HexPi – Hexapod Pi Robot
May 26, 2014
HexPi has all his legs attached and calibrated!!!
It turns out that the servo horns have enough play in them (where attached to the chassis and legs) that along with the differences between servos I could not hard calibrate each leg precisely for centering at 1500us.
Eye-balling it, it looks like there is an approximately 1-1.5 degree difference between the center position of the legs; if needed I will be able to calibrate them precisely by offsetting the center pulse length for each servo.
To preserve my sanity, I labelled each servo lead as follows (from front to back):
L1A-L1K-L1H ……… R1H-R1K-R1A
L2A-L2K-L2H …….. R2H-R2K-R2A
L3A-L3K-L3H …….. R3H-R3K-R3A
Here is HexPi with all his legs!
(Click on the image to see a larger photo)
May 22, 2014
Here are some photos of HexPi with the chassis assembled and all six “pelvis/shoulder” servos calibrated.
Top view of HexPi, looking at it from the back:
(Click on the image to see a larger photo)
Top view of HexPi – the front of HexPi is on the right side of the photo:
(Click on the image to see a larger photo)
Bottom view of HexPi, from the back:
(Click on the image to see a larger photo)
May 21, 2014
(Previous State of HexPi – click on the image to see a larger photo)
I’m back working on HexPi after a fantastic three week vacation!
I dismounted the three legs I previously mounted on HexPi, as I want to re-calibrate all the “pelvis/shoulder” servos. It turns out that my initial idea of where to center the legs was sub-optimal.
This time, I am going to pay attention to the limits to travel imposed by the servo holders. Later, I may remove some metal from the servo mounts to allow for greater freedom to travel – I consider the unnecessary travel limits to be a design flaw, caused by using off-the-shelf brackets for the hips. I’ll take some photos to illustrate what I mean later.
I might even add some limit switches to detect leg collisions, and footpad switches to detect when the leg is on the surface the robot is travelling on.
Meanwhile, I added support for the inexpensive HC-SR04 distance sensor to the RoboPi firmware, object, and library, so I will mount one or two on HexPi. (I am also adding Ping and SeeedStudio range sensor support).
The “Better Mousetrap” servo tester is working, I plan on posting some photos (and maybe videos) of it working soon. It is really nice to be able to test up to eight (soon 16) servos at once.
Update #1
All six “pelvis/shoulder” servos are calibrated, and the chassis is now fully assembled.
Time to start adding and calibrating the legs
Once I add the legs, I’ll add the battery packs, and RoboPi for initial testing (the Raspberry Pi will be added later).
[Navigate the pages in this article by clicking on the numbers below]
