Close
0%
0%

Thor

OpenSource 3D printable Robotic Arm

Similar projects worth following
Thor is an Open Source and printable robotic arm with six degrees of freedom.

Its configuration (yaw­-roll­-roll­-yaw­-roll­-yaw) is the same one that is used on most

manipulator robots that currently exist in the market.

In its upright position, Thor is about 625mm and it can lift objects up to 750 grams.

The project started a year ago as my Final Degree Project called "Design and startup of an

Open Source and printable 6DOF robotic arm" but a lot of things have changed since the

presentation day.

The main purpose of this project was to create a robotic arm that could be used in

universities and schools to teach robotics instead of using simulation software or low

accurate models. Having this in mind, the final prototype had to be affordable and, of course,

Open Source.

If you are interested in this project, please consider joining our mail list.

To develop this project, I have been involved in several areas of expertise:

Hardware choice

In order to have a nice accuracy on the moves without raising the price too much, I chose stepper motors for the movement transmission. I had to choose stepper motors with mechanical reduction for the second and third articulation in order to manage the torque generated in these articulations. For the two last DOF I chose smaller steppers with the purpose of reducing their weight.

For the electronics I chose an Arduino Mega as main controlling board. And I designed a "shield" to make possible the control of 7 steppers. I chose pololu A4988 as stepper drivers.

For the transmission of the third, fifth and sixth articulation I have used GT2 Pulleys and GT2 Belts.

Finally, I chose optoisolators and a micro-endstop in order to establish a home position for the first five articulations. Using optoisolators instead of mechanical stop allows it to rotate more than 360 degrees without colliding with the sensor.

As you may have realized, all of the hardware I have chosen is commonly used on DIY 3D printers.

3D Design

I used FreeCAD software in order to design the pieces of the robotic arm. There are 37 different printable pieces in Thor. For every piece there have been a lot of iterations behind the final one.

Some design solutions are inspired by commercial robots, others are the result of hours of work and others are just serendipitous ideas.

The final design had to be compact and safe. I wanted to hide every motor and every single wire to make it more aesthetic. Also, I designed covers and protections to reduce the risk of entrapment while the user is manipulating the robotic arm.

None of the 6 DOF is a end-effector. There are a lot of commercial and DIY grippers/vacuum/hooks around the web and for each use something different is needed. Instead of integrating a gripper into the design I decided to make an adaptable design making an interface between the end-effector and the robotic arm. This way you have 3 options to include a tool on Thor:

1. You can modify the interface, placing the mounting holes wherever you need them.

2. You can modify the tool, matching its mounting holes with the interface ones.

3. You can design an intermediate piece that fit with the interface on one of its sides and with the tool on the other side.

I like the whole robot, but I'm especially proud of these 3 solutions:

- Semi differential (5&6 DOF)

This kind of transmission allows it to do two types of movements using two small gears that act over a big one. When the small gears rotate in the same direction, the big gear bloks and rotates around the small gears' axis. When the small gears rotate in opposite directions, the big gear rotates around its own axis.

This way, actuating over the two small gears allowed me to place the steppers as lower as I could, transmitting the movement with GT2 belts and reducing the torque generated by the motor's weight.

- Support point of 2nd DOF shaft

While I was designing the second and third articulations I was worried about the available space. The motors tookup a lot of space and there wasn't a single small area to establish the second support point for the second articulation axis. The solution was one of the serendipitous ideas I have mentioned early. Why not use the shaft of the stepper of the third articulation as the support point for the second articulation axis? And voila! It worked well.

- DIY Bearing

The design of the 4th articulation had an issue, I didn't find a commercial bearing adjusted to my needs. After days of searching, I decided to make my own. Inspired by a solution that some users used for the Ciclop's bearing, I designed the bearing using 6mm airsoft balls as bearing balls. The result was precise, almost without backlash (after some printing tolerance tests) and very cheap.

Electronic Design

Before this project,...

