Milli (#9) - New Mount for Helix - and Crawling Video!

 Milli - a Bio/Math inspired crawling single actuator robot millipede - now with teeth!

Milli - Millipede robot - now with teeth

From the last post, you may have noticed that I had tried to print some "nodes" as a possible improvement to Milli the millipede. Well, that thought was not as successful as I had hoped. There turned out to be no flexibility and too much tension on the helix. With this in mind, I reopened the milli design and set to work on some design updates:

• Offset the motor
• add a gear reduction to the motor
• add a bearing to main helix mount - to remove force on the motor itself
• try out a new power source
• Add teeth - because we need to add more Nightmare Fuel!!

CAD Design - new Motor Mount and Gear

The new design is a new motor mount, with drive gear, a bearing mounted main gear, and new 12V "shore" power.

3D printing gears remains a dark art to me, but here you can see that they are functional. In this print, I have used a larger Modul of 1.5, this makes larger teeth. I believe for my quick 3D prints this is a more accommodating solution (more tolerance for rough prints!).

I applied a liberal amount of hot glue to attach the pre-existing helix mount to the newly printed main gear. That may qualify this for a frankenstein prototype, no doubt.

Sloppy gears? certainly room for improvement - on the next iteration.  The thicker base layer print creates an uneven gear surface.

Another issue was the 'pede wheels' that I kept from the previous build. They function as wheels, however in the video you can see that the left wheel consistently unwinds and falls off, Doh!

Ready to Test!  Let's get crawling!

Milli - Robot Millipede on a colorful background

For this round of testing, I am using "shore power". The robot is tethered to a 12V DC source. I reused a momentary switch - also from a printer - in order to easily turn on and off the robot.

Testing goes as expected.  The motor gear hold, which is surprising.  However the front left pede-wheel falls off frequently.  The motion of the helix, continues to be smooth.  I believe the addition of the bearing smooth this rotation off, taking load off the motor.  

Video on YouTube:

Next Steps:

With this success, time to think of what would be in the next iteration.

• Onboard 12V power supply
• micro-controller with PWM output
• Motor Driver board
• remote or autonomous operation (Min1 project may come first)

Current Thinking in the Robot Labs

These are things currently being thought about at the TENRA Robot Labs...

Late nights and early mornings in the lab!

3D Printing

Ah... the beauty of 3D printing for robots!   It is the magic of having an idea in your head, which leads to, a sketch on paper, a drawing in CAD, and then a solid print - or many prints:

1st - A quick sketch on paper

2nd - A quick draft in OpenSCAD

3rd - a quick print - a solid Object!

Finally - Repeat as needed

I really love this aspect of Rapid Prototyping. I find it one of the more enjoyable steps in robot building. At each step in the process, you can find opportunities or issues with your design. It is fantastically easy to go back a step or two and add improvement to the design.

A good example is in this very simple build of "Node", pictured above.  "The Node" is an experiment in the Milli Project. I wanted to explore the ability to move in two directions on the helix in order to improve the mobility of the millipede robot. 

At each step in the process, I found that new aspects of the build became evident.  Even in the first print, i realized that I need more of the 'feet' to touch the printer bed. A quick change to parameters in the CAD, and it solved the problem. Finally, a quick print of three copies of the nodes. I assembled them and was able to perform my tests. The result was that the solution would not work as is due to certain bindings. I was having a difficult time imagining this complex relation in my head.  The rapid development of the concept provides a fast path to success or failure, both are acceptable!

Min1 - what should it do at night? 

Min1 is the Minimal Viable Robot for the Totally Not Evil Robot Army.  But, I am just thinking, what can this robot do at night.  It is just sitting around, with nothing to do.  Idle cycles are the devil's playground, right? So perhaps it needs a few teeth and a sensor to seek things (like ankles) out and make stabby motions.  ;)

Min1 - outfitted with ultrasonic sensors and stabby teeth

Min1 ready to Roam - seek and stab

Min1 Robot - as a Franken-prototype ready to go

Tilt!

Tilt, the balancing - 2 wheel Robot, is sitting prominently in the TENRA Robot Labs. It is patiently waiting a more powerful power source.  Twelve Volts DC is not going to cut it, it needs more power to keep the recycled printer motors from stalling.  I have just recycled a 19V power adapter that I will use as 'shore power' for the next round of testing. 

I am really liking Tilts new 'eyes'!

Robot RoundUp - Mid 2020 Edition

Here are some interesting finds from the internet about Robots of today:

The Totally Not Evil Robot Revolution has Arrived

National Geographic Magazine has an excellent feature read on how robots are entering out lives. It covers many different aspects of the social impacts of robots in work and life. It is a very nice read:  The Robot Revolution has Arrived

BionicSwift

Festo continues to amaze with their bio inspired everything!-- in this case a bio mimicked bird, the BionicSwift!

