@@WilsonFabrication Water is all you need. Ideally something that will drip continuously on the cutting area, else you have to stop and reapply frequently. For making holes, a dremel and set of cheap diamond bits works very well. For that I keep the part almost entirely submerged in water.
This lad looks like it'll be quite big when finished, I look forward to seeing it. Always interesting to see hefty ones that can stomp around the house scaring the burglars, rather than just those little table-top robots that are so popular.
It's not all about backdrivability, but inertias!! The fact that the joints are technically backdrivable does not necessarily mean dynamic motions can be easily achieved. I suggest you watch the interview "The World's Safest Humanoid Robots | Eric Jang, 1X Technologies" to learn a little bit more about the implications. Love the rolling joint design!
Just goes to show that there’s always more for me to learn! Thank you, I’ll definitely take a look. As for inertia, it really doesn’t feel like the leg has substantially more inertia than given by the mass of the leg alone
@@WilsonFabrication Totally, I'm still learning everyday! I suggest you dig a little into reflected inertias, you'll find the relevance of the type of motors and reduction ratios.
Before I thought the motors were the wrong choice but I think I was wrong. Those DC motors could work pretty well with some good drivers. Im used to bargain h-bridges but if you have current sensing you could make something much smarter than just a servo. The design looks really stiff and durable too.
Honestly until it was all assembled I was quite worried about the motors too. The new motor controller boards are going to be a game changer. They’re so feature filled that I will probably make a video showing them off and how they can be used for a TON of other things than just driving motors
What kind of forces do you think it can take? I have an 80(ish) pound tripod dog (back left leg, still got most of the meaty thigh).. I've been thinking of building a prosthetic (not that he seems to need one). I wonder if I could fix that knee joint to work with gas springs.
The joint is really strong, way stronger than I expected it to be even. I bet I could put my weight on it if I had a good way to lock its position. The motors on the other hand seem to stall at 25lbs of force, but if you used gas struts you could totally get more force out of it. The CAD modes are up on my website if you want to take a crack
This is a very promising looking project. I think the only problem I may see with it in the future, is the foot. Although it is difficult to tell from the video, it looks like it could be a fragile part, which may not last well in walking conditions, especially outdoors. Also, wouldn't the backdriveability of a motor also depend on whether the robot is on or not? I'm not sure if your demonstration was representative of real conditions the actuator would be in
I agree with both your points. I have a more standard looking "nub foot" designed and printed which I will do the majority of the testing with, but the compliant paw was too cool to pass up. And yes the motors will act differently once powered up, but I think it will depend more on how the motors controller is tuned. Right now I have them on a PID loop, so the farther from their target the more current will be given. This should have them act like a spring. The PID loop can be tuned for positional holding or for reaching their target quickly, or have two stages of PID values to get the best of both worlds. I could also set a deadzone around the target angle where the H bridge motor driver shorts the motors causing them to be very difficult to back drive. Programming isn't my specialty so we will have to see how it goes
Not enough testing to ensure it'll hold. I plan on replacing the main hip axis with a 1in steel tube for the needle bearing to ride on instead of the carbon
Thank for giving the people what they want: a build montage. And thanks for posting your L’s as well as your W’s. That’s honest and useful. And that’s engineering, baby! 🦾
This is something I have definitely been worried about. I will likely replace that carbon tube with a 1 inch steel pipe for durability. Luckily that is a very easy change
I'm actually ditching the pro mini and replacing it with an esp32! And just like you recommended I'm using the castellated edges to solder the esp directly to the board. I'm planning on using the esp32-c3 due to its cost but the footprint of the board is shared with its more powerful siblings so if I need more horsepower I can easily swap it out
you have no idea. Timeline gets so long and overwhelming to rebuild that I export as step then reupload and keep going to clear out the timeline. This is probably the 5th timeline restart
Great progress so far, careful about the carbonfiber dust it can be dangerous
Thank you! I did wear a mask but I should really start wearing gloves and wetting the tube before cutting next time
@@WilsonFabrication Water is all you need. Ideally something that will drip continuously on the cutting area, else you have to stop and reapply frequently. For making holes, a dremel and set of cheap diamond bits works very well. For that I keep the part almost entirely submerged in water.
I LOVE that you shared the failure! I'd enjoy more details about WHY the regulator failed and what you are doing to avoid that issue in the future.
This lad looks like it'll be quite big when finished, I look forward to seeing it. Always interesting to see hefty ones that can stomp around the house scaring the burglars, rather than just those little table-top robots that are so popular.
Keep it simple, then upgrade is the best decision! I am excited about your project - rolling contact joints are the way to go for legged robots
Look up smoke-stoppers; many people that assemble their own drones use them to avoid frying electronics when plugging the battery in the first time
It's not all about backdrivability, but inertias!! The fact that the joints are technically backdrivable does not necessarily mean dynamic motions can be easily achieved. I suggest you watch the interview "The World's Safest Humanoid Robots | Eric Jang, 1X Technologies" to learn a little bit more about the implications.
