Some legged robots end up moving with ponderous deliberation, or wavering in unstable-looking jerks. A few unfortunates manage to do both at once. [MusaW]’s 3D Printed Quadruped Robot, on the other hand, moves in rapid motions that manage to look sharp and insect-like instead of unstable. Based on an earlier design he made for a 3D printable quadruped frame, [MusaW] has now released this step-by-step guide for building your own version. All that’s needed is the STL files and roughly $50 in parts from the usual Chinese resellers to have the makings of a great weekend project.
The robot uses twelve SG90 servos and an Arduino nano with a servo driver board to control them all, but there’s one additional feature: Wi-Fi control is provided thanks to a Wemos D1 Mini (which uses an ESP-8266EX) acting as a wireless access point to serve up a simple web interface through which the robot can be controlled with any web browser.
Embedded below is a brief video. The first half is assembly, and the second half demonstrates the robot’s fast, sharp movements.
We love it when robots show some personality, like this adorable little quadruped robot that can make small jumps.
Thanks to [Baldpower] for the tip!
That actually looks a bit freaky – almost lifelike
Kicks up a phobia, doesn’t it?
Yes kind of cool. Now I want one large enough to ride!
It seems snappy simply because the video is sped up. SG90 servos can’t normally move that fast.
Reduce playback speed to .5 seems much more plausible.
Agreed, The rocks being kicked up seem much more plausible at that speed. It still looks pretty awesome at that speed.
It honestly had not occurred to me that there might be something else (something not part of the build instructions, I mean) responsible for how good the motions look. I suppose one way to find out would be to build the thing and see for myself, and another would be to ask the author. I think I’ll try both.
You can find another video here with fingers that shows that it’s at normal speed : https://www.youtube.com/watch?v=U6wg6iCAS0g
Nope. Watch how fast his thumb is tapping the icons at about 1:08.
Looks like normal speed to me, care to elaborate ?
Why does he uses both an Arduino Nano AND an ESP8266? Looks like the Arduino is just used for I2C and uses just 2 pins (SCL and SDA), which ESP8266 can also manage, no?
Because he has no idea what he is doing, that’s why.
To me it seems he know exactly what he’s doing while you are just standard Internet trash. Shut up or put up – show us your design.
Here you go:
https://hackaday.io/project/3300-kubik-quadruped-robot
https://hackaday.io/project/3301-pico-kubik-quadruped-robot
https://hackaday.io/project/6050-tote
https://hackaday.io/project/6321-katka-a-mammalian-robot
https://hackaday.io/project/7168-logicoma-kun
https://hackaday.io/project/9065-tote-zero
https://hackaday.io/project/9927-tote-had
https://hackaday.io/project/19603-d1-mini-tote
https://hackaday.io/project/19921-spiderwing
https://hackaday.io/project/158981-kubik-m0
I’ve been showing my designs for the last 5 years. You just have to look.
Now that sounds a bit harsh.
Based on the video I would not say that! And I must say that the video looks very explanatory regarding the construction, so regarding that it is a job well done. Sure the ESP8266 could easily replace the arduino but I’m sure that there is a very easy.logical explanation for the fact that he didn’t choose that path.
When we see something we are always judging from “hey, why didn’t you do that”, a pitfall of commenters around the world. Yet what the commenters don’t know is that a project evolves. It starts with one thing and that is being expanded into something else. Creating strange combinations of parts that sometimes could have been easily combined. Yet those insights always come at a moment in time when the project has taken to long and the builder is no longer interested in doing a complete redesign in order to save one part. I think we’ve all been there.
I think this is a very nice build, thanks for posting.
The movements looks very nice, even if the video is sped up, it does look very natural, speed or no speed, the movement are excellent.
I think, maybe, he wanted to mantain the two basic functions separate: inverse cinematic servo control on one side, and communications on another side. It is true that the esp8266 can perform pretty well with all robot functions, but is still cheap use two processors for dedicate tasks.
Those who can’t do criticize. Very cool design!!
Sure it’s not “those who can’t do…” become writers? Or is that “those who can’t do…”become coaches?
Hey Deshipu,
I’ve been a fan of your work for awhile. You’re good mix of artistic and hardware pro. The fact we both seem to prefer SMD makes you nifty in my book.
But there are many reasons for redundancies. Most likely, the builder started with a PCA9685 and an AVR board, then, at some point thought adding WiFi ability would be great. Often, a hacker’s time is fairly precious–I’ve been in the position where I must decide, “Do I go back and do this best? Or continue with redundancies to a finished product?” I’ve learned to accept imperfections a part of completion. In fact, as I’ve moved into the world of professional software development I’ve learned–perfection is the enemy of complete.
As an important citizen of the hardware community, you might take notice of Linus Torvalds personal realization:
“…My flippant attacks in emails have been
both unprofessional and uncalled for. Especially at times when I made
it personal. In my quest for a better patch, this made sense to me.
I know now this was not OK and I am truly sorry.
The above is basically a long-winded way to get to the somewhat
painful personal admission that hey, I need to change some of my
behavior, and I want to apologize to the people that my personal
behavior hurt and possibly drove away from kernel development
entirely.”
https://lkml.org/lkml/2018/9/16/167
He’s a pretty smart guy. And so are you.
–Ladvien
It is a very impressive build, and it’s also very well programmed. It’s also much better than anything I have built myself, even after so many attempts.
I was merely referring to this one part of the design. Whatever the reason, whether it’s the result of lack of knowledge or a vestige of previous design, it is a design mistake that adds to the cost of the robot, to its weight, and complicates debugging.
ESP8266 is a great SoC but it does not have the hardware timing the Nano can provide that can make a rock-solid PWM generation for 12 servos.
“But For this project i choose Arduino Nano, because don’t need much of pin i used, it’s small and don’t need FTDI to program it.”
Although $22 may seem a bit much for what’s being used. The PCA9685 does the heavy-lifting as it were.
Someone needs to check the suggested 50$ BOM. Pretty close but unless filament and printing free somewhere.. Shipping … Othewise pretty cool. I use Amazon to cut down some BS and slow boat. Vid does seem to be sped up a bit. At least none of my cheapo SG90 run that fast especially under any load. Ignore cranks about using ESP8266 and NANO and 9685. I do it as well.
Anyone know the software used in the video to do the 3d work (movement and pin placental etc)?
Possibly Solid Works. I’ve seen kinematic demonstrations with that before.
Something similar to this could be a good platform for a backyard weed hunter robot. Bluegrass only…be gone crab grass and dandilions!!
It would be difficult to weather-proof it.
Someone should TOTALLY add quad rotors to make this fly! Strange weird spider beetle thingie that crawls out from under somewhere and lifts off ????