Weight Plate Loader

Earlier this year, around August, I built a full-scale prototype weight plate loader, out of my interest in weightlifting. I was tired of manually loading weight plates so I did some research to see if there was anything to load plates automatically and was surprised to find that no one had created anything to load weight plates onto a barbell. Weightlifting is a sport that's been around since the first Olympics, so I was surprised to see no innovation for loading plates.

The only thing that existed was a barbell jack, which only lifts the barbell, working a bit like a lever. Even with the barbell jack, the loader needs to bend over, lift the plate and physically put it onto the barbell. Below is a video of how a barbell jack works.

The first step was to get a design up and work out how I was going to lift the barbell. I immediately thought to simple machines and thought to use some kind of lever. I was working on this project at my parents' work which is a warehouse with all sorts of tools and material I could use. It's called CNC Shop and basically all my parts except for the elctronics were from there. I was also under a time constraint, as I was entering this project into a competition and started a bit last minute. I started building and proofing my ideas only about two months before when this should've been a 6 month project.

There are 5 main weights of weight plates, but I decided to use only three (5kg, 10kg, 20kg) because this needed to be done quickly and it was only a prototype. I also selected these weights specifically because they had the most weight combinations possible.

I had some ideas about how I would store the weight plates after months of research, but I hadn't actually started trying things out physically or looking at materials. I learnt heaps from this project. A lot about physics, how motors work and simple machines, but I also learnt how to use heaps of tools and how to apply my ideas to real life and make them happen. There were certainly many mistakes on the way. I broke many taps, getting a lot of them stuck in the extrusion, broke drill bits, made mistakes with the CNC machine, broke the lathe at one point and many other things I don't need to list.

I learnt how to:

- a lathe

- a CNC lathe

- Designing on Fusion 360

- A bandsaw

- A CNC router (touching up on this one)

- rack and pinion

- use a Raspberry Pi

- Wire up drivers to motors

Other skills I touched up on inlcude programming Arduino, using powertools, soldering, wiring stuff.

The Barbell Lifter

Getting the barbell off the ground was no easy task. It had to be strong enough to lift and hold 105kgs. I decided to go simple and just use a simple first-class lever (which is just like a see-saw).

I spent a long time researching different options to actually lift the barbell. I looked into pneumatic, hydraulic and electrical systems for lifting the barbell. In the end, pneumatic was too weak, hydraulic was very expensive and I had zero experience with these systems.

I selected an electrical system because I already had prior knowledge and equipment with electricity, from my extensive background in robotics. Given no time constraints, I might've gone with a hydraulic system as it is very strong, but I also needed ease of control and precision for this part of the machine/robot. The barbell needed to be lifted to an exact height, otherwise the weight plates wouldn't line up with it.

I used a stepper motor, partially because it would work great for the job, but mostly because I saw it was already connected to a ballscrew (which is the stick connected to the motor on the top of the leaning thing on the right). A stepper motor was more expensive, but it was easier to control. Unlike DC or AC motors it doesn't need a limit switch to stop it. The motor also needed a driver to control the movements, so that added to the cost. All of the motors except for 3 of the motors were NEMA 23 motors. The motors driving the y-axis are NEMA 34 motors (from memory) as well as the motor driving the barbell lifter.

The video below shows a detailed explanation for how a ballscrew works, but if you want a simple explanation, it is just a screw that turns rotational movement (ie. the spinning of a motor shaft) into a linear, straight-line movement (up and down movement of the barbell).

I had a few problems with the Barbell Lifter, such as there not being enough range of movement during the testing stage and the motor being too weak to lift the required weight, but more of that later. The lifter was made of steel, welded together which I chose because it's strong and needed to handle heavy weights. There is a bearing as a pivot point, because it was smooth and was able to reduce friction.

Weight Plate Holders

The weight plate holders are made of aluminium extrusions, CNC-cut acrylic plates, steel shafts and trapezoidal screws.

I used the aluminium extrusion for half the frame and weight plate holders, because it was easy to put together, compared to wood or acrylic, and could be adjusted easily. I needed it to be adjustable because I was still testing and little things needed to be tweaked here and there. The extrusion slots in with each other, a bit like how Lego all fits together. I still had to cut and tap (thread) the ends of the extrusion but it was much easier to use that than if I used steel.

