I think I finally figured out the best way to put the knee servos at the hip for the #fluffbug #robot. I think I should be able to pull this off with just the servo horns that come with the servos, some parts made of PCB and some paperclips. I'm still undecided on how to do the knee joint bearing, but right now I'm thinking of simply using a nylon bolt with a nylon washer.
I finally decided to try and make the #fluffbug #robot cosplay a little bit as the Flathead robot from the #cyberpunk2077 game. I laser-cut new legs that are more similar in shape to those in the game, together with an extra pair to be added on top. The cardboard "face" is temporary for now, as I designed a shield to put there, but am still waiting for the PCB.
I might add some greenstuff to make the legs more 3D and paint it later.
#fluffbug #robot #cyberpunk2077
I finally decided to try and make the #fluffbug robot cosplay a little bit as the Flathead robot from the #cyberpunk2077 game. I laser-cut new legs that are more similar in shape to those in the game, together with an extra pair to be added on top. The cardboard "face" is temporary for now, as I designed a shield to put there, but am still waiting for the PCB.
I might add some greenstuff to make the legs more 3D and paint it later.
More progress on the #fluffbug #robot with smart servos: I finally wrote a proper library for all the basic functions of the servos, and it even works. I also set the IDs of the servos, and fine-tuned their positions. Now I should just throw my walking code onto that, and see how that works, and see if I can improve it using the speed/time limit functions of the servos. We will see.
https://hackaday.io/project/190377-sly-bug/log/217800-servo-library
Thank you everyone at #hackadayberlin for a great experience. We have to do this more often! Thank you to the organizers who worked hard to make it all happen, and thank you to the participants who made it all great. It was great to finally put the faces to the nicks. There are now at least ten more #fluffbug robots out there in the wild, and soon maybe more, with all the people interested in building them. And I have lots of inspiration for new stuff.
All the parts are here and assembled. I copied some code from the big #fluffbug to see if everything works, and it's all good, except for one tiny annoyance: the servos come from two different batches, and some of them rotate in the opposite direction to others. I could fix that by swapping the leads on the motors and pots, but I recently added a Servo class to Fluffbug so I can reverse them in code. I just need to adapt the new code to the different pins and tune the trims on the servos.
I'm preparing to run some workshops on building the #fluffbug #robots, and one data point that I needed is how long it takes to assemble one. So I recorded myself assembling half of the robot (the other half is the same), to get some idea. Of course I will take that x4 for people without experience.
I'm preparing to running some workshops on building the #fluffbug #robots, and one data point that I needed is how long it takes to assemble one. So I recorded myself assembling half of the robot (the other half is the same), to get some idea. Of course I will take that x4 for people without experience.
I decided to also make a PCB for the leg parts. With #fluffbug I went with acrylic, because it's both cheaper and nicer, but this is a one-off build, I don't expect people to make it, especially since those servos are out of stock for a few years now, so I may as well just add this to an existing order. I got lucky that it got approved without an extra panelization penalization.
Of course as soon as I ordered the PCB with the coin cell battery holder on the back, I find out about the 14250 cell batteries and the exact same battery holder design that the big #fluffbug uses, only scaled down. I suppose I will be doing some bodging.
With a proper PCB, and the oscillator soldered to it instead of being connected with questionable bodge wires, the same 24MHz crystal suddenly works. Who would have thunk. That means I have a working camera shield for the #fluffbug in barely 4 iterations, though I think I will probably need to make one more version to move the resistors at the back out of the way of the fpc connector, if I want this to be widely used.
I was thinking about making a version of #fluffbug robot with a raspberry pi pico for a while now, and even tested it a bit, but now I decided to pull the trigger on it, especially with the pico w available, and the camera working. So here is #moobug, a bovine cousin of the fluffbug. I kept it pretty simple, just the battery management stuff (protection and charging) on board, plus servo connectors. Almost everything is SMD. And there is room to connect extra-wide shields – muzzles?
I think I have figured out what to do to actually be able to use the S2 Mini camera shield with the #fluffbug robot, despite not having any free PWM channels. I can provide the external clock for the camera by adding a 24MHz oscillator on the shield. Then I don't need the xclk pin, and the esp32-camera library doesn't need the PWM. So now I have three next steps: a pull request for the esp32-camera lib to make xclk optional, corresponding PR to #circuitpython, and version 4 of the shield.
And now the most recent #fluffbug is fixed and assembled. I think this is where I stop with the hardware, and try to focus on the software for a bit. I realized that even if I get the camera shield to work, I can't use it with this robot, because it uses at least one PWM channel, and I use all eight PWM channels for the servos. This makes me think about the picow, but if I got that route, it will be a new project probably. Name it #moobug or something. In either case I will need the software.