Read more »

sla - 864.73 kB - 08/03/2016 at 14:50

Download

sla - 991.29 kB - 08/03/2016 at 14:50

Download

sla - 29.57 kB - 08/03/2016 at 14:50

Download

sla - 745.20 kB - 08/03/2016 at 14:50

Download

sla - 133.68 kB - 08/03/2016 at 14:50

Download

sla - 424.89 kB - 08/03/2016 at 14:50

Download

sla - 34.65 kB - 08/03/2016 at 14:50

Download

sla - 90.51 kB - 08/03/2016 at 14:50

Download

sla - 255.26 kB - 08/03/2016 at 14:50

Download

sla - 110.53 kB - 08/03/2016 at 14:50

Download

View all 39 files

  • 4 × Kg of filament
  • 1 × Nema 17; L=40mm; Holding torque: 39.22 N.cm
  • 3 × Nema 17; L=34mm; 5.18:1 mechanical reduction; Holding torque: 121.2 N.cm
  • 3 × Nema 17 L=34mm; Holding torque: 21.57 N.cm
  • 2 × GT2 closed belt 208mm (104teeth)
  • 1 × GT2 open belt 1m
  • 1 × GT2x20 Pulley
  • 2 × GT2x40 Pulley
  • 1 × 16014zz Bearing
  • 11 × 625ZZ Bearing

View all 51 components

  • Control PCB Wiring Diagram

    AngelLM2 days ago 1 comment

    I just made a Wiring Diagram for the Control PCB v1.0 board.

    You will find the full resolution image & source files at the Github repo.

    I want to remind that there is a Google Groups community with 90 members in this right moment where you can ask, share and read about this project! Join us! :)

    Check soon my fork of the grbl repo to get the lastest firmware for the ControlPCB board.

    Best regards,
    Ángel LM

  • Thor's community is now online

    AngelLM01/15/2017 at 16:30 3 comments

    Hi!

    Since a few weeks I'm receiving a lot of questions/feedback/information from many different directions (email, Twitter, Hackaday, Thingiverse, GitHub, Wevolver, etc.).
    One thing is clear, this project has become known, at least more known than I expected at the beginning, and I cannot do other than thank you! Thank you for your likes, your comments, your feedback, for sharing the project, contributing to it and for building a Thor. You rock!

    Also, from the very beginning I wanted everyone to became part of this project and share all the information. I think that the current platforms I'm using are not the ideal ones to develop a community project. Using a lot of platforms is convenient for spreading the word, but not for focus the information...

    That's why I decided to make the Thor mail-list community. I have been in other communities of Google Groups and it works so well. Everyone can access, everyone can ask/answer/show, it's easy and clear and it's free. In this way I want to share all the info I have with you all. And you will be able to help, give/get feedback and share your ideas, progress and modifications with other community members.

    From now I'll group all info there and I'll be answering the questions there. I'll be posting there my progress and changes too (I will still doing that here too).

    See you there!

    EDIT: I forgot to say that I also did a Frequently Asked Questions in order to answer them :)

  • Danny's mods incorporated!

    AngelLM12/19/2016 at 16:05 5 comments

    Hi!

    Time ago, Danny made some modifications to largest pieces, splitting them and making them printable in 180*180mm bed area printers.

    He made an awesome work, and I wanted that modifications to be in the main repo, so I added it! Also, I migrated the design to FreeCAD in order to have the source files too ;)

    EDIT 21/12/16: I have migrated another 2 modifications to FreeCAD :)

    On the left, my remix of Art4Body modification made by Danny. Designed for low height printers, thank you again Danny :)

    On the right, the BaseBot modification made by Ctrl-Alt-Dude for 190*190mm print area. This will allow more users to print Thor. Thank you too! ;)

    Are you having issues with other pieces?

    Season's greetings for all!

  • New ControlPVB PCB arrived!

    AngelLM12/16/2016 at 14:33 1 comment

    Yay! I received them 3 days ago... Exciting!

