn't settled on what 3d software until recently and even still I am fighting Solidworks vs Inventor.
Now step 2 will become the actual planning phase, because I must begin looking at leg length and gait.
Random Musings
- My current working plan is to use upper leg length equal to the entire length of lower leg.
- My "hip" is created with 2 servos glued together, at least at this point. This was a stupid idea really... but at the time I had no idea that I would have a 3d printer to do this. I have 3 of them completed so I need to contemplate if I should try to undo them and print something or just make the other 3 in the same way.
TL;DR = I all but randomly chose some leg piece and printed it 6 times. Time to design around them.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.
It's a good start, it's nice to have something tangible you can hold in your hand.
Have you already thought about what you are going to use for the brains and for the servo controller?
Are you sure? yes | no
Very nice to hold something beyond a bunch of servos and failed cuts.
For brains.. hmm
What I had planned is each leg controlled by something like an attiny85, then have some SoC running linux with RTOS kernel patches. The SoC would communicate in primitive movement commands to each controller. I thought this would add flexibility and distribute the workload.
BUT that plan was years ago. I am now under the belief that SoCs are so powerful that the individual controllers would limit the performance. In addition I think the added circuit complexity might outweighs the cost of dedicated servo controller, and there might even be SoCs with enough PWM that a servo controller is unneeded.
In summation. Yes, but I am still quite undecided on basically every aspect. I do know I want it running linux for a myriad of reasons, and that the main control program will be written in C++ and any controllers in C(assuming I can still remember how)
Are you sure? yes | no
I find it very useful to have a dedicated servo controller -- not only makes the programming easier, but also you can test it independently of the rest. Not sure if a servo controller per leg makes sense. A single atmega328p can control 20 servos and act as i2c slave at the same time without much trouble -- I have some code for that at #Servo Controller, if you are interested.
Are you sure? yes | no