3D-printed RC-Car [*WIP]

15/07/2020

*WIP = Work In Progress 

Foreword

This project has been on my mind for quite some time, but has unfortunately been left behind due to its level of difficulty and demand. Procrastination [en.wikipedia.org], is very strongly present in every challenging project, and it couldn't be avoided in this project either. However, when you remember to consider all the things you'll learn during making of a project, you have a lot more motivation right away. In many cases, to start a project it's enough to shift your gaze from failures to new things and learning experiences that come up during the project

Here's a dictionary for the previous image:

  • Suunnitelma = plan
  • Toteutus = implementation

Additional attention to the following links! Almost every link on this page is an affiliate link, which means I get a very small amount when you buy products from those pages through the link! So if you want to support the site and you are about to buy that product anyways, then by buying through these links, you are also supporting my site without paying anything for it! If you absolutely want to avoid supporting this site, I recommend avoiding those links!


The goal of this project


Every project should have a some sort of goal to work towards to. It's important to set a time limit, because otherwise the project tends to stretch indefinitely because you didn't set the time for its completion. I have noticed also that for hobby projects, it is a good idea to have first really low expectations, in order to get it started. For some strange reason, Newton's first law also seems to apply to brainwork too, as it is necessary to just start the project immediately and not to be left to the state of considering about starting it. If you manage to do this, the project is already afoot and further work on it is so much easier.

First goal that I set for this project: 

  • Make a forward moving car, that turns with a servo motor.
As you can see, the goals for this project have been set very low. However, it is important that low goals do not remain as the only goals in that project. In this case, you would to see the situation where a plexiglass with two wheels and a servo, which you push forwards and backwards by hand, achieves all the goals you have set and further development seems to be a sheer waste of time. So we set the final goals for the project. Here, however, care must be taken that the outcome does not seem too difficult to implement, while ensuring that even a little new things have been learned when the outcome is reached.

The final goals I set for this project:

  • At least a speed of 40 km/h [24.9 mph] (while maintaining driveability and stability)
  • The car moves forwards and backwards, turns and brakes.

The progression of this project

Selecting the parts

The first step was to choose a suitable platform for the first version. The choice was very simple, as I only had two 300mm x 300mm x 6mm plexiglass sheets on my shelf that I bought from Biltema a long time ago. Next I printed the bracket for the rear motors. This version has two brushless motors at the rear of the car that are directly attached to the tires. This seemed to be a much easier implementation than designing a fixed rear axle. But in this design, it is good to keep in mind that the two non-sensored engines will rotate at different speeds, so the car may not drive straight. However, this simplifies the project in the beginning by using two engines.

The next problem is that there are only four brushless aircraft engines on my shelf. Two are the well-known A2212 2200 kV motors [eBay.com] (I don't recommend to use it for non-aircraft use!) The rest two are the slightly lesser known Alomejor C4250 550 kV motors [Amazon.de] designed primarily for electric longboards. (The description on the C4250 engine are just pure lies, but the product still seems to be a high-quality and affordable non-sensory engine.) 



To make the project even more challenging, there are only two HW30A Drone / Quadcopter ESCs [eBay.com] left at the bottom of the shelf. These ESCs are designed entirely for aircraft use and can't do for example reversing, let alone handle the voltage of an 6S Li-Po battery. Fortunately a 3S Li-Po battery was left over from a previous project, which works well with that ESC.

To steer the car, we need a controller, which I fortunately bought some time ago when I was thinking about this project. However, the Flysky FS-I6 [Amazon.de] is not designed for RC cars (it's for airplanes), but because it was cheaply priced, I decided to choose it, and try to make it work as a car controller with moderate effort. I also recommend buying rechargeable AA batteries! They are easy to change from one device to another, and when the battery runs out, you can put them into a charger rather than the hazardous waste -bin!


