Tension Biped - A new take on a 9g Servo Walker

- A new take on a 9g Servo Biped Walker

Similar projects worth following
"9g servos dont have enough torque for a biped" - Every robot forum everywhere.

Using spring tendons to aid, I aim to prove this as incorrect.

Mk 3 is an Original Design 8 DOF Biped "Chicken" Walker using tensioned parallel bar legs.

Using plain rubber bands - Ive replicated what is normally seen on your trusty work light, parallel bars, with a spring loaded to hold the weight.

the rubber bands support Just enough weight to allow the servos to hold the weight of the bot, and just little enough to still be able to lift the feet.

standing at 14" tall (fully extended), its a completely unique class of Biped robot.

All Original Design 3D printed chassis , using 8x Tower Pro 9g servos.

Mk 3 is an Original Design 8 DOF Biped "Chicken" Walker using tensioned parallel bar legs.

All parts Easily Printable without supports

8x Tower microservos, 2 for each ankle and hip, knee actuated by long lever dually by hip and ankle.

The majority of the weight is supported by a plain rubber band tensioned across the back of the "knees", its just enough force that the servos can both lift the bot, and still lift the opposing foot.

Uses only standard "garage" hardware- M3 and M4 metric hardware, every Home Depot and Harbor freight has this stuff on hand..

Small Finish nails for Joint pivot points- any small nail with a matching drill bit would work.

  • 8 × SG90 Servo Tower pro Micro 9g Servo clone
  • 1 × Entirely 3D Printed Chassis All Original Design, All easily printable without support.
  • 1 × Arduino Nano 328 5v/16mhz The Brain board for the bot itself
  • 2 × nRF24L01 2.4Ghz Transceiver 2.4ghz Datalink one on the bot, one on the base.
  • 1 × Arduino Mega 2560 The brains of my "control station"

View all 7 components

  • More Code and Electronics

    Hooptie J11/05/2015 at 03:56 0 comments

    This week I've been spending time in 123D.Circuits and Fritzing Getting all my wires mapped, and testing out bits of code.

    I bring you The robot-side Schematic - Thanks to!

    We're looking at - The 8 Robot Leg Servos, Battery pack for power, Arduino Nano uC, CommonAnode Dual Color LED, and the NrF24l01 2.4 module.

    The receiver Powers on, and lights the red LED, once the radio gets a link from the transmitter/controller side the LED lights green, it will flip red if the signal is interrupted or out of range.

    Transmitter/Controller side Coming Soon!

  • Brraaaaaaiiinns..

    Hooptie J10/30/2015 at 19:20 0 comments

    Got Great work progress on the Electronics this week!

    Basic "Control Station" and 'Bot Controllers are up and running on Arduino-

    "Control Station" Electronics

    • Arduino Mega 2650
    • NRF24L01 2.4ghz transceiver.
    • Sticks and buttons!
    • USB Powered+debug
    • printed control housing? - long term

    Bot side electronics

    • Arduino Nano 328 5v/16
    • NRF24L01 2.4ghz transciever
    • custom Power Harness
    • 6v battery powered (4x AA or AAA)
    • 8 Servos

    I have Radio Link working well, and servo direct control functional, time for some wiring and Animation work!

    - More updates inbound! (i took a whole mess of pics)

  • Ive been away ! Updates Inbound!

    Hooptie J10/23/2015 at 22:19 3 comments

    Sorry all, There's been some progress, but ive been totally wiped out by Real life -

    spent 2 weeks sick with a throat infection, no progress-

    torn rotator cuff - yeah, a 5'10 150lb guy probably shouldnt move a full sized file cabinet alone without a dolly, my dad thanked me for getting it out of his house, but recovering from a throat infection , then tearing my rotator cuff, ive seriously hurt my ability to Day-to-day anything, so, very little progress for almost a month

    that all said - the last week or so has been a whirlwind of progress -

    New Control systems Have arrived! - Arduino based Brains and controller, we'll be seeing a Nano328 taking the place of the Raspi+I2c Board, and a Mega 2560 on the base station end, NRF24L01 Wireless 2.4 radio link .

    i DID get extras of EVERYTHING this time, ive got an Extra Nano, and a couple of spare 2.4 transceivers,

    Ive already got basic "stick to servo" controls functional, Sidenote* Arduino is SUPER easy to get started with..

    I also ordered a Bucket of Extra Servos to replace the ones ive damaged already (both waist servos, and one ankle servo have developed dead spots and/or grindy gears) , with enough leftover for a second bot.

    that said, A Second Bot is under construction - calling him "One", started calling the original bot "Zero".

    plans are to test against each other for part and software revisions, a walking "spare parts supply" and possibly to have an extra robot for demos or purchase possibilities.

