Bionic NXTPod 3.0 – the Gripper

3 06 2011

Creating a gripper really slowed this project down, since I only had one motor port available for it – each NXT needs two for its linear actuators and the master needs one for the NXT-PF cable for communicating using the LEDs, leaving only one port on the slave. Since the weight of a motor turned out to be too much for the moving part of the structure, I used pneumatics instead. the problem with that is the need for at two things: a switch operator and a pumper – two motors.

The following image cannot be displayed: Gripper's multi functional motor - operates both the switch and pumpThe solution came in the form of a differential, which contributes the input power applied to it to one of two outputs depending on which one has a smaller resistance (check this 1930s tutorial for a much better explanation). So when the motor starts turning, it first moves the least resistant part – the switch. Once the switch is all the way to the opposite side, it blocks that output on the differential and makes the motor move the pump – when the motor is spun forwards, the actuator goes up, and when it goes backwards, it goes down:

The following image cannot be displayed: NXTPod's gripper module - motor and actual grabberThe gripper itself is pretty simple, and opens/ closes when the actuator moves the middle part up, pushing the two “fingers” to opposite sides. Its size allows it to grab either the blue or red ball that come with the NXT 1.0 kit.

The following image cannot be displayed: The actual gripperDisclaimer: This site is neither owned nor endorsed by Festo Group. The Bionic Tripod 3.0, on which this project is based, is copyrighted by Festo.





Bionic NXTpod 3.0 – the actuators

20 05 2011

They’re not as compact as the LEGO Technic or Firgelli ones, but these actuators should do their job just right. They consist of a 16 bricks long axle with a freely moving worm wheel on it, which moves along a series of four racks. When the motor is spun, the gear wheel goes up, pulling or pushing the attached pyramid leg along with it. With all of the stuff built around it, the actuator can move up to 10.5 Technic bricks up and down, in about 10 seconds.

The following image cannot be displayed: One of the NXTpod's four actuatorsThere’ll be four of these controlling the entire NXTpod, and probably another motor in the middle powering the pneumatic hand on top. More coming soon!

Disclaimer: This site is neither owned nor endorsed by Festo Group. The Bionic Tripod 3.0, on which this project is based, is copyrighted by Festo.





Bionic NXTpod 3.0 – Research and Tests

2 05 2011

Festo, founded in 1925, is a German engineering-driven company based in Esslingen am Neckar. Festo sells both pneumatic and electric actuators, and provides solutions from assembly lines to fully automated full automation solutions utilizing Festo and third party components. It also has a kind of R&D department, the Bionic Learning Network, where they’ve created some amazing projects including SmartBird (“bird flight deciphered”), AquaJelly, Robotino XT and much more. [source]

They’ve also created the Bionic Tripod 3.0. This arm-like machine is driven by four linear actuators, which move semi-flexible cables (“steel spring rods”), to which the grabber is attached. The video below is from Festo’s official YouTube page, and explains the tripod much better than I just did:

Now what I’m going to do is make it out of Mindstorms NXT.

The rods

… are probably going to be one of the hardest parts to this build. The ones that Festo used are flexible, strong and don’t stay bent. When I went digging through my Lego Techinic, I only found three, long flexible parts! So I compared them to find the best one.

The following image cannot be displayed: Will they stand up?This first test was to determine which of the three would stand up best, carrying its own weight. The pneumatics cable came out worst, which, unable to stand up, wouldn’t work for this kind of application. The other two were very close. The right one was pretty strong, but it’s also thick and tends to curve a bit too much. The left one, however, stayed pretty straight (the top one was bent before), especially the lower one.

The following image cannot be displayed: Will they bend?This test was to see just how flexible the rods were. Too flexible wouldn’t be strong enough, and very rigid wouldn’t be… flexible enough. This time, both the right and middle ones lost, because they bent far too much to be able to support a lot of weight. The left one won again.

So, I turned out picking the left one. The only problem was that I only have two of them. So, I went to this store on BrickLink, and got twenty more.

The linear actuators

… might  cause problems as well. Since I don’t have any, and don’t plan on buying any in the near future, I needed to design a linear actuator from scratch :(. I ended up with this design:

The following image cannot be displayed: Linear actuator designBasically, the rods will be attached just above the worm wheel, which, when spun, will either push it up or down along the gear racks to the side. It isn’t very strong yet (the gear racks tend to break), but it works fine. Also, The back part of the motor is free, so the four motors should be able to attach to each other fairly easy. The other ones aren’t done yet, because I’m still waiting for materials.

The next post on this will probably be in a few weeks, when I have the first prototype done.

Disclaimer: This site is neither owned nor endorsed by Festo Group. The Bionic Learning Network, SmartBird, AquaJelly, Robotino and Bionic Tripod 3.0 are all copyrighted by Festo.





Easter Egg Sorter (EES) – Hardware

1 05 2011

EES is almost finished!. It just needs sensors and a few changes to the gearing and connections, so it runs a bit smoother. Therefore, the pictures below might not look exactly like its final version.

The following image cannot be displayed: Front view of EESThe idea behind EES is that easter eggs will be fed to it by the conveyor belt. At the beginning of the belt, a color (and possibly light) sensor will give EES information about the egg, so the arm knows when to grab it, and where to sort it. The egg is picked up by EES’ grabber (not entirely finished yet), which is controlled by a PF motor, after which four NXT motors bring it in the right position, so the grabber can drop the egg. The whole process between grabbing and dropping will theoretically take about 6 seconds, so, optimally, the robot should be able to sort 10 eggs per minute.

The following image cannot be displayed: Zoom on bottom EESIn earlier robot arms, I had the motors very close to the part they were supposed to move (shoulder motor in the base, elbow motor in the lower arm, wrist motor in the upper arm, etc). Since that never worked too well (it got too heavy for the lower motors), this arm is very lightweight, and almost all of the motors are below shoulder. This way, everything should be able to move faster and smoother.

The following image cannot be displayed: Upper gearing and rotation transmission EES The problem was that those joints still needed power, and the solution came in the form of universal joints – 6 in total. The great things about those kind of connections is that they can move the rotations at any angle lower than about 35-40 degrees. That way, the joints were able to bend individually and smoothly. Lego universal joints and gears in general do have one big problem though, especially when a lot are used in line, which is backlash. So, it might be hard to get the robot to move to exact positions.

More info coming soon!








Follow

Get every new post delivered to your Inbox.

Join 34 other followers