You're building an ODRI robot? Tell us about it!

[thomas.flayols]

Hi,

We are often asked how many builds this project motivated. We ourselves are quite curious.

If you assembled/bought an ODRI robot or just used part of the actuator design / electronic, we would love to hear about it!

Here are some ideas of what would be interesting to know:

• Who are you? Do you work for a lab / company?
• What did you build? Did you buy an assembled robot, a robot kit, or source the parts yourself?
• You can upload pictures here and links to projects or scientific publication.
• Of course, your experience and feedback would be very much appreciated.

At some point, we might add a section to the GitHub with content from this thread, if that’s OK with you.

[Roy]

Here’s my work so far on building a Bolt biped!

Who the heck am I?
I’m an undergraduate student who wanted to work with bipedal research and just learn as much as I can. In the past I’ve had experience with SLIP models and the Cassie robot from past research assistant experiences. I’m especially grateful for the amazing documentation from ODRI (really helps me not over-spend on my budget by avoiding mistakes). I’m also an one man & self-funded team so I’m doubly grateful for the responses and activity on this discourse!

Bolt Robot
For sourcing the parts I relied on the GitHub list for starters and then put together my own huge Notion spreadsheet for the exact vendors available to me. My Notion page is also were I keep notes on the project, like small modifications I made to the robot’s design, etc. I took advantage of some previous projects to reduce reuse and recycle as much as I can too to cut down costs (like the crappy stand I built).

Current Progress
Right now I’m still working on testing all the motor modules now that I’ve fully assembled the robot. I tested each individually but now I want to make sure they still function with the fully assembly.
I’m also reading up the papers associated with Bolt so I can better understand the underlying theory and software. As for electronics, I ordered a masterboard and micro-drivers from Seeed and also bought a cheaper IMU from Lord that still had built in EKF. Both orders are still backordered due to the supply chain issues, but I’m still really occupied with the motor module testing and paper/code reading!

[Yelk11]

This looks amazing! Im excited to see your progress, and one day see it work!

[Lhum]

Hi,

I am Elham. I am working with 4 other students at McGill University to build Solo12 under Prof. Hsiu-Chin Lin supervision. We found it difficult and time consuming to order every part by ourselves, so we decided to purchase the SOLO 12 KIT from PAL Robotics.

Since it was the first time we assembled such a robot, unfortunately we broke some parts, so we need to print them again. In addition, the microdriver arrives later due to the chip shortage. These reasons slightly delayed the process.

I will update this post when we build the Solo12.

I would like to thank @fgrimminger and @jviereck who helping us a lot with this project and solving our problems:)

[EtienneAr]

Hello,

I represent INRIA Paris research lab. We just finished assembling one Solo12 here.

To simplify the sourcing, we ordered two kits from PAL Robotics (one for immediate build and one for spare parts). We were overall quite satisfied by their service : They were a bit slow to replace some defective/missing parts in our kits, and they lacked a bit of communication, but they always did it without asking any question. And the kit also contains many parts that have already been pre-processed by them, so it saved a lot of built time for us.

Here is a picture of our robot :

The GitHub for the assembly was very exhaustive, but also thank you @fgrimminger, @thomas.flayols and @paLeziart for the help throughout building and testing !

Glad to join the community,
Etienne.

[laniakeawhite]

Hello,
i with my classmates,on process of building Bolt.
Here is the low leg assembled latest:

lower leg control.mp4

Many other things are on delivery, so now we spend most time to test communication,controll.
Later we will impelemets the dynamic model for Bolt,and try to contorl it using HZD.

By the way, isUS summer vacation now？ It seems very few people online.

[erich.hoo]

I have builded a Bolt hardware this month.

It’s not an easy job for me, but thanks the perfect documentation from ODRI , i have done it and only a few mistakes.

Right now I just use TI Evaluation Board Electronics to testing the motor modules. I have sended email to info.de@beta-layout.com for replicating the MicroDriver v2 boards a few days ago, but not got a repley yet.

I have a question about MicroDriver v3, does MicroDriver v3 avaiable for 6 dof bolt now?

and, thanks to the odri team, you are creating a very good community.

[yunifuchioka]

Hello!

I’m a master’s student at The University of British Columbia in Canada, supervised by Professor Michiel van de Panne. I built a Solo 8 Robot starting in Summer 2021, with another student to help with electronics fabrication earlier in the project.

Back then there was no option to buy a pre-built kit, so we sourced everything ourselves, which was a lot of work to do during pandemic lockdowns and chip shortages. We had to ask PCB manufacturers to DNP some components that they couldn’t source, which my teammate sourced and SMD soldered on himself!

We recently finished a research project using this robot: OPT-Mimic As mentioned briefly in the paper, the modular design was incredibly helpful for quickly repairing and replacing components, especially when testing high-impact motions controlled by experimental RL policies! I timed myself and it took me 10 minutes to replace a leg that I suspected was problematic, as opposed to having to send the robot back to the manufacturer for repairs as might be typical for other robots.

