I am interested in building a hexapod like the Lynxmotion CH3-R, but would like to know how the servo motor was chosen (i.e. the specific quantitative mathematics/engineering involved). The link to the CH3-R kit is given below:
I read the leg torque tutorial, but I am not entirely sure how to do the math for a radially symmetric hexapod. Can I assume that the weight acts at the geometric center of the hexapod (center of mass), and that because in the worst case scenario at least 3 legs will be on the ground at a time, the reaction force on each foot is a third of the entire weight of the robot? From there, how would I calculate the torques required for each joint on the hexapod’s legs?
The weight of the CH3-R is 4.60 (I’m assuming this is in kg?). My hexapod will most likely be no more than 3 kg, but it would be great if someone could help me make sure (with mathematical proof) that the HS645MG is indeed sufficient for the static and dynamic forces the robot is subjected to.
The servos used for most Lynxmotion hexapods were chosen largely based on what was available at the time (i.e iterative process rather than calculations). The calculations required a quite extensive: robotshop.com/blog/en/robot- … orial-3587
This does not include actual motion and covers only static / standing positions.
The weight of the CH3-R is actually in imperial units (lbs). As with all walking robots, the weight of the center is critical, as is the overall size so the lighter the better.
To give you an idea of how much additional weight the 645MG servos can support: lynxmotion.com/images/hi-res/j601.jpg
This is pretty much at their limit of operation.
I have tried many(!!) different servos, as I found the calculations could only get you so far. The servo that has come out on top for me, in terms of price point and performance is the Solar Servo D772 high voltage digital metal gear servo which runs about $22 and some change. It has an aluminum body and can be driven directly from a 2cell lipo battery. The torque 417(oz./in) @7.4v is quite nice. It is a coreless motor as well. Depending on sensors and load, I am able to run for 15min or so on the solar servo before giving them a rest. At that point they are pretty warm to the touch. The hexapod I have been working on is being built to be autonomous to a certain extent - closed outdoors course. Runs on an arduino mega (firmata), raspberry pi 2, win 10 IOT, GPS, MPU-9150, xbox wireless, 5 ultrasound sensors and WIFI.
It is kind of funny seeing plastic gear, plastic case, < 200oz/in torque servo being advertised as “ultra-torque”
Beware of Ebay, I have seen these servos on there for $50+!
Also check if the servo has some loose space when moving. I bought servos with this issue, i can turn the servo one degree or so when they are not moving. Also try to get “robot” servos. They have at one site the servo nob and on the opposite a bearing or something else. So you can attach the leg parts on two points, not only onto the servo nob. I found some good cheap servos at hobbyking.com
If you do get a servo without a bearing/shaft on the other side, I made the following part that can be 3d printed to replace the bottom plate of the servo to allow a bearing to mount.
