Blog #6 - Mechanical Build P1 - Robot On Rails

1. Introduction

     Hi! This will be my sixth blog for the 1 Meter of Pi - Design Challenge. In the last blog which you can find linked above and at the end of this blog, I covered my design and build of the Light & Sensor module. In this blog, I will finally do the main mechanical build, the first part of it. The main mechanical build will be a 2 part series. In this first blog, I'll cover the rail system that the robot will ride on and everything relevant to that, while in the second blog I will go over mounting my modules, LED lights, water storage and everything else needed. That blog will come at a later date when everything else is completed and can be mounted.

2. Idea

     Let's start with the name of this blog, Robot On Rails, what kind of rails and why do we need rails. To answer this question we need to go back to my first idea for this project and general concept. The idea was to have a robot that could move up & down between two big essentially pots on both sides of it. So we have the question here on how to solve that movement problem for our robot. Let's take a look at the original sketch first.

     On the picture above you can see what I envisioned. Some kind of a rail system or something else that would allow our robot to go up and down between those 2 big pots. There are 2 things to think about when we are talking about a mechanism like this, what will propel our robot and on what will our robot move. These 2 things are of course dependant on one another, here are a few examples of what we could possibly do for this project:

• Flat surface - wheels - onboard motors
  • This option would mean that we would have a flat surface and a small motorized cart that would carry our robot around
• Flat surface - wheels - belt driven
  
  • This option would again see a small cart carrying our robot but instead of it being motorized we would have a belt pulling the robot up and down this system
• V slot extrusions - wheels - belt driven
  
  • This option would see us using the 20x20 extruded v slot aluminum extrusions and rollers which would fit in those rails, while there would be a belt pulling our robot. This method for moving something can be seen a lot of machines like some 3D printers, which is where I got my idea from.
• V slot extrusions - wheels - threaded rod driven
  
  • This is the same setup, just instead of the belt driving the robot, we would have a long threaded rod turning and moving our robot. This method would give us a lot of torque because of it's construction, but would be really slow and we would need a really long threaded rod for that
• Smooth rods - linear bearings - belt or threaded rod driven
  • This would be a really professional setup but required really long rods, which would mean we would have to get some thicker ones to compensate for the sag, and we would also need linear bearings. As for the propulsion, we have the same options as for the v slot extrusions

3. Plan

     Those are just some of the options, we can certainly find more options which are better or worse at something than this here. The option with which I went is the V slot extrusions with small rollers that is belt driven. There are multiple reasons for this. The rails are pretty strong, so at 1m there won't a lot of sag with them, all of the part can be found easily or made easily using 3D printing, those rails aren't expensive. As for the propulsion, I will be using a stepper motor which will drive the robot by pulling a belt. This gives us pretty good accuracy and much more speed compared to the threaded rod. Also, these extrusions are like LEGO and I can mount other stuff to this frame later on in the project.

     All that's left to plan our are the measurements for our rail system. I wanted to have 1m of movement up & down the rails, while the distance between the rails will be determined by the gap between the wheels when we mount them to the robot box. So I'll go over that in the assembly section of this blog. Let's first design all of the mounts and brackets needed to assemble everything. There are 3 main assemblies we have to cover, the wheel assembly, the belt & motor assembly and the leg assembly. Besides that, we have a couple of things that will be needed for the assembly which don't really fall into one of the main 3 assemblies.

4. Design

     Now we get to the main work for this part of the project, the design. To design all of the parts I'll be using Fusion 360 and will be printing them out using my Ender3 Pro using Creality White PLA. For all of the prints I will be using standard settings with a 0.4mm nozzle at 0.2mm layer height and will be using Cura 4.6 for slicing all of the models (I know there are newer versions available, but since there were a lot of bugs in the 4.7 I'm not upgrading until I'm done with this project). You can find all of the 3D models here: Rail System. The models come as they are and you are free to do with them whatever you want! To begin the design, let's start with the wheel assembly.

Wheel assembly

     This assembly will go over the parts needed for the robot to slide on the rails, this will include the wheels for the robot and the brackets that are used to mount the wheels to the robot box. The wheel assembly will consist of 6 wheels, 3 on each side, with 2 being on the top and 1 being at the bottom of the rail. In this way we have a really stable construction, we can turn it upside down and it would still work. To begin, let's start with the wheel. There is a subfolder for all of the models for this assembly, you can find that here: Wheel Assembly.

Finished wheel assembly

     With all of those printed out, the only thing that is left to do is mount it to the robot box. I mounted the 4 upper wheels at the corners to get the best stability, while the third one went in the center between those two. If you've missed it above, all of these models are available on the following link: Wheel Assembly. You will find both the STP files and STL files there, you are free to do with them whatever you want, the models come as they are without any guarantee, but they did work for me.

     The wheel assembly turned out great! I was afraid if I got the measurements right, and they are pretty pretty close. The lower wheels could maybe have been just a tiny bit close for a better fit. But it glides down the rails great like this and there's no way it can fall of them. To attach the mounts to the box I just drilled some smaller holes and used a tap to thread the holes in for the M4 screws. The weakest part of this build is the plastic box itself, which I found it flexes a bit, but it is holding up really good.