Oh yeah! I almost forgot the servomotor! Fortunately, I had these on my shelf laying around! So what I used is the MG996R servomotor [Amazon.de] with an incredible amount of torque and speed! I can definitely recommend this servo for all projects, because I think the price is really cheap and the quality of the motors does not disappoint anyone! Each servo comes with a load of accessories like horns, screws and even vibration dampeners! I used to make a pretty big mistake when using SG-09 servos in almost every project, because I thought that I had bought them already and it would be waste of money not to use them. If this situation feels familiar, then you are not alone. It is a so-called 'sunk cost fallacy', which everyone will encounter at some point in their lives. This phenomenon therefore means that decisions are made on the basis of past losses and not on the basis of how the decision would affect your future. It's good to be aware of this phenomenon so that from now on when it comes up, you won't waste all your time and money. I recommend watching Better Than Yesterday's video about this subject if you are interested in that phenomenon. [YouTube.com]

The first version


Great! The parts are now selected and we should start designing the brackets for the parts to keep them attached to the base. (And I know that you are thinking about it but no, Duct tape doesn't count.) Fortunately, we have 3d-printers that help making our ideas to reality in an instant. Thanks to my ultimate laziness, a very innovative 90 ° corner piece with a pair of mounting holes for the plexiglass and motor came out of the Fusion 360 in about five minutes.


Warning: unneeded theory ahead!

However, when I got the corner pieces out of the printer and made a little test drive with two tires, I noticed that the ESCs were heating up a lot (didn't pass my touch test), and I switched the A2212 engines to the C4250 engines immediately. This is because the kV of the A2212 is 2200, which means that the motor is designed to reach really high speeds even with a small battery. (With a 2S LiPo battery, this does at about 16280 rpm when calculated at rated voltage.) Brushless motors are really bad at low revs (especially without sensors), and cause a lot of strain and load, especially when the gear ratio is 1:1.

The C4250 motors have a kV rating of 550, which means that it does't get to such high revs with the same voltage provided (2S LiPo with a nominal voltage will get max. 4070 rpm.) and it'll get more torque out with lower revs. The larger diameter of the motor also helps to produce more torque, as the arm of the force is considerably longer compared to smaller motors.


Second version

Now that the motors are replaced with new ones, new brackets must be designed to keep them firmly attached to the plexiglass. After a few iterations, a very interesting looking part popped out of Fusion 360. With the first version, I didn't remember (again) that the screws might have some sort of heads and might collide to each other. But fortunately for us, a 3D-printer spits out new parts in the blink of an eye, and you can't do anything than learn from the mistake! So what did I do differently? I moved the base plate down a bit, and modeled the M3 screws with small counterbores (I hope that's the right term for the holes that allow the M3-screws to go a bit further down) so they wouldn't stick to the M4 screws that go through the plexiglass plate! 

Take note! If you want to test the construction and improvement of the prototype yourself, here is a link to the Fusion 360 file on the rear axle bracket! [a360.co, Autodesk's own distribution platform for Fusion projects.] Note! From that link you can also download that part in STL, STEP, SketchUp, Inventor2019, etc. format! Just click Download in the upper right corner and select the desired format for the file! 


The root of all causes, that is, the turning mechanism


Now that we got the transmission problems figured out, it's time to focus on turning mechanism. This feature is really the root cause that made the beginning of the project stretch with the assumption of 'turning is really hard to accomplish'. However, when I spent about 30 minutes on Google, I realized quickly that the turning mechanisms aren't terribly complicated in any way. From this moment it took something like an hour, after which a working model was already done. And again, I'd wish I had done this sooner. A lot of times you just tend to make a sheep into a wolf in your mind, and therefore you don't want to touch it.

Take note! If you want to take a look of the 3d model of the turning mechanism, you'll find the first version from the upcoming link! Note, however, that this is the first version, not the final one! [Fusion 360 shared project link, a360.co] 

As it can be seen quite well from the picture, the turning mechanism is not terribly complicated. The only downside to this design is the the inconvenience of assembling it. Generally, this type of turning mechanism uses a bolt that extends completely from top to bottom. However, I designed the turning part so that it has 3d-printable threads above and below, so all parts have to be fastened at the same time, and the pivoting part can't be removed without dismantling the entire system. In the next version, using a bolt would be a necessary change I'd make. You've probably also noticed the massive size of the mechanism compared to the tire. This is because I had only 8 mm x 22 mm x 7 mm bearings [Amazon.de] laying around, and those crappy small tires were found on the same shelf that came with the motors of the 3d-printed Mecanum-wheeled robot (previous project of mine). Those tires and motors should be avoided like the plague in this kind of projects! (If you still don't believe my warning, I embrace that the motors are bad because of the weak, poorly designed shafts, weak reduction gears, and motor inefficiency.) Sure, they can work on lighter systems, but there are a lot of better options!)