And their Bionic Mobile Assistant:

These robots are starting to look more and more like the helper bots of the future. Tools that assist. Perhaps Festo is becoming a better 'Boston'?

Practicality / Pragmatism:

Continuing on the pragmatic path. Stretch by Hello Robot is also ready to enter the domestic robot market.

Stretch by Hello Robot

And on the not so Totally Not Evil side...  Autonomous Warfare.  :(  Our humanities advancement in the ability to automate everything will always be driven (it seems) by the need to build more effective machines for warfare.  Here is a good video summary by Digital Trends:

Back to the safety net of the economy! Here is a more commercial perspective, In this view point from Forbes Magazine on Where Robotics Are Heading

Finally back to this! - THE SNATCHER!

A team at  the Seoul National University of Science & Technology has developed an ultrafast biomimicry Chameleon Tongue! (article in IEEE) This is a great vision on how robots can snatch packages for robot delivery of our future.  of course in the Totally Not Evil Robot world, it could be used to enhance porch piracy! (gads!)

How could you NOT want a Robot with a 30" Tongue?!?

Tilt! - A Balancing Robot (#6) - Prototype Construction Complete

The planned 3D printing is complete, here is a view of the assembled Prototype of the Tilt! Self Balancing Robot.

Tilt! -3D Printed - Recycled Printer - Self Balancing Robot (prototype)

Now onto the wiring and testing!

Tilt! - A Balancing Robot (#5) - Initial wiring and software for testing

The first leg is completed and structurally sound enough to start testing the wiring and software needed for control.  (hmmmn... did I say leg?)

Tilt! Balancing Robot - first prototype leg

over the past weekend, I found many many videos on YouTube for Balancing Robots. There seem to be dozens of them!  The majority of these are copies of other projects, just re-implemented. Even less of these videos support code to help in the own development.  And with inspection of these codes, the majority are copies of the same code. It makes me wonder, who is the originator of this code base that is getting copied so many times?

Cargo Programmer

A Cargo Programmer is a person that copies code and scripts off the Internet and tries to build a system without fully understanding it. In 2020, this seems to be massively popular. Sites like StackExchange and the many Instructables and Git repos make this almost too easy.  The biggest concern is that these 'cargo programmers' do not understand the logic and activity behind the code.

The origination of the term is from the concept of Cargo Cults. Now we see that same concept being applied to 'magic technology' that is appearing on the internet. Same concept goes back to nomenclature of 'Script Kiddies', who were n00bs! that just cut and paste computer hacking code without really being a Computer Hacker. The point is, examples and code on the internet is great - but please take the time to understand them as you use them!

Mea culpa

OK OK! after all that ranting - for my first tests, I will do the same thing!!  Just to get things started and to learn more about the systems. However, I did take time to examine each of the libraries and code to make sure I understood it. As well as had to do some minor adaptation for my hardware.

In searching the internet and git, there are a few copies of the exact same code. I found that Makertut, has the code on git, and even though it is clearly copied, it is cleaner than some of the other code sets (even from James Bruton) for this project.  Then I was surprised to find it was a copy of kurimawxx00.  So, who was the original, (Frank?) honestly I don't know and I don't think it was either of these.  They both contain code for a motor controller LMotorController - in which I can not find the original creator.  It is simple code to control output on a L298N H-bridge, but why have special code for that?

Special shout-out, through my internet searches I have found that Dinh Tung “Tony” Le, AKA mechatony, Gets some special credit. He gets extra credit as he had to build all of the libraries from scratch in order to get school credit for his project. Tony has removed his code from the internet. But the explanations of the difficulties of this task are spelled out in his videos. this concept was often overlooked by the 'copies'.

#include "PID_v1.h"

#include "LMotorController.h"

#include "I2Cdev.h"

#include "MPU6050_6Axis_MotionApps20.h"

The result after all of this, I was able to install and compile the basic libraries for MPU6050, I2Cdev, and PID (and LMotorController - roll eyes!). The guys that wrote these libraries are awesome for their support of the maker community.

With some (also copies) example code, I was able to get my test leg moving and reacting to the pitch change of the MPU6050.  The motor accelerated with the larger pitch and reversed when pitching backward.  This is the basic proof-of-concept needed at this stage of the rapid prototype.

Next steps

Print out and assemble the second leg of the Robot (see image), and then design the chassis or mounts for the microcontroller, mpu6050, and L298N H-Bridge. for a battery I will be using a Talentcell 12v/5v battery bank.

Tilt! Balancing Robot - Prototype layout

Here you can see all of the components that will be part of Tilt.  Notice that the motors, belts, rods, rubber stoppers, and screws are all recycled from Printers.  Also, the bearings are from discarded inline skates.

One can dream...