Belt & motor assembly

     This assembly will cover the propulsion part for this build. This includes the motor mount, the belt tensioner and belt grabbers. The gears I have on hand are both over 20mm in diameter which means I can put them outside the V slot extrusions and still use them no problem. In this way, I can use the maximum length of the rails for the robot to travel on them. I had a bit of a different plan in the beginning for mounting the motor and belt tensioner, and that would be to a piece of wood that would act as a wall. I changed along the way to have a frame made out of the v slot extrusions, so I had to redesign 2 mounts. I'll still upload the files for those mounts I won't using since they may be useful to someone. You can find all of the models for this assembly here: Belt & Motor Assembly.

Leg assembly

     The last big assembly for this part of the build is the leg assembly. Since the robot will have 2 wheels underneath the rails and we will need 2 cables running underneath the robot, we need a way to somehow lift up the whole construction. My solution for that is printing out 4 screw on feet with height adjustment, so I can make sure that everything is level and that there is no wobble in the whole construction. I've designed these legs so they can easily be redesigned so that they can be screwed onto something or just mounted differently in general. You can find all of the files for this assembly here: Leg Assembly. As usual, you have all of the STP and STL files available and are free to do with them whatever you want.

Finished leg assembly

     Each set of the legs took a little longer to print. In the slicer I put 4 outside walls to make it stronger and to make the wall around the 8mm hole thicker. This is important to have meat into which we can cut it when we are tapping the hole. Each leg, by that, a leg and a foot, needed about 2 and half hours to print. Here is how they look printed. The 8mm screw is used is 50mm long.

Other

     Besides the parts and assemblies I've talked about above there are 2 more parts necessary for the final assembly that don't really fall into of the assemblies above. Those are the mounts for the end stop switches and L brackets used to put the whole frame together. These are simple and small prints, but I made the walls on the L bracket extra thick to make it as strong as possible. You can find STP & STL files for these models here: Other 3D models.

5. Assembly

     With all of the parts designed and printed out, the only thing we have left to do is assemble everything. The first thing I assembled is the wheel assembly for the robot. The reason why I did this so I can know the exact measurement to which I'll need to make my frame so the robot glides nicely on it. Here are the pictures of the wheel assembly which you could have seen above as well.

     After this part of the assembly was finished, it was time to turn my attention for putting together the whole frame. The wheel to wheel distance came to around 178mm, so that was the dimensions that I was aiming for when putting this together first. Here are all of the fasteners and parts needed for this part of the assembly.

     On the left we have the L brackets, screws, washer and nuts needed for them, in the middle we have the motor assembly, and on the right we have the belt assembly. To make the frame, I used 4 pieces of 20x20 extrusions, two with the length of 993mm, and 2 ones I had laying around that are 280mm and 425mm. I didn't shorten those to the same length, as it gives me more room for mounting things later on. As for why the length of the rails is 993mm, the guy I ordered them from cut them to that size on purpose, since, if the package is above 1m, the postal services charges way way more for shipping where I live.

     Above you can see some of the close ups of the construction. On the first picture you can see the frame held together with the L brackets and the belt tensioner doings it's thing. On the second picture you can see the legs mounted to the construction, the height adjustment works great, and on the last picture you can see the part where the motor is. Next to it, you can also see the limit switch. Here is a big picture of how the whole construction turned out.

     I am so happy with how third turned out. In the picture, the legs are at their shortest setting, and the lower wheels are still clearing the ground by a couple of cm. They can go much higher if needed. After doing the first test assembly, I took the whole thing apart and with help from my dad, we slowly put it together to make sure everything was aligned properly and centered. After that assembly, it turned out great! The only mistake I made was tightening the belt too much, but after loosening it up everything went smoothly.

     Here are 2 gifs of the first gliding tests I tried. On the gif above, it's just me pushing the cart up & down the rails. It's really going really nicely. The only thing you can hear is like it's climbing a small step and that's because of the line where the printer met at that layer, so there's a bit of a seam, but nothing too dramatic.

     On this gif we can see our astronaut/artist/gardener enjoying his first ride on the rails, and he's pretty happy with the ride. The only thing to do is to try out the whole system by running the stepper motor. I'll wire up a small test circuit with a step driver and see how it moves. Let's first take a look at what we will need from an electronics standpoint.

6. Electronics

     As mentioned multiple times, I will be using a stepper motor for this part of the build. As you know, driving the stepper motor isn't as easy as just plugging it into power and it turning, but, to be honest, it's not much harder than that using small stepper drivers. They became really popular and low cost specially because of 3D printers using them. There are multiple options for widely available stepper drivers, here just a list of the basic few:

• A4988
• DRV8825
• TMC2208

     They all do the same thing for the most part, driving a stepper, but they are rated for different current of course and have additional features compared to one another, like silent stepping, sensorless homing where it measures the current going to the motor to home and it so on. The one I have and will be using is the DRV8825.