    I ordered them to PCBWay and my experience was great! They are fast, kind and very cheap :D And the quality of the PCBs is incredible, very professional. I'll work with them again for sure!

    As you see I have designed it for THT and SMD (1206 package) components. I never used SMD before, but there's a first time for everything haha.

    Now, it's time to wait for the arrival of the components... I think that it will take about a month or two... I don't know if I'll be that patient...

    Hope you like it! :)

  • New electronic board very soon!

    AngelLM12/07/2016 at 21:43 2 comments

      Woah, more than 300 followers! Hello everyone!

      As I said weeks ago, I needed a new electronic board to control Thor... The one I was using had many mistakes (a man may learn wit every day...) and I got tired of botching it. The initial idea was design a DIY PCB board, but due to its limitations and how cheap is to manufacture it in China I decided to design a better control board.

      Of course using OpenSource software, KiCad again.

      First of all, some pictures:

      It took me a few hours to understand the values that I had to set (track width, drills diameters, etc.) in order to send it to the manufacturer... but it won't happen the next time >:D

      Anyway, the new features:

      1. First and the most necessary: Reverse current protection circuit. No more burned drivers because of mistaken power input polarity... It will save my wallet!
      2. This board allows to control up to 8 stepper drivers (instead of 7 of the previous version).
      3. Smaller size: Less than 100mm*70mm.
      4. Easier assembly: the Silkscreen is a nice helper!
      5. Added 3 confirmation LEDs (Power Input, PCB Power, Arduino Power).
      6. Added more 12V aux outputs for the fans.
      7. Added a switch to power off the steppers and keep the fans working.
      8. The connections between the drivers and the male pins that connect with the Arduino Mega are now well driven. This has been possible thanks to the 2 layers board, and it will allows to simply connect a ribbon cable with a 2x18 connector without "hacking it"

      And I think that's all! I'll send it to the manufacturer as soon as I finish another PCB I'm designing.

      Of course, I shared it on the Thor's Github repository. If you are reading this close to the publish date, you will find the new PCB files in the ControlPCBUpdate branch. Feel free to take a look and share your opinions!

      Hope you like it!

      Best regards,
      Ángel LM

  • Thor at Bilbao Maker Faire!

    AngelLM11/21/2016 at 16:11 1 comment

    This weekend I was at Bilbao Maker Faire showing Thor. The event was great and I met a lot of interesting people.

    I met Antonio, the creator of DIMER, who was located next to me at the faire, so we managed to interact our robotic arms with each other... And here is the result! (Maybe the first marriage proposal between robotic arms?)

    As you can see in the video, Thor is moving faster than weeks ago. I increased the speed of the slowest articulations about 50%. Soon i'll test the maximum speeds that Thor can support.

    Another great new is that I have been using Thor for 8h non-stop working and it didn't broke or melt neither! Yay!

    Next things I'll be working on: Test the maximum speeds, making a conveyor belt that will interact with Thor, design a new electronic board and add a new degree of freedom (hehe >:D).

    Finally, I just saw the awesome work of Danny... Just go to his project page and see it with your own eyes...

    Best regards!

  • Overheating problem fixed!

    AngelLM11/07/2016 at 21:12 0 comments

    Hi everyone!

    First of all, thank you for your 100 likes 276 followers & over of 15k views. You rock!

    We have another thing to celebrate today hehe. The problem about melting gears and malfunction due to steppers overheating is solved :) Let me explain the steps I followed to fix it.

    1. I measured the temperature of steppers in action... and it went over 70ºC. PLA starts deforming at 60ºC... So I had a real problem there. Not only the gears were deforming, also the Art1Body part were deforming and causing the malfunction.

    2. I appreciated that my design didn't had refrigeration holes, so the heat produced by the steppers only could be dissipated through the plastic (bad idea because it was causing the deformation). After facepalming myself I redesigned the parts that contain steppers and introducing 40mm fans into the design. In this way, the motors would have forced cooling.