Love the rolling joint design!
Just goes to show that there’s always more for me to learn! Thank you, I’ll definitely take a look. As for inertia, it really doesn’t feel like the leg has substantially more inertia than given by the mass of the leg alone
@@WilsonFabrication Totally, I'm still learning everyday! I suggest you dig a little into reflected inertias, you'll find the relevance of the type of motors and reduction ratios.
Good luck, I can't wait to see where this goes.
You should believe in a current limited bench power supply. This helps keep the magic smoke inside.
Before I thought the motors were the wrong choice but I think I was wrong. Those DC motors could work pretty well with some good drivers. Im used to bargain h-bridges but if you have current sensing you could make something much smarter than just a servo. The design looks really stiff and durable too.
Honestly until it was all assembled I was quite worried about the motors too. The new motor controller boards are going to be a game changer. They’re so feature filled that I will probably make a video showing them off and how they can be used for a TON of other things than just driving motors
4:07 "and for you, i only have one thing to say"
Wilson: proceeds to prove by moving parts.
Those with something to say: Now we have nothing to say😂😂
someone buy this man a silicone mat
Great video would like to see if the paw would work make it more animal like when walking could be a possibility of opening the doors to other areas.
What kind of forces do you think it can take? I have an 80(ish) pound tripod dog (back left leg, still got most of the meaty thigh).. I've been thinking of building a prosthetic (not that he seems to need one). I wonder if I could fix that knee joint to work with gas springs.
The joint is really strong, way stronger than I expected it to be even. I bet I could put my weight on it if I had a good way to lock its position. The motors on the other hand seem to stall at 25lbs of force, but if you used gas struts you could totally get more force out of it. The CAD modes are up on my website if you want to take a crack
I love this!! Can’t wait to see the next video
Engineering videos this make my happy in a way i cant explain❤
nice job!
COOL SIR. NICE LEG
Impressive!
This is a very promising looking project. I think the only problem I may see with it in the future, is the foot. Although it is difficult to tell from the video, it looks like it could be a fragile part, which may not last well in walking conditions, especially outdoors. Also, wouldn't the backdriveability of a motor also depend on whether the robot is on or not? I'm not sure if your demonstration was representative of real conditions the actuator would be in
I agree with both your points. I have a more standard looking "nub foot" designed and printed which I will do the majority of the testing with, but the compliant paw was too cool to pass up.
And yes the motors will act differently once powered up, but I think it will depend more on how the motors controller is tuned. Right now I have them on a PID loop, so the farther from their target the more current will be given. This should have them act like a spring. The PID loop can be tuned for positional holding or for reaching their target quickly, or have two stages of PID values to get the best of both worlds. I could also set a deadzone around the target angle where the H bridge motor driver shorts the motors causing them to be very difficult to back drive. Programming isn't my specialty so we will have to see how it goes
Shoes! Let it wear shoes!
Are those needle roller bearings on carbon fiber? How does that hold up?
Not enough testing to ensure it'll hold. I plan on replacing the main hip axis with a 1in steel tube for the needle bearing to ride on instead of the carbon
Thank for giving the people what they want: a build montage. And thanks for posting your L’s as well as your W’s. That’s honest and useful. And that’s engineering, baby! 🦾
Using needle roller bearings right on carbon tube like that will fail pretty quickly with any load or use.
This is something I have definitely been worried about. I will likely replace that carbon tube with a 1 inch steel pipe for durability. Luckily that is a very easy change
@@WilsonFabrication For weight, use aluminum tube. Polish it and use PTFE sheet for bearings. Rolling element bearings are only needed for high speed.
Great video
Thanks! I tried a number of new things this time and I think turned out well
this is what my generation would call a camel
I don't really understand the Arduino Pro Mini. Why not just solder the MCU directly to the board? Btw. you might want to use some more modern MCU-s.
I'm actually ditching the pro mini and replacing it with an esp32! And just like you recommended I'm using the castellated edges to solder the esp directly to the board. I'm planning on using the esp32-c3 due to its cost but the footprint of the board is shared with its more powerful siblings so if I need more horsepower I can easily swap it out
This is great!! Keep it up!! 🎉🎉😂
Thank you! Will do!
0:54 thepress inserts mid build 😨
🎉🎉
wow fusion timeline scary
you have no idea. Timeline gets so long and overwhelming to rebuild that I export as step then reupload and keep going to clear out the timeline. This is probably the 5th timeline restart
Liked, subbed, and algorithm stimulated! 8^D
Thank you! Genuinely appreciate it
2:39 I….. uh. NM.
Ha ha! Debug time
Hello, we are a professional brushless motor reducer motor factory
Can Bus lolll