I used acrylic panels for the other half of the frame, because I needed a lot of holes. I first measured where I needed holes, then designed the panel on Fusion 360. I had to learn how to use Fusion first, then started designing. I've found it a very powerful program to use, much preferable to TinkerCad I had previous experience with. The panels needed to include holes for the motor to attach and there were 3 motors. I cut it out of thick acrylic on the CNC machine.

the gallery below shows the process of creating the acrylic panel part of the frame.

I also used acrylic for the L-braces attaching the T-shape weight plate holder to the steel shafts. They were also designed on Fusion 360, just cut out of thinner acrylic. I had a few prototypes, after making a few mistakes with the CNC machine. For example, the hole being too small to fit the shaft, the hole being too big, forgetting to put a hole in for the trapezoidal nut. But I persevered.

This is one of my faulty L-brace designs which didn't have a hole for the trapezoidal nut.

Below is my mass production of the CNC cut L-braces. It actually took quite a while to do this because each linear bearing had to be screwed into the brace with at least 2 screws, each screw with a nut on the other side. Then I ran out of nuts which was a problem, but managed to get them done.

The steel shafts were for the l-brace attached holders to slide left and right on. I had two steel shafts per weight plate, for stability, which meant a linear bearing to go on each shaft. So there were 6 bearings in total per frame.

The stepper motors driving the sliders were actually from an old 3D printer and there were just enough for what I needed so that saved a bit of cash. These motors' shafts were connected to the trapezoidal screws via a coupler I made myself. There weren't any of the size I needed, and didn't have time to buy any, so I learnt how to use a CNC lathe to cut and drill the couplers to the size I needed. So in the process I also learnt how to make my own couplers. They worked out well.

Assembling it all together:

MDF Track

The MDF track is essentially the Y-axis which drives the weight plates onto the barbell.

There were a few options of how I could do the y-axis, and deciding which idea to go with was probably the hardest part. In the end, I went with the track idea because it was just a track that the frame could roll on. The most important part was that this was detachable, and you could take it off. All my other designs were fully attached, making it hard to transport and move. This thing was really heavy, without the weight plates, so being able to detach it was important. The track is just two 1cm wide lines engraved in the mdf, using a CNC router once again. Because this was a simple, easy design, I used ArtCam to design it, which was just more convenient.

I attached bearings to the bottom of the frame, which acted as the wheels.

The driving force also came from a stepper motor, connected to a rack and pinion, which allowed it to be detachable. This idea ended up working great.

Below is a very zoomed in shot of the bearing 'wheel' in the MDF track, held in with a steel brace I cut using a bandsaw.

Wiring and programming

The wiring part took a while, just because each motor had to be hooked up to its own driver and there was 9 motors, each with four wires. I probably spent over 8 hours just wiring. It was pretty wild and intense.

Below is my wiring diagram just to get an idea of the amount of wiring I had to do.

I soldered where I could and after checking each wire was in the correct place, I then had to connect them all to the PiLeven I was using which went on top of the Raspberry Pi 3 I used. I used the Arduino software to program the robot. The only part not done on the Arduino was the User Interface(UI). For this, I used Python because Arduino didn't have anything to use with a touch screen and could only be done in Python. I actually had no experience with Python before this and learnt, using code from the internet that people had already developed and just modifying it slightly for what I needed. I then sent the response from the user to the Arduino software to tell the robot what to do. It was a bit complex but worked in the end.

What the touch screen looked like. I had no experience with one but thought it would be cool.

Testing

I tested as I built and as soon as the whole thing was finished, I found many issues with it. It is still a prototype, and worked pretty good for that, but had some practical issues. Some obvious ones are the large size and heavy weight. It is cumbersome and bulky, making it not the most practical. The dimensions are approximately 2.4mx1.2m for the base. Each weight plate holder alone is over 15kg each, so is extremely difficult to transport and move.

Awards won

I won a major bursary in the Science Talent Search in the Inventions category and was selected as a finalist in the Australia-wide BHP Foundation Science and Engineering Awards competition.

Acknowledgements

I got feedback from multiple people including coaches, loaders, potential users, gym owners and a scientist.

I would like to thank Josh Balia, Dr Eric Drinkwater, Adam Kabbas, Matthew Williams, Wayne Zhou for the time they took to review my Weight Plate Loader. I would also like to thank my dad, Dan Sriratana for teaching me how to use all the tools and materials. Also thanks to my family, friends, teachers and everyone else who supported me. It means a lot to have people backing me.