    3. I tested it. And it worked well, now the steppers don't reach 40ºC and it goes well! Just in case I printed the gears using ABS to prevent the deformation. The test was performed at a 10h non-stop maker faire called OSHWDEM at A Coruña (Spain).

    So, check the new design at GitHub repo before print it! And if you have printed it already don't worry. Take the dremel, do some holes and glue the fans! (I did this hahaha)

    Aaaand... there you have a video of Thor doing a manipulation job. I think that the next step is increase its speed, don't you?

  • Beeing part of this project

    AngelLM10/10/2016 at 12:12 0 comments

    Hi everyone!

    I have received some applications from Hackaday users that wanted to be part of this project. And I think that it's important to explain how to be part of this project.

    In one hand, this project is completely Open Source and shared on its GitHub's repository. This mean that you can share your modifications and beeing part of this project by doing pull requests, giving feedback or helping solving issues.

    In the other hand, there is the "Project Team" on Hackaday. I don't really want (and I think no one wants) to have a Team with a lot of inactive users. I think that this Team is a nice place to share opinions, test designs, test features and things like that. Obviously, the team has the labour to gather all this information and make it public too!
    In this way, everyone could access to all the info and the Team members would have a place to share their progress.

    And the final question: How to be a Team Member? I think the best solution is accepting the requests of the users that contribute. So, if you want to join this Team, I suggest you to go to the GitHub repository, clone it, make something nice and send a pull request!

    What do you think about this?

    ¡Regards!

  • Inverse Kinematics

    AngelLM08/18/2016 at 18:44 0 comments

    Today I'll share the solution of the Thor's Inverse Kinematics.

    Inverse Kinematics (IK) allows to find the robot's joint parameters (q1, q2, q3, q4, q5, q6) that do robot's tip to position according to a specific spatial location (p,[n,o,a]).

    To do the Thor's IK calculus I have used the kinematic decoupling procedure. This procedure allows to calculate, the joint parameters (q1, q2, q3) that place the robot's tip in a point of the space first and then calculate the joint parameters (q4, q5, q6) responsible for the wrist's position and orientation.

    Geometric method: q1, q2, q3

    We have to bear in mind that the point that the IK will position won't be the Tool Center Point (TCP) but the 5th joint axis' center (Pm). Therefore, given an orientation and position of the TCP (noap_tcp) and because the 4th, 5th and 6th joints cross in a single point and the TCP's Z axis crosses Pm, then:

    Geometrically we obtain the following equations:

    where:

    Rotation matrices: q4, q5, q6

    Thanks to the forward kinematics the Pm's position and orientation due to q1, q2 and 13 can be obtained. As the position is no longer needed, it can be possible to operate only with the R:

    Which is the same as:

    I have used the SymPy tool (mentioned on the Forward Kinematics log) to do the symbolic calculus. The results are the following:

    Hope you like it!

  • Thor's new brain received!

    AngelLM08/18/2016 at 16:20 0 comments

      As I said, Arthur Wolf from SmoothieWare sent me a SmoothieBoard to test it in Thor.

      And I just received it! It seems pretty awesome and they have included a LCD Screen in the package! Yay!
      The board can control up to 5 steppers, so I'll need to add another driver in order to control one more.

      Next steps will be:

      1. Study the board
      2. Study the firmware
      3. Test the board with the standard firmware
      4. Modify the firmware to adapt it to Thor
      5. Add an additional driver
      6. Test it all with Thor

      Regards!

View all 12 project logs

  • 1

    Print it!

    3D printer: I have used one with 300x220mm of printing area and 0.4mm nozzle.

    Material: I printed every part with PLA.

    Printing Profile: I recommend to slow down the outer shell speed at least to 50% of its normal printing speed in order to get accurate results on holes, gears, etc. For some pieces supports are needed.

  • 2

    Prepare materials!

    If you want to include the home sensors, you will need to make the PCBs and weld the components.

    It's also recommended to weld a wire extension to the steppers wires, specially for the last three ones.

