New Mini Humanoid : Poppyrate!

Hi !

We are pleased to introduce to the Poppy Community and to the Flower Team our new poppy-based project: Poppyrate! Our Poppyrate is a 25 DOF robot, motorized with 23 XL-320 and 2 AX-12A. Our purpose is the same as the Poppy project’s one: Make accessible robotics through an Open-Source, 3D Printable, Transportable and Inexpensive robot. But unlike most projects and the INRIA, we are ZeCloud Robotics a little company of two employees. To reconcile our two opposite objectives, profit and popularization, we have decided to find different funding plans (notably in the supply source) than the users and sell Poppyrate kits at general public price!

The mechanics design is almost over with a new poppy-like skeleton of 45 cm, adjusted for the new motors’ dimensions, sized for the new torques capacities and 3D printable on the most affordable printers. The electronics cards conception is over and the cards are currently being built. They will adapt the voltage and the intensity to the different motors’ needs. Finally we hope we will get the full programming libraries of Pypot and offer the first beta version as soon as possible.

For deeper Information you can follow us on our Website. You also can follow us on Twitter and Facebook.



I don’t want to get you down, but a 45cm robot with XL-320 (that can only produce around 0.8N.m of continuous torque without overheating)… good luck.

It is a great project but I really advise you to play with a Darwin-Mini.

Your project is to build a Darwin classical (45cm height) with motors of Darwin-mini.
I am convinced in passive robotics but the issue is not the torque as said Xevel, the issue is the global robustness of XL-320 (3.5 times less powerful)

But what can be interesting is building a mini-Poppy-Torso like this. A low cost Poppy-Torso could be really a good programming platform !!

Hi @poppyrate and welcome to the Poppy family!

In your robot you have XL-320 and AX-12A how do you manage the difference between them? You have 2 different voltage sources or DC/DC? You have 2 different serial networks or only one?..

@Xevel @Nicolas @Thot

Thanks for your support !

For the Mechanics:
Don’t worry we made a lot of sizing calculus and we are tring to explain you our reasoning :
Torque = Force * Length, here Force = Mass
PoppyrateTorque = PoppyrateMass * PoppyrateLength but the PoppyrateMass = 0.4PoppyMass
PoppyrateTorque = 0.4PoppyMass * PoppyrateLength but PoppyrateLength = 0.5PoppyLength
PoppyrateTorque = 0.4 * 0.5 *PoppyMass PoppyLength
PoppyrateTorque = 0.2
PoppyTorque. So we can replace the MX-28T (0.25 kgf.m) by the XL-320 (0.04 kgf.m) and the MX-64 (0.6 kgf.m) by the AX-12 (0.15 kgf.m) and save a lot of money :blush:

For the Electric Supply :
We use one 12 V electric supply and a few bucks that convert the voltage for the XL-320. But we are writting an article on our website to explain this deeper with schemes :wink:

Welcome to Poppy Humanoid’s little brother !

Well… I guess we’ll see.
Just keep in mind that XL-320 have comparatively more backlash than MX, have plastic gears, and are less rigid all around.
AX use a position controller that is basically a glorified P control loop, quite different when it comes to keeping a position from the full PID of MX and XL.
And the torque values you refer to are stall torque, using more than about 1/5th of that continuously without additional heat dissipation risks overheating.

Aside from that:
How do you plan to attach the XL320 to your 3D printed pieces? The Ollo rivets have way too much play to be used in such a big robot.
What’s you plan in terms of controller board? And battery ?


We try having a good price-quality ratio so we have to make some concessions and use the XL-320 with its advantages and its withdraws. The industrial plastic XL-320 are made of, satifies us, is sufficient for our needs and is completed by the rigidity of the carbon skeleton. They are set with mechanical stops on the skeleton and without almost no fastening. We only use a few OLLO rivet for the XL-320 rear horns but it’s the only way to easily anchor them.

The project has just begun and the main difficulty we must face is to coordinate the AX-12 and the XL-320. So some changes into the pypot librairies (just like the controller protocol you mentioned) will have to be done to match the code with the new motors specifications.

The factor of safety is the same as the Poppy Flower Team chose. Thus even the biggest torque will be 4 times lower than the stall torque.

We plan to put batteries but it’s not the priority and just like the Poppy we use an Odroid Card. If you want to see more about the electronic architecture you can refer to this article.

Thanks for your warm welcome. We used to meet at events with another robot and I hope I will see you again soon with a Poppyrate :grinning:

Hi @poppyrate !

Great project ! I am sure this will be both very useful for many users and set inspiration for other creatures.
Like @Thot I would advise to seriously consider a torso version of your creature, which will be both cheaper and more robust, and suited to many uses.

Why not using this regulator ?

1 Like

Because the regulator would be oversized, a capacity of 2.5A for 4 XL-320 is a lot for their needs but I am sure it can work !

yep but it costs 10 bucks and you are sure it will work

Hi @Oudeyer !

Thanks for your cheering :smile:
We begin this project with all the skeleton. Because obviously it’s cool to have legs. But the final goal is to have cheaper version of Poppy so as making a torso version seems easier, after the first complete prototype version of Poppyrate. we’ll probably make a torso version too.
A nice feature could be to have a system to easily add legs sort of pluggable system. The difficulty with that kind of thing will be to connect all the wire and electronics to the lower body.

Geeks in a garage

For the moment we use this it is convenient because it has a integrated voltmeter but it’s too big so we decide to make our own small regulator board. It’s not that easy so maybe we’ll try a smaller solution afterwards like the polulu board.

Welcome @poppyrate and good luck for your ongoing development. Looks like a promising platform !

@poppyrate, I saw your videos on twitter, and I just wanted to say that is really a great project. For servomotor strength, the way to find the good model is not so mechanical… I do not know yet how to do this…
Congratulations and go on. Do you have an estimation of the total cost of a poppy rate ?

Hi Thot,

They have a website and it shows an estimate. Great project.

Here is a video reference to Frankenstein


Hello Everyone,
@Yassine @edil @Thot Thanks for your cheering and for relaying our website and video.
It’s been a long time since we have given some news, the project has slowed a bit in september. We were two behind this alias this summer. Paul and me Aymeric. At the end of august Paul has started a new internship in Montréal and i needed to work for my clients.
Obviously we’re continuing the project, its’ just a temporary slowdown.
And to give you some fresh news of the first move of Poppyrate I wrote a new post. :smile:

By the way i just discovered the Raspoppyno project and it could be a great addon to the poppy creatures.
Un petit message en français pour préciser qu’on est français mais que pour une grande diffusion sur le forum on avait choisi de s’exprimer en anglais sur ce post.

1 Like