Here is the datasheet for the driver that is on this breakout board as well as a link going into a bit of a detail when it comes to using the module:

• https://www.ti.com/lit/ds/symlink/drv8825.pdf
• https://www.pololu.com/product/2133

     Driving this stepper driver is pretty simple using a small microcontroller. On one side of the board we have connections for the motor and the power supply for the motor, while on the other side we have the control inputs. We have reset, enable, sleep signals, we have inputs for microstepping and most importantly we have the step and direction pins. By sending a pulse to the step pin we can turn the stepper motor, while depending on the logic level of the direction pin, we will be turning it either one way or the other. Here is a simple schematic of this breakout board.

     The connections between the components are simple as you can see from the diagram above, all we need are 2 digital pins from the Arduino Nano. I haven't played with microstepping features of this breakout board yet, but I'll look into that in the future. The power supply pin for the motor accepts a wide range from 8 to over 40 volts, but I'll be going with 12V so I can use that at the rest of the project. All we need now is a short piece of code and we can finally test it out.

#define dirPin 2
#define stepPin 3


void setup() {
  // Declare pins as output:
  Serial.begin(9600);
  pinMode(stepPin, OUTPUT);
  pinMode(dirPin, OUTPUT);
}


void loop() {
  // put your main code here, to run repeatedly:
  Serial.println("Direction 1");
  digitalWrite(dirPin, HIGH);
  
    // These four lines result in 1 step:
    digitalWrite(stepPin, HIGH);
    delayMicroseconds(200);
    digitalWrite(stepPin, LOW);
    delayMicroseconds(200);
  
  
}

     As you can see, the code is incredibly short, the only thing I did was send a pulse to the DRV8825. The frequency of that pulse signal determines how fast our motor spins. All that was left now was to connect the motor and see it working.

7. Testing

Edit media

Dimensions x Small Medium Large Custom

Subject (required) Brief Description Tags (separated by comma) Video visibility in search results Visible Hidden

Parent content

Project R.A.G. - Blog #6 - Mechanical Build P1 - Robot On Rails

Poster

Upload Preview

Ok

Cancel

We have the motor spinning, let's just see how it works when we have it all mounted to our rail system.

Edit media

Dimensions x Small Medium Large Custom

Subject (required) Brief Description Tags (separated by comma) Video visibility in search results Visible Hidden

Parent content

Project R.A.G. - Blog #8 - Mechanical Build P2 - Robot On Rails Project R.A.G. - Blog #6 - Mechanical Build P1 - Robot On Rails

Poster

Upload Preview

Ok

Cancel

And here we can see the stepper moving the cart along the rails. During testing and preparing for another video, I managed to short something out on the motor driver so I had a bit of a situation, as you can see on the following photo:

     In other words, I'll be needing another motor driver. Besides just sharing all of the things that went right, I think it's important to also show some of the stuff that went wrong, like here. I'll just order a new (or 2 to be safe) and continue on with this build. To wrap it up, here is a picture of our astronaut mounted on the rails.

8. What's next?

     Next on the line is the main box which contains the Raspberry and the relays for powering all of the modules. Besides that I'm working on another module similar to the Light & Sensor module that I already covered, this one being the Water & Rail module. As the name suggests that module will take care of watering the plants and moving the robot up and down the rails. It will also have an important role when it comes to the tools on the top of the robot as well, but more on that when I get to that blog. When those 2 modules are completed, all that's left to do is mount everything together in this project, where comes in play the second part of this blog and then we can play around with software and program it to work with us. The only thing besides that, that is left is the tool holder and the tools themselves, which is a one day job on the design and production aspect. A lot of cool stuff on the way, though, the deadline is approaching really fast.

9. Fun news from the ISS

     In this segment of fun news from the ISS, I came across a cool little experiment that was done a few years ago, or to be more exact, onboard the expedition 23 mission which took place in 2010. They experimented with growing white spruce seedlings onboard the ISS and got some cool results as you can see below. As the other similar missions, the goal of the experiment was to see how gravity, or in this case, the lack of, influences growth and genetics of wood formation.

10. Summary

     This was the part of this project that I was the most worried about besides the software side. This part is essential for this project to have a point, because if the arm can't move down these rails, we can't really reach that many plants, the only thing we could to was to put the robot in the center with plants around that it can then go and work around them. These v slot extrusions are really an amazing thing, they are like a grown up version of LEGO for making stuff. I can't wait to finish up the main box to play around with this part of the build by controlling it from the Raspberry. All of the relevant links for this competition, link to the next and previous blog as well as my GitHub where I'm uploading all of the models can be found underneath. Thanks for reading the blog, hope you liked it and found it interesting!

Milos

Relevant links for the competition:

• 1 Meter of Pi
• 1 Meter of Pi - Design Challenge
• 1 Meter of Pi - Meet the Challengers

Link to my GitHub where you can find all of the files used for this project (code, 3D models,...):

• https://github.com/MilosRasic98/1MeterOfPi

Link to my Project Collection:

• Milos Rasic - Project Collection