XL_Quad little 3D printed quadruped robot


I have been talking about the USB2AX Pro, DXLD dynamixel device and SSSlidar (Small Solid State Lidar) that I desigend recently, and here is a robot that will use them all.

3DOF legs, 12 servos, RPi3 + USB2AX Pro.

Full album, including all the details of the development process : https://goo.gl/photos/okZSp7jzWNRkBvD88

It’s up to 21cm long (body is 18cm long but the legs can protrude on both front and back), 12cm wide and 13 to 19 cm high.

It will be 480 - 500g when all the missing stuff (1200mAh 2S1P battery, DXLD, USB2AX Pro, sensors…) is added.

It has potentiometers and torsion spring in the joints, and uses them to get a sense of the torque exerted by each leg servo.


Impressive. I’m looking forward to the progress of this robot. You will need an extra board for the ADCs that will read the potentiometers, or are you planning to add that to the DXL board?

The DXLD Dynamixel device that I talked about in another thread will do that. I will put one on each leg, to read the output from two potentiometers.

Hi! It’s a VERY interesting project, what amazing thing… Could you post a video of it while moving? (even if you’re just testing it).

I would ask you so much things about your project, I am starting a similar project around a biped basis with a friend and we would know for example how are you driving the XL 320 from Raspberry? Is it hard? Efficient? Stable?

Do you plan to use pypot? Do you plan to document your project, do you want to see a community around it? We are very excited, here, in Normandy!


Thanks :slight_smile:
There is currently one video of the bot getting up, and it’s the last thing added to the album linked to in the first message (I update it in real time).

Sadly, it shows that my little quad cannot jump as-is. Since I haven’t monitored the movement, I can’t affirm if that’s due to too much weight, or not enough speed (=legs too small). At first glance I would think the second. We’ll have to see when the whole thing actually is driven (right now that was just using the Dynamixel Wizard) properly to see if it can accumulate some energy in the spring and use it properly.

I haven’t tested anything with the Pi3 yet, and since the servo are going to be driven through the USB2AX Pro, things might be different from what others have experienced.

I might document the project, what is about sure is that I will answer questions. Don’t hold your breath for me to release stuff by myself, I’ll be very busy in the comming monthes…

I have no plan on fostering a community by myself around it, it’s not the same level as building a Poppy : too many custom metal parts, hand made in my case though they could be done with a lathe and mill; and too much post-processing of the parts (including fine adjustement of the tibia for perfect mating, revork of the surface of some parts with a soldering iron, and soldering of the plates of the chassis…).
The work Matthieu did on Poppy to make it replicable by other people is extensive even though it is not necessarily evident to everyone.
I will not put that effort in to get there myself. I have already spent multiple hundreds of hours on it, and will spend again a long time on firmware/ software just to get it working the way I want…