  • 3

    Make the electronics!

    Make the Control PCB and weld the components!

View all 6 instructions

Enjoy this project?

Share

Discussions

adam wrote 03/09/2017 at 14:00 point

I have started printing parts for this. As i do so, I am wondering if anyone has thought about scaling up the design to accommodate NEMA 23 and 34 motors?

  Are you sure? yes | no

Olaf Baeyens wrote 01/10/2017 at 23:04 point

Maybe Danny has better results in the 3D printing quality but in my prints I have noticed that for accurate movement your "theoretical" calculated movement will not be accurate enough for reliable automated percision movement. Joystick control yes, but repeating the process I doubt it. 

Also every 3D model that gets printed will be slightly different.

I think in the software must take into account imperfection to compensate. But first I think we need a mechanism to actually measure the imperfections first.

  Are you sure? yes | no

AngelLM wrote 01/11/2017 at 18:00 point

Hi Olaf! What is the calculated movement that are you talking about? About repeatability I didn't test it yet, but I got almost equal movements doing a process two or more times. I'll upload a video soon :)

Printing tolerances is an issue hard to deal with... Every printer has his own tolerances and are different from other printer, as there are only one 3D model to print probably there will be some printed pieces better than others.

Of course doing a closed-loop control with sensors (potentiometers or encoders) will ensure that repeatability.
If it helps you I applied the trial and error method to identify my printer's tolerances before print them. 

Best regards!

  Are you sure? yes | no

Olaf Baeyens wrote 01/11/2017 at 18:20 point

It is only intended as feedback so we can improve the Thor ;-)

I am not that far yet, I am now building the code (proof of concept). I have now my base that can actually rotate. But when you look at the side, then you see that the plastic top is a bit tilted. This will affect the precision when I ask the robot to rotate around its axis. The payload may shift up or down depending on the angle.

I have been toying with the idea of a Gyroscope+accelerometer, but I am still busy setting it all up (buying the electronics, wires, .... since I am a complete N00b)

Another thing I am thinking of is to use a photo-detector with a ring hat has multiple varying spaces. If one space is 5 steps to cross and another 10 steps, yet another 15 steps  then it has a sense of direction but also a sense of how off its position is. It could compensate real time during the movement.

Example: If the software expect a gap that starts at position 105 but it already detects it at 103 then the robot knows that it must add 2 additional steps to get to the correct position. If that gap is in the middle of normal operation then the robot can work longer with less need of a recalibration getting to the home position.

The hardest part is to develop that software and squeeze it into that small memory space :-)

  Are you sure? yes | no

Olaf Baeyens wrote 01/11/2017 at 18:51 point

Ah yes: "Calculated movement" Your forward kinetics formulas.

  Are you sure? yes | no

AngelLM wrote 01/11/2017 at 21:59 point

Of course! I really appreciate your feedback :) I think it's the only way to progress! 

About the forward kinematics and control: to be honest I estimated the steps per unit (the steps that the motor has to rotate to complete a degree in this case) in firmware by trial and error, maybe it is the main issue about the forward kinematics control. I'll upload soon the steps I followed to do it. About repeatability, do you have some backlash between pinions and gears?

Best regards!

  Are you sure? yes | no

Olaf Baeyens wrote 01/12/2017 at 18:38 point

If I can find time this weekend I will test accuracy. I did try to hold the base with my hand to see if it broke or halted the stepper motor, but it has decent force. But I did manage to make it stop by hand. (The plastic itself did not bulge)

I also discovered the mystery why my optical ring came lose. It turned out that the screw that held the stepper motor cog in place was a bit too long. When I assembled it I did notice but it probably pushed the optical ring from the inner cog base.

I am wondering when you move a heavy object how it behaves too. We may have to modify the stepper motor code to accelerate slowly and also stop more slowly to keep accuracy. I am wondering if we can put a weigh senors in the base so that the robot motor controls can optimize its speed depending on teh laod it carries.

