hey man, im loving the videos, this is my first one and i subbed instantly. just a tip, your videos would be alot better if you could fix the weird effect, kinda like autotune on your mic every once and a while keep it up!
The kinematic model can be done with different coordinate systems if you find the transforms to translate and rotate between the different coordinate systems.
Very good point, and it's basically what I did to get the leg axis to agree with the body axis. Though, the mental model is so much easier if you don't have this issue to start with.
@@scheffield yeah, it is best to keep it as simple as possible. For my hexapod I tried to work the leg kinematics out using a Denavit-Hartenburg (D-H) representation and managed to get an inverse kinematics equation for the joint between the body and coxa, but the coxa to femur and femur to tibia joint movements are linked, both joints together control the position of the foot within two axes which makes it very hard to use that method to get an equation for each joint angle so I had to instead find the transform into the coxa's coordinate system, including the first joint angle and then use trigonometry from there and it just made it so much easier. Interestingly for the body to coxa joint the equation came out the same as if I had just used trig from the start. The D-H representation makes the forward kinematics really easy but the inverse kinematics are a lot more difficult.
I love the project . You can run power straight to the servos from a power source and then only send the signal through the board to keep current down .Would this help ? Like all + and - to a battery and just signal going to board. these powerful servos can pull a good bit of amps each .i had trouble powering 5 though esc bec
Why bother with the buck converter at all? Most servos like that can be run straight from a 2S lipo, it would make it easier and make the PCB smaller. Edit: then you are limited to the max current draw of the lipo, not the max current draw of your buck converter. Servos and motors in general can draw a lot of power, especially when just starting moving, something that can be done with cheap servo driver boards like the PCA9685 is to stagger or offset the PWM signals, so that the signals aren’t sent to the motors at the exact same time and the motors aren’t starting to move at the exact same time, this might help reduce the peak current by spreading it out.
Love the staggering idea. Something to try out next time I get the bot out. As for the buck converter: I wanted to challenge myself. And I think I achieved that :)
Jeez your channel is underrated. It's really interesting watching you face challenges and how you end up overcoming them. Also nice editing!
Very nice. Winter is coming, and your dog needs TPU boots.
hey man, im loving the videos, this is my first one and i subbed instantly.
just a tip, your videos would be alot better if you could fix the weird effect, kinda like autotune on your mic every once and a while
keep it up!
The kinematic model can be done with different coordinate systems if you find the transforms to translate and rotate between the different coordinate systems.
Very good point, and it's basically what I did to get the leg axis to agree with the body axis. Though, the mental model is so much easier if you don't have this issue to start with.
@@scheffield yeah, it is best to keep it as simple as possible.
For my hexapod I tried to work the leg kinematics out using a Denavit-Hartenburg (D-H) representation and managed to get an inverse kinematics equation for the joint between the body and coxa, but the coxa to femur and femur to tibia joint movements are linked, both joints together control the position of the foot within two axes which makes it very hard to use that method to get an equation for each joint angle so I had to instead find the transform into the coxa's coordinate system, including the first joint angle and then use trigonometry from there and it just made it so much easier. Interestingly for the body to coxa joint the equation came out the same as if I had just used trig from the start.
The D-H representation makes the forward kinematics really easy but the inverse kinematics are a lot more difficult.
I love the project .
You can run power straight to the servos from a power source and then only send the signal through the board to keep current down .Would this help ? Like all + and - to a battery and just signal going to board.
these powerful servos can pull a good bit of amps each .i had trouble powering 5 though esc bec
Yes, it would help a lot, I think. However I chose to challenge myself. 😋
Why bother with the buck converter at all? Most servos like that can be run straight from a 2S lipo, it would make it easier and make the PCB smaller.
Edit: then you are limited to the max current draw of the lipo, not the max current draw of your buck converter. Servos and motors in general can draw a lot of power, especially when just starting moving, something that can be done with cheap servo driver boards like the PCA9685 is to stagger or offset the PWM signals, so that the signals aren’t sent to the motors at the exact same time and the motors aren’t starting to move at the exact same time, this might help reduce the peak current by spreading it out.
Love the staggering idea. Something to try out next time I get the bot out.
As for the buck converter: I wanted to challenge myself. And I think I achieved that :)