I hope to slowly give back to the wonderful efforts from the ODRI project and the community by sharing our experiences/advice with building the robot in North America. For starters, here are the suppliers that we used, in addition to some local/university-specific resources not listed below. We tried to keep them as local as possible:

• 3D printing: Xometry (SLS Nylon for legs, FDM ASA for body)
• Motor controller boards: Bitelle
• Masterboard: JLCPCB
• IMU: Hoskin Scientific
• Timing belts: Belting Online
• Bearings: EZO Bearings (US office)
• ODRI Encoder kit: PWB Encoders
• Motors: T-gravity
• Wires: Kabeltronik
• 2mm motor phase connector: HobbyKing
• Motor reamers: Misumi
• Misc. Electrical components: Digikey, Mouser, Newark, RS
• Misc. Mechanical components: McMaster-Carr

Many thanks to ODRI and the community,
Yuni Fuchioka

[amo]

Hi,

I’m from the Dynamic Locomotion Group, MPI-IS.

We built a Solo12 robot. We sourced the parts and assembled the robot ourselves. It is a bit time-consuming, though.

As you see from the picture, our robot is mounted on an optic table, ready for dissection.

In fact, we use our robot in an unconventional way, for a very different purpose.

We turn it upside down, fix its body (so it cannot run away), and connect the feet to a platform with ball joints. Basically, we have an over-actuated parallel robot/motion platform.

We do this on purpose. We need a high-acceleration 6-DoF motion simulator for another study. After extensive search on the market, we just can’t find any affordable option for our needs. Convinced by the amazing high jump of the solo robot, we decided to build our own motion platform on top of it.

12 motors working together on a light platform can achieve highly dynamic motion. We tested up to 10Hz vertical oscillation with an amplitude of +/- 10 mm. I believe I can push more, as the motor current is not saturated yet, besides it is ok to overpower it for a short duration. By the end of the day, I didn’t want to fix a broken robot, so I didn’t push more.

More about our project:
YouTube
GitHub
Paper

We have been using ORDI robots/modules for different projects in our lab. Many thanks to ODRI and the community, especially @fgrimminger who has always been very supportive.

An Mo

[easybuildtech]

Hello All,

After a long time in the making, I have recently completed an independent source and build of a Bolt.

I am a hobbyist and really appreciate the documentation and support of the community here.

I used an FDM printer for the body structure and am looking forward to testing how the material performs as well as possibly extending the design with the power management board and battery operation.

[rkourdis]

Hi there!

I’ve recently left my position at Ocado Technology working on robot arm simulation to build a Solo-12 full-time and learn about dynamical systems, control and optimization.

I started part-time work on the robot a while ago, but since I wanted to learn more about agile systems it made sense to take a few months and do this project properly!

---

Electronics

I’m soldering the electronics at home.

The PCBs are made by JLCPCB and I sourced the components from various distributors. After applying solder paste and placing the components, I use a hot plate to solder the boards:

---

3D Printing

I’m making the large 3D printed parts at home and have ordered the plastic pulleys from a Multijet printing service.

Initially, I bought 4x of the encoder kits from PWB which include the aluminium parts, but to save cost the rest of the metal pulleys were printed on an SLS machine (again, by an online service):

---

IMU

I’ve recently bought the IMU for the robot. The suggested model, 3DM-CX5-AHRS, is now old and has been replaced by 3DM-CV7-AHRS which has overall better characteristics.

The new IMU is also half the price than the previous one - £550 vs > £1200!

I haven’t yet figured out what modifications are required to make the new model work, but I can make a post with the details when I’ll get around to it.

---

Simulation

Recently I’ve integrated the robot in MuJoCo using high fidelity models exported from Solidworks:

I’ve spoken with the people at DeepMind and hopefully I’ll contribute the model to the MuJoCo Menagerie after doing system identification.

You can find the model here!

---

Current Progress

I now have completed one of the legs, all electronics / wiring apart from 2 uDrivers, and have all parts to complete the robot over the next few weeks!

After that I’m planning to focus on learning about agile robot behaviour using the Solo.

It would be great to collaborate with people that are working on these algorithms, so please reach out!

You can find me at rafael.kourdis@gmail.com or over on LinkedIn.

[erich.hoo]

Hello all,

Based on the documentation and community support, I recently built a Leg Test Stand V2. I made a slight change, instead of using the 6 axis force-torque-sensor, I chosed Omron’s E3Z-T61A as the ground contact signal switch. I chosed the AEDM-5810-Z12 encoder kit and didn’t modify the codewheel. because I used a slightly longer motor shaft, shortened the motor pulley AT3, and modified the cover of the upper leg module shell to fit the codewheel and encoder.

Now it can take off and jump at a height of about 40cm, from 26cm in the squat to 66cm in the highest position. Perhaps by optimizing the parameters, it can jump a little higher.
jumping test video

During the jumping test, the upper leg module hit something next to it for some anomalous reason, with high speed , and both times caused the AEDT-9810-Z00 encoder to be damaged, no longer usable, and a new encoder had to be reinstalled. The shell of the upper leg was good, the codewheel was good too, i had no idea why the encoder be damaged.