  Are you sure? yes | no

dannyvandenheuvel wrote 01/11/2017 at 19:07 point

The main reason I did not build the robot so far, I tried to make the software almost complete and experiment with it to see what are the main bottlenecks and issues to have a perfect robot. After all proofs are well done I build it for 100%. I printed all parts and I see some problems with the heating and the precision. For the heating I gonne modify some parts with build-in fans like AngelLM already done, for the precision I am thinking to make a feedback with a encoder or something I can measure to angle. In that way I can make a closed loop controller for the stepper motors. If I make it absolute I don't have to home, just for the first time setup. I have to luck to have a proffesional stratasys uprint se printer, the accuracy is a lot better than the hobby printers. I didn't have any problems printing the parts, some parts I made a little smaller or broken into two parts because the dimensions of my printer area. Olaf, if you wish, I could make that part where you have some troubles with be split in two parts, I will mean, the upper shrew blocks could be the second part with a little inlay plate to fit inside the second plate and can be glowed togetter. Then you can print both parts without structures.

  Are you sure? yes | no

Olaf Baeyens wrote 01/11/2017 at 21:26 point

You are reading my mind, I too am focusing on seeing how the robot behaves first so I can pair the software and the mechanics so that act as one entity. If only I had more time after my day job :-) 

I am interested if you could modify the model to avoid having support material to be the Guinea pig :-). I probably wasted 25% or more plastic on that support structure. The plastic is not the issue but the time to print it is more of an issue.

I think what is important for people that wants to copy the Thor is that it can be printed in 12 hour maximum 3D prints. The longer the print lasts the bigger a cache that a print fails. In addition, people are less interested because the part is too big so people give up to soon.

I really do enjoy both of your work greatly. It challenges me and it is very rewarding when parts comes together and start to acvtually move.

  Are you sure? yes | no

vader333 wrote 12/17/2016 at 09:51 point

Hi there, I'm looking to build a Thor, and am ordering parts now. I would like to modify the model to fit a hitec servo, as I'm reusing the servos in another project. Do you think it is possible to do so without altering the mechanism too much?

  Are you sure? yes | no

JustinShaww wrote 12/18/2016 at 05:06 point

Can I ask where you are ordering your other parts from? I am also trying to put this build together and quite frankly I have no idea where to begin! 

On a more general note, I downloaded the ListofMaterials.txt file but it looks like steppersonline.com is no longer a working domain and also the counts for the quantity of each motor is off (for example, the .txt file says you'll need (2) 40mm Nema 17 servos but the components list on this page says you only need 1. I've also checked the hub and it calls for only 6 servos (versus the seven in the txt file and the components section of this page) and of those 6, half of them are Nema 17 and the other half are Nema 14. What is the difference? Are the 14s just smaller/lighter? Does it matter which ones you get, or will it affect the performance/strength of the arm?

Any help is appreciated, 

Justin 

  Are you sure? yes | no

vader333 wrote 12/18/2016 at 08:16 point

I'm getting most of my parts off taobao.com. Nema is a size standard; Nema 17 means 1.7 square inch face plate. Nema 14 is 14 square inch face plate. Download FreeCad and the fcstd file and look for yourself.

Motor size doesn't necessarily correlate with performance.

  Are you sure? yes | no

JustinShaww wrote 12/18/2016 at 23:57 point

You're a life-saver Vader! Thanks so much!

Sidenote: How many of each motor do I need? The hub says I need 6 total, and this hackaday page says I need 7 total... Does it matter?

  Are you sure? yes | no

vader333 wrote 12/26/2016 at 06:39 point

7 based on the model.

  Are you sure? yes | no

Olaf Baeyens wrote 12/26/2016 at 21:03 point

I also have a hard time to find the components, they are spread everywhere.

I ordered the motors from this site: http://eu.stepperonline.com/ but so far I have not been at a point to actually test the motors. It took some time to receive them. Also be careful, some people seem to have had a import tax, I did not.

There are 6 degrees of freedom so in theory 6, but one of the joints needs 2 motors that work together to get the needed power to get the arm up.

There are 3 motors that have a planetary 5:1 gear ratio, those are pretty expensive (I got them for €31 each).

Also the big ball-bearing for the base is hard to get. I found one on Amazon for €31 but I never received it and Amazon canceled the order because of undeliverable. I got refunded. 

  Are you sure? yes | no

dannyvandenheuvel wrote 01/11/2017 at 19:22 point

You can find a lot of things at reprapworld.com. For the motors, I used three types,
for the base I use a little more powerfull type 17HS16-2004S1 and for the first arm and second arm movement I use 3 geared motors 17HS13-0404S-PG5 and for the wrist and tool rotate and up/down movement I used 3 motors 17HS13-0404S. All motors I ordered at steppersonline. Please consider to take they last 3D parts because they will be corrected with extra fans to cooldown the chambers.
The next comming days I will do a full reupload off all my modifications. I will make a new list of materials where you can find all parts at the moment. I changed some 3D parts so you can use more standard products that you will find on the internet.

Greetings,

Danny

  Are you sure? yes | no

dannyvandenheuvel wrote 11/19/2016 at 00:39 point

Hi all! on my page the first glimps of the graphical controller!
Basics has been fullfilled! Now the big works arrives making the
simulator and after that the real TCP socket connection between this
piece of software and the real robo arm!!!!
Take a look and give any comments as you like, all feedback is welcome!

  Are you sure? yes | no

AngelLM wrote 11/21/2016 at 15:31 point

As I posted on the YouTube video, it blowed my mind. Awesome work. You lifting this project to a new level!
Best regards!

(I'll download it right now to see it in actiooooon :D)

  Are you sure? yes | no

dannyvandenheuvel wrote 11/08/2016 at 21:21 point

BTW, it will be a version for PC and MAC! 

  Are you sure? yes | no

dannyvandenheuvel wrote 11/08/2016 at 21:19 point

Hi all! Update!

I've been working hard on the graphical GUI controller at the moment and must say it's moving fast! Inverse kinimatics is working great! There will be Joint jogging and Cartesian Jog with translation along or rotation about X,Y and Z axis. Camera views has a lot of improvements. Watch out these comming days for the first real release! :)

  Are you sure? yes | no

Olaf Baeyens wrote 11/05/2016 at 22:29 point

Hello, am I wrong that Art4Optodisk.stl is missing in your files list?

I am builkding your robot :-)

  Are you sure? yes | no

AngelLM wrote 11/07/2016 at 10:25 point

Hi Olaf! You were completely right, I just uploaded the stl file of Art4Optodisk to the Files List, I'll update the github repository later.
Thank you for the appreciation :)

  Are you sure? yes | no

Ethan wrote 10/16/2016 at 19:34 point

Hello Thor team. I am proud to be building one of these awesome robot arms. my goal is to control it through rhino and grasshopper so that driving the movements can be para metrically controlled. First got to build Thor though. Its taken me a while but I am up to the largest parts now . Thanks to danny vdh for breaking some parts up into smaller pieces. Ill look for those now. I have an ultimaker 2 and some parts are just too big. In the future i hope to also find someway to laser cut some of the parts... so that the fabrication can be faster. and maybe to have an acrylic window to see into the inner workings! I look forward to seeing the builds of others. :) ethan

  Are you sure? yes | no

dannyvandenheuvel wrote 10/15/2016 at 07:48 point

I made some modifications on thor to print it on smaller area 3D printers, I will put the step and stl files later this day on my github page! You can find me on the github page of Angel. There will be a link to my github! If you have any questions don't hesitate to give me some feedback! Success in making a real Thor! (Needed me two weeks to print the whole robot!)

  Are you sure? yes | no

Olaf Baeyens wrote 10/13/2016 at 22:04 point

I want to print this robot but I discovered that my printing bed is not big enough. Is there a way to remodel the 3D parts in such a way that it reduced printing time and can fit on smaller 3d printers?

