CARA robot dog

[Sergei G]

Printable, original design. Seems to perform well.

https://youtu.be/8s9TjRz01fo

Best Regards,

-- Sergei 

[Alan Timm]

What a neat application bldc + capstan!

I thought he looked familiar, he's the one that did that capstan video last year:

https://www.youtube.com/watch?v=MwIBTbumd1Q

I'll have to take a closer look.  Between this guy and Kristof and his Kind Robot Humanoid using captan (rope) drives there has to be something there.  :-)

Alan

[Chris Albertson]

I’m impressed, not because this is the best, but because he hit the cost vs. performance curve as a good point.

I’d say this has the minimum performance to be useful.   I mean that in a good way, I more powerful robo-dog would cost quite s lot more.  SO this might be the minimum cost, usable robodog.    I’m guessing the cost to be about $2K for this.  Which is pritty good.

I built a robot dog using R/C servos and I must admit they are very underpowered and too slow for dynamic stability, so a servo-dog must take short steps on flat ground.      His dog is just barly over the threshold of dynamic stability, which again is good if you are looking to keep costs down.

But if yoiu need a robot that can walk outdoors, climb up curbs and stairs, you need MUCH more power and speed and lower mass.    What I found with 3D printed plastic is that it’s strength-to-weight ratio is low

This guy is one of the few I have seen who really knows how to design with 3D printed plastic,  It is NOT like metal.  Metal is a solid, printed plastic is basically hollow, it is a skin with foam inside, or it SHOULD be.  Notice how thin=ck his parts are.  For best resolt you want thick skin over light infill and all the compound curves you can get,

But notice where his project stopped, teleoperate walking on a flat floor.   The next phase of the project is much harder.  An uneven floor requires sensors and 100X greater compute power because you have to plan footfalls and you can’t use open loop pre-programmed leg motions.  Once this is done, that you need the same localization and route planning and obstcal avoidance you would use on a wheeled platform

This is a great compromise project. It is exactly strong enough to serve as a software development platform for studying walking.  Mine was too weak and slow for that but only cost about $500.   His, I’m guessing ios at $2K.   But tfor serius perfromance you need to spend coster to $300 for each of thre 12 motors and controller.  More like mid four digits for the final robot.

Multiply those numbewrs of 4 or 5 for a humanoid as they need much greater power to weight ratio

I think he founbd a good compromise design.   He did exactly what engineers should be doing.

-- 
You received this message because you are subscribed to the Google Groups "RSSC-List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rssc-list+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/rssc-list/SJ2PR12MB7798E061A2BAD5F6CD6A0985C955A%40SJ2PR12MB7798.namprd12.prod.outlook.com.

[Thomas Messerschmidt]

I like it! But why does he stress over the 8:1 ratio? 

Thomas

On Jul 15, 2025, at 9:03 PM, Alan Timm <gest...@gmail.com> wrote:

What a neat application bldc + capstan!

--

You received this message because you are subscribed to the Google Groups "RSSC-List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rssc-list+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/rssc-list/e8fb4a35-a0c6-47de-8194-f10bba3ca932n%40googlegroups.com.

[Chris Albertson]

I wondered the same thing.   8:1 is a good compromise ratio but anything close to that would be as good

In fact one might even argue that a non-constant ratio, where the larger drum is elliptical, might posible y work better.

I also think his cord is too large. He does not need 2,000-pound test line.    I would use liter cord but a winding pattern that got between the two drums many more times so there are 6 or 8 connections between the drums and not just two.

Finally, does the cord need to be non-stretch?   Maybe a small amount of elasticity would be best

But his goal seems to be teaching, not a commercial product.   YouTube is the product so his design is very easy to see visually.  He’s very good at storytelling and visual design

To view this discussion visit https://groups.google.com/d/msgid/rssc-list/8DD028CC-0264-411F-826B-ECEC09F2FD98%40gmail.com.

[Thomas Messerschmidt]

A small amount of flexibility would help prevent robot breakage and help prevent human injuries. 

Thomas

On Jul 16, 2025, at 1:25 PM, Chris Albertson <alberts...@gmail.com> wrote:

I wondered the same thing.   8:1 is a good compromise ratio but anything close to that would be as good