    Ive Also got Mk3 UNTETHERED - with the new Arduino + 2.4ghz link system, it was as simple as slapping on some AAs in a battery pack and Off to the races!

    I'll get some more in-depth updates and pics later!

  • Foot/Ankle Mechanical redesign pass 1 - Done!

    Hooptie J09/25/2015 at 05:31 0 comments

    Finally got some time to work on the ankle/foot attach design -

    I made a stop at Harbor Freight this week, purchased a "Small metric hardware" variety pack - and have been installing M3 & M4 hardware throughout.

    No more random wood screws- That said - A few less limitations on the Design process.

    Left New - Right Old

    its 5 parts, up from 1.

    Its a Bit more complicated, But much easer to print, although 5 parts instead of one, they all fit just fine on a single plate (minus the footpad) though.

    After battling my Q3D for a few hours today -

    It fits together Great - 0.25mm gaps in the design, snug fit after printing.

    The leg bones fit quite a bit easier, but , I cracked the servo mounts on the new footpad, probably will need a "strengthening" pass and a definitely a new print.

    Its held together by M3 hardware now, 3x 5mm, 1x 10mm. The heel side of the foot pivot is now an M4 x5mm

    More printing to do, been outlining some software ideas, More updates as they come!

  • Success and De Feet!

    Hooptie J09/04/2015 at 22:06 0 comments

    Update on the foot re-design process - New "foot pad" is designed, now working on modifying the ankle parts to accommodate them -

    layed-out some quickie adaptor plates to mount them in the mean-time.

    1 Foot printed, looks 100%!, printing "left" now, and the ankle-adaptors next -

    Also - Have spent the last few nights (several till 3a+) trying to wrap my head around using python to control it- Updates on that later.

    I'll be out of town for a few days so no updates, but im taking the 'bot with for some show&tell, who knows maybe i'll be struck by some inspiration on how to get past my python woes.

  • Servo Brace Progress

    Hooptie J09/02/2015 at 06:14 0 comments

    As Promised - Update on the "waist" servo braces.

    basic design done and printed - I'll definitely be making a cosmetic pass later, im not a fan of the looks of these, but it works substantially better than the previous "fancy" ribcage.

    And the part count grows by 2. - from a 2 piece "ribs" to 2 parts for the servo mount, and 2 for the "center" - again , no joke, Anyone who wants to help name the parts are welcome to comment - ive taken to calling this the green section the "ribs"

    And - In the flesh

  • Legbones connected to the ... El-bone?

    Hooptie J09/01/2015 at 18:44 0 comments

    after breaking a "Tibi-fibula" (i dunno what to call the legbones, any zoological anatomy folks wanna help?) - anyway - after breaking a leg bone during R-O-M checks , i decided to do a minor re-work on the tendon hooks. - New legbones in White/translucent

    while i was at it , i modified the El-bones?Kneebones? whatever - to include a tensioner on the rubber band tendon. (new legbones in white, new "elbones" in green)

    the Elbow/Knee now includes ramps that pull the joint parts together when the tendon is stressed, instead of the tendon slipping down between the parts, its increased the lateral stability of the legs noticeably.

    I also Got most of the design on the upper servo brace/bracket done and printed- i'll add another log with the specific there later today.

  • Wired and Hot

    Hooptie J08/30/2015 at 09:00 0 comments

    Spent some time today working on the wiring and software side of the bot.

    I've hooked up my Raspberry pi along with an Adafruit I2C Servo controller to get some basic range of motion and movement checks.

    Everything looks Pretty good, Standing on its own power, and able to crouch lean, etc -BUT.

    Chassis side I need some minor revisions:

    - on the feet, there's an issue when all the weight is on one foot, it flexes the servo horns just enough to snap out of place from the bottom of the ankle, and it falls back onto both feet 1/2 the time, it can only stand on one foot when this works... - pondering the right "fix" for this- may need a bigger revision of the foot and ankle attach.

    -On the "waist" servos - They definitely need a brace, as the servos promptly stripped out their mounting screws and started getting floppy- a rubber band was enough tension to keep it all functioning, but a small brace/battery rack is likely coming soon.

    - on the legbones, there's some range of motion issues when it comes to fully "tucking" the legs, ended up breaking a "femur" side inner legbone from it. Easy fix though, just need to angle the "tendon hooks" 45 or so.

    Also, i didnt print any spare leg bones last revision, so this ends Range of motion testing for the day.

    On the Software side - i got a BUNCH done-

    Test app for the walker in python, uses a USB gamepad for input, basic UP/Down , Lean Left/Right, No animations , but got a stable workable input for testing.