One of the parts requires me 31 hours printing time, but when I removed support structure I got reduced to only 19 hours. And I think if I can slice it in 2 halves, have the bottom plate printed upside down so I don't need additional support structures, then I can print it way better, faster, ...

of course the slicing must be done in such a way that it also has support structures to keep the parts together when you glue it.

I have the Prusa i3 mmk2 printer. 

  Are you sure? yes | no

dannyvandenheuvel wrote 09/24/2016 at 09:49 point

Hi all you robot geeks! As a member of the 'Thor' project I'm proud to be part of it! I am building one myself for the moment and made some modification to print it on a smaller area of 3D printers. I made some addons and I am working on a graphical GUI to control, program and emulate together with to electronical part for Thor. So take a look at our pages to see the progress. 
Everything will be ported to github of Thor when ready.

If you have any questions or improvments in mind don't hesitate to ask!

  Are you sure? yes | no

Olaf Baeyens wrote 10/14/2016 at 23:05 point

Hello, A few days ago I discovered Thor, and discovered that my 3d printer is too small for the big parts. So I was thinking to take the models and cut them into peaces so they become more printable.

I am a complete noob in the CAd software and 3D printing.

  Are you sure? yes | no

AngelLM wrote 10/19/2016 at 15:01 point

Hi Olaf!

I shared the source files, so it could be possible! But if you are not familiarized with CAD software maybe it would be a bit hard...

I know that the size of some pieces are a problem... and many people asked for smaller pieces that fit in their printers' bed. Right now I'm working on other issues, but I know that there are other users that have some work in progress in this topic. Let's wait to see their work finished! :)

EDIT: Source files are available on the GitHub repository ;)

  Are you sure? yes | no

trevorjtclarke wrote 09/23/2016 at 21:34 point

Hi! Really liking the project you started! You noted "6mm airsoft balls as bearing ball" have you tested this yet? Im curious how stable it is, especially after hours of use. 

thanks!

  Are you sure? yes | no

AngelLM wrote 09/24/2016 at 14:06 point

Hi trevor! Since I made that printed bearing (months ago) and with a lot of use, it's stable as the first day :). Also, you can apply some oil or some kind of lubricant in order to reduce the friction between balls & the printed parts ;)

  Are you sure? yes | no

dannyvandenheuvel wrote 09/23/2016 at 20:33 point

I am trilled to be a team member of the Thor project, did some addons like a connector box and  hand controller and changed some parts to print it on a smaller print enviroment. I am also working on a graphical GUI to program and simulate 'Thor'. It will be multi platform. See some of the progress when I am making one for my own use. Credits for this nice piece of robot development 'Thor' goes offcourse to AngelLM !
The comming days we disguss some ideas and make a list of 'things to do' :)

See ya!

  Are you sure? yes | no

pkmcnc wrote 09/18/2016 at 12:25 point

Hi!

I see that you used gearboxes with 6mm shaft, is it correct? I only can find 8mm shaft everywhere.

Well I found it. The ID is 17HS13-0404S-PG5

  Are you sure? yes | no

AngelLM wrote 09/21/2016 at 12:21 point

Hi!

Sorry for the late reply :( Yes, that ID is the same of the gearboxes I used :)

  Are you sure? yes | no

Dan Royer wrote 08/30/2016 at 22:44 point

https://hackaday.io/project/12971-robot-overlord/ would be delighted to add your robot to the 3d controller.  it already supports several arms.

  Are you sure? yes | no

AngelLM wrote 09/04/2016 at 18:50 point

I was looking for a GUI to control Thor, this sounds great! I'll take a look :)

  Are you sure? yes | no

Dan Royer wrote 09/17/2016 at 18:12 point

Please call or email me through marginallyclever.com.  A brief call would be enough to explain what I need from you in order to add the support.  I'll do the work, I already know the code.  I just need some dimensions and a few tricky STL files.  (Easy for you, you already have everything in 3D.)

  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