Warning! The next two paragraphs include a large amount of weighing between different ESCs and tires!

And back to the topic. So the next step would be to buy new, bigger and more durable tires and redesign the turning mechanism. Also, one problem I mentioned earlier, namely the lack of a reverse, is one of the biggest problems that needs to be fixed. Since the current ESCs are not able to reverse, the project will require the purchase of new ESCs. I have tried to compare different kinds of controllers, which would be able to handle the 6S LiPo battery pack's voltage, slow down fast (intentionally), and even to change the direction of motor rotation, when needed. When buying from the Hobbylinna (Finnish seller for all things RC), the price level would have been in the order of about 150€ per ESC (and you need 2 pieces, ie a total of 300€), which makes no sense. With that money, you could buy two ODRIVE V3.6 56V controllers! If you are not familiar with ODRIVE, it is one of the most advanced motor drivers available on the market today. And on top of all that, it is capable of running two motors simultaneously, using different encoders and even regenerating braking energy! But still, because my budget is minimal, those delicacies go unpurchased and I look toward eBay pages. The cheapest ESC I found costs about 15€ and, according to the description (which is usually not worth relying on too much), is able to handle the voltage of a 6S LiPo battery. However, the product information does not mention whether that ESC is capable of driving in reverse. However, the title says that the ESC is suitable for driving a 1:8 RC car, so one could assume from this that it 'should be able to do it'. That ESC can be found at this link if you want to take a look. As a result, I can say that I will leave this to think for a moment whether 30€ is worth for those Chinese controllers. However, the 24V version of ODRIVE would cost only 129.00€ and would also be suitable for future robot projects. Even if my bank account was empty after that purchase, I would still have a brushless motor controller suitable for robotics projects.

Even though there was a very negative atmosphere left over from the recent paragraph, there is still an urgent need to mention that the current tires are really bad. There doesn't seem to have any traction when accelerating and the car seems to 'push' at every turn. The result of the comparison is in the same category as with the ESCs, and 4 pcs of 1:8 wheels for the buggy were found on eBay. [Here's a link to eBay.com again!] The only thing that makes the tires a little questionable is the material of the rubber. The material should be really soft so that the car gets the best possible grip even with cold tires. However, 15€ is a bit too good of a price for four wheels, which again makes you wonder where it has been saved.

However, so that the project does not stay in place like the current version of the car, it is necessary to make updates to something. However, let's look at two short videos about the car's current state of moving on the carpet and floor to get a good idea of the situation of this project. I apology for the poor quality. 

As the videos show, driving is a bit challenging due to the lack of reverse, and it doesn't have quite enough of grip. However, the center of mass is already very low, albeit a little too much behind. It is likely that the pushing would be slightly improved by moving, for example, the electronics and the servo a little further. 

Another problem that can be seen from the video is the very rudimentary transmission for turning. Fortunately, it is easy to design new rods for this problem which transmit power from the servo to the tires. Of course, the pivoting parts had to be redesigned, as their attachment points were too far away and the pivoting was not sufficient at that time. The bearing will be (hopefully) made with my on shelf 3x6x2.5mm bearings, which should be good enough, but the small size brings some challenges to printing.

Uusi kiinnityspiste = new mounting point
Aikaisempi kiinnityspiste = the previous mounting point

But now, in order for the project to move forward, I will immediately order the new wheels. The situation of the motor controllers still needs to be clarified a little more, but I am afraid that they will not be available in an instant, at least cheaply. Unfortunately, again, I have to return to do an another chunk of my military service, which slows down the progress of the project for a while, but at least after that, the new parts should have arrived! 


The final version of the turning mechanism


The turning mechanism is still in development and I'm starting to run out of time. I'll try to get the final version done this morning. The current version works otherwise, but the adjustable part is a bit too short. I made the parts 5 mm longer, and that should be just fine for it to fit well.

New changes in the picture by number:

  1. New rotator with a mount for an interchangeable arm
  2. Interchangeable arm
  3. Helper linkage (to allow the mechanism to turn smoothly)
  4. Adjustable link to allow the adjustment of the steering angle

To my luck I got the turning mechanism finished at the last hours. You can find a video about the turning below. That mechanism was way more complicated that I thought and it took a lot more iterations than I expected. It still isn't the best it could, but at least it works now!