Home Creators Posts Import Register Favorites Logout
haven't archived this post yet. have a subscription? use the importer!

Content

The last, for now at least, miniDog video will be up for Patrons next Tuesday and go out publicly the week after that. I was intending to make it have more functions but having changed the foot profile a couple of times it's become apparent that there are multiple issues with the speed, power and accuracy of the servos I'm using.

The cost of an upgrade I'd be happy with is more than some pretty decent brushless motors to build openDogV2 with back-drivable 'quasi-direct drive' reductions.

miniDog v2 is staying in one piece because it ultimately works ok. I'd like to return to it later but I'll be designing a new robot dog which combines all the R&D I've done including my other brushless motor test dogs which I built last year.

The blue shoes pictured are just PLA - this was actually a good diagnostic tool because they are slippery, so I can see where the feet slip and where the errors in the gait are.

Files

Previews only

Comments

Anonymous

I still love your channel. Your enginering mindset (try, theorize, solve, build again). Would you be able to make a seperate (compilation) video of all the thing you learned? e.g. hip joint straight above feet, compliance with springs not required, servo's not accurate etc.

XRobots

Yes, the first 8 mins of the the first openDog V2 video is a complete recap of everything from all previous robot dogs!

John N Nelson

Keep up the great work. I haven't used trig for a very long time and I have a project I'd like to finish and seeing your programming source on this project will be a very big help. Very impressive work.

PatchBOTS

Great project James! I've really enjoyed this one

scott arendt

Really enjoyed Mini dog's :) Wondering if you would be willing to recommend a servo that would do it justice? Say something like the Dynamixel's series? If so, which would you think are the best option? Thanks!

XRobots

Yes probably if you want the absolute best. I was going to just try some standard rc sized 'brushless servos' at some point, but I have't really ever used anything better than digital rc servos in any projects.

scott arendt

Might be worth contacting your local supplier for Dynamixels, they're pretty cool little guys. Obviously expensive, but you might be able to get a sponsorship as your projects are certainly motivation for other people to go out and buy them to build their own robots :) I'm not sure if I stated my question well enough, but I was hoping you could suggest a specific set of servos as there are many different ones, even among the Dynamixel ones. Thank you again for the cool videos and projects!!

XRobots

Hi Scott, I'd really have to do some testing before recommending you buy 12 of them. Ideally you want the ones with lots of speed and power.

Anonymous

I've seen this video few minutes ago: https://youtu.be/utDagouxM5U . I really like your ideas whith all the robot dogs, but many problems that you were not able to solve were caused by joints... In my opinion you should try something like this. It surely is over-engineered, bit it might work. You can backdrive it, no backlash, very accurate. It is a very interesting option for the big robot dog (even if it is probably too much for the small one). I don't know how the elastic feature could be implemented in this design (an elastic rope is not the way to go: you won't be able to decide how much could it stretch). The big brushless motor of the big dog should be perfect! And, above all, over-engineered things are wonderful :-) . Thank you for all your ideas...and good luck for your projects ;-) !

XRobots

Yes I was just watching that video. It doesn't need elastic to be complaint - it is back-drivable as he mentions in the video, so that will make the motor act as a spring as per my test dogs: https://youtu.be/bHCzrDr-t3w . The plan for any new robot dogs will be very similar, but using a much higher torque motor, and the lowest reduction possible. This is how MIT cheetah works.

XRobots

Yes I've seen those. It probably weighs about 1Kg though, so it can make much faster moves with less inertia. Once we get over about 5Kg we'll need some compliance so make it more dynamic. miniDog is right on the edge of that, but anything bigger will need a Quasi-Direct Drive reduction like my test dogs: https://youtu.be/bHCzrDr-t3w

XRobots

Yep - I'll be doing something with brushless motors and Quasi-Direct Drive in my next build. But with an off the shelf motor and probably a 5:1 or so belt reduction. I should probably do a servo testing video or something, but even that is quite expensive to compare several models and test tolerances and accuracy between several of the same model.

scott arendt

I would really like that, but understand it's expensive! Does the UK have a goto robotics reseller? I like trossen, but its a US company and may not be able to sponsor you..?

XRobots

There are various sellers, generally if I went to a manufacturer direct or authorised reseller they'd want a specific product used in the project. It might be better to have a different sponsor for the video. I'm going to be doing a couple of ads for 'RAID: Shadow Legends', which is totally irrelevant but it helps out with the motor for openDog V2. So that's always an option too.