    I'm currently limiting the speed on the servos- 3deg/sec slow - i took a video but it was godawful boring to watch, movie magic later....

    walking will be static animations to start, then down from a few steps to procedural half steps eventually.

    For now, part redesigns and printing replacements should keep me busy for a few days at least.

  • Updated Hip Linkages

    Hooptie J08/19/2015 at 18:09 2 comments

    Continuing basic range-of-motion checks, i noticed that the linkages for the hips would hang up on the hip mounts, did a minor redesign- it got rainy and humid here yesterday, so my 3D printer got really finicky- it only took 3-4 tries to get the linkages printed, but finally, success.

    New Hip Linkage - Slimmed for clearance.

    And installed on Mk3.

  • Updated Pics - as of 8/18

    Hooptie J08/19/2015 at 03:19 0 comments

    Some Updated Pics!

    Had a minor redesign of the Leg bones today and the updated Hip design was done printing this morning, - Pictures are up.

View all 11 project logs

Enjoy this project?



stefane.lemay wrote 05/08/2019 at 16:23 point

Since there were no update in a very long time, I'm assuming that this project is shelved or on hold...

  Are you sure? yes | no

RigTig wrote 03/11/2017 at 10:45 point

When 'everyone' says it cannot be done, I reckon its great when a person just says 'Just watch me'. Well done!

  Are you sure? yes | no

delosangelesmaria224 wrote 11/24/2016 at 02:14 point

And when will we see it walking ????

  Are you sure? yes | no

Jonathan Lussier wrote 03/28/2016 at 02:17 point

Any updates on the project?  Did you manage to post any of the design files online?


  Are you sure? yes | no

Jarrett wrote 11/05/2015 at 17:48 point

More videos! I want to see some Robocop stuff happening :D

  Are you sure? yes | no

Hooptie J wrote 11/05/2015 at 18:45 point

Request acknowledged! - *note to self, i need a 1/10scale staircase for testing

  Are you sure? yes | no

richard wrote 11/05/2015 at 15:29 point

Hi is there any github or other page where i can download the stl files?

  Are you sure? yes | no

Hooptie J wrote 11/05/2015 at 18:45 point

Not as of yet Richard, He's still really "alpha" stage, also, im still working on a good way to get the v3bot here out and seen, i'm just not quite ready to let him out into the wild yet.

  Are you sure? yes | no

richard wrote 12/04/2015 at 07:16 point

is now the right time?

  Are you sure? yes | no

The Big One wrote 09/08/2015 at 14:53 point

Neat project!  You are saying that you are trying to get python control working... if you want, you can take a look at some of my code.  I use Python over a serial port (Bluetooth, FTDI + XBee, etc) to control my hexapod.  I use a framing protocol which gives me 255 commands with variable payload and checksumming.  The protocol implementation is in both C / C++, Python, and Java (and it would be easy to port to other procedural languages as well).  Take a look at #Stubby the (Teaching) Hexapod for details and code.  Feel free to ask if you have questions.


  Are you sure? yes | no

Hooptie J wrote 09/09/2015 at 16:46 point

Thanks for the reference! - as to the Python issues im having.. well, the extent of my coding experience before I picked up the Pi was some HTML and applescriping back in the OS8 days. That said, ive got "control"  i can move the servos, address them individually, Ive got Pygame seeing my control pad, and can pass commands from the controller from my computer to the pi using sockets&pickling in python. I can do all these things individually- im just kind of stumbling a bit tying those into a cohesive functional program.

for example - I've got an interactive python script that lets me map a servo to a controler axis and move it- but im not quite to the point where i can move multiple servos different ranges simultaneously, nor do i have any "animations" mapped out.
im struggling with managing what classes and subclasses i need, or the best way to make use of them, - its kind of the nature of the beast when self-teaching anything though, "best practices" only come with lots of practice, ive got a few ideas, but not enough time to practice them at the moment-.

  Are you sure? yes | no

The Big One wrote 09/09/2015 at 17:44 point

Hah, fair enough.  I'm not a python expert by any means; I find it is great for small scripts, but when I start writing larger programs with multiple classes things get confusing.  Oh well.  :-)

Anyway, awesome job so far, doing a walking robot with el-cheapo hardware is sweet!

  Are you sure? yes | no

deʃhipu wrote 09/10/2015 at 06:31 point

I found it very helpful to use iterator generators for anything that has to happen in parallel. For instance, my robots move all the feet that are on the floor backwards constantly, and in a separate subroutine, they make steps with each leg in turn. That's two separate iterator generators, using "yield" where you would normally use "delay()", and then there is the main loop which simply iterates over them all every frame. You can see the code at

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates