REVERSE ENGINEERED EASEL: PROGRESS III

November 17, 2021

REVERSE ENGINEERED EASEL: PROGRESS III

This week, I continued to reverse engineer the easel. I completed component E and worked a lot on component I, which is the wheel. The wheel is by far, the most complex part because it has ball bearings and injection molded plastic parts which are difficult to take apart without damaging them. Luckily, I was able to find schematics online which helped me create a realistic ball bearing without having to take mine apart.

In this blogpost, I will break down my 3D modelling process in Rhino into key steps. I faced many challenges measuring and replicating complex forms but I was able to work around them. I would have liked to finish the entire easel model by this week but I improved my Rhino skills and learned a lot about ball bearings.

1) Completing the 3D Model of Component E

Two photos showing different angles of component E

Completed Phillips screw from last week

Last week, I modelled a little phillips screw. This screw is used four times for the component that allows the horizontal canvas tray to slide up and down.

The exterior of the component is a continuous 1mm folded sheet of metal. A rivet holds four sheets of spot-welded metal inside the object. And the sheets of metal, measuring 1.5mm in thickness each, hold a spring in place. Out of the four sheets of metal, only the two in the middle have an extruding ring that you can pull with your finger. When you pull the ring, the spring compresses against the rivet. Though I didn't take out the rivet, I could peek at the inside of the component to figure out the shapes of the interior parts.

Process screenshots of component E in construction (exterior part and rivet)

2 drafts of the spring and the final version

At first, I was uncertain about how to replicate the curvy folded metal exterior. After considering a few over-complicated methods, I realized I could draw the edge of the folded piece and extrude it. Then, I rounded all the edges to match the original.

I also made a compression spring to fit inside the component. I used the Spiral and Pipe commands for this part.

Component E in process/ making the interior components

Different views of component E

After modelling the exterior metal component and the spring, I modelled the sliding pieces of metal that hold the spring. I measured the original very carefully and I am confident that the curves and angles match.

Some of the commands I used for the 4 interior metal pieces were circle, Tangent (circle), FilletCorners, Trim, extrude, and polyline.

Those are the same commands I used for modelling the exterior/rivet. But with the exterior/rivet models, I additionally used the commands (not in this order) Offset, FilletEdge, Cylinder, Ellipsoid, BooleanUnion, and BooleanSplit.

2) Rendering Component E

3) Learning about ball bearings for Component I

Photo of wheel with a ball bearing

This week, I realized that the wheels for the easel were a lot more complicated to reverse engineer than I'd initially thought. For instance, I can't take apart the ball bearings in the caster wheel. There are four ball bearings per wheel, two in the holder that attaches to the easel and two in the wheel itself.

As I was planning my approach for modelling the wheel, I tried to guess what the interior of the ball bearing looked like. To process my ideas, I made some really rough sketches that show the balls moving freely inside the component.

Sketch - predicting the ball bearing interior

As I thought about how the wheel moves, I noticed that there might be a conflict with having the balls move freely. When the wheel rotates, the balls would hit each other, which might prevent the wheel from rotating smoothly.

I looked up diagrams of ball bearings and realized that they were different from my prediction. Nearly all of them had retainers (also called separators or cages) that keep the balls spaced out.

The only bearing without a retainer is called a full complement bearing, and it is only used for high loads at very low speeds where friction is not a problem. They're usually found in gear drives, rolling mills, and machine tools. So, it is unlikely that the wheel I'm modelling will have a full complement bearing.

Render by Factory Handling Solutions that shows
different types of ball bearing retainers

Even though my prediction did not match what many ball bearing interiors actually look like, I was glad that I had considered the problem of having ball-to-ball contact.

After reading this article and this other article, I realized that there are multiple reasons that ball bearings use retainers (aside from the reason I first considered). For instance, retainers:

help retain grease around the balls
restrict radial movement and maintain equidistant separations between the balls, thus

reducing wear between elements
reducing heat build up
extending bearing life

allow bearings to move at higher speeds

After learning about the importance of retainers, it became important for me to accurately 3D model the wheel. Even though I couldn't disassemble the bearing, I didn't want guess or cheat at making the structure because I might miss out on an important feature. So, I inspected the bearing, determined its code (608zz ABEC-7), and recreated it by referencing schematic diagrams and photos.

Schematic diagrams and images of ball bearing (608-zz abec - 7) that I referenced while making my model

Photo of a typical ribbon retainer

By looking at the diagrams and photos of the 608zz ABEC-7, I determined that I had to create a specific type of retainer called a "ribbon retainer."

My ribbon retainer model would be similar to the one shown in the photo on the left, but it would hold seven balls instead of eight.

Fun fact: The 608zz ABEC-7 ball bearing is commonly used for roller skate and skateboard wheels!

4) Modelling ball bearings for the caster wheels (Component I)

Looking at all the screenshots, my process looks more straightforward than it actually was. I had to think a lot about the commands and the order in which I used them.

Here's a rough explanation of my process:

I used polar array to equally separate seven spheres along a flat ring. I duplicated and enlarged one of the spheres to be the size of the sphere holder part of the retainer. I kept the original sphere in the same spot. I extended the walls of the ring and used them to BooleanSplit the enlarged sphere. Then, I used the Shell command on the enlarged and cut sphere. I used BooleanSplit again to discard the part of the ring that goes through the sphere. I used polar array to give the remaining six spheres their own shells. I used BooleanSplit again to discard parts of the ring. Then, I placed a shallow cylinder at the center of the form and used it to split the retainer (not the spheres). In real life, the two halves would have been rivetted together. On the top half, I used FilletEdge on all the hard edges. Instead of repeating the process on the bottom half, I deleted it and replaced it with a mirrored version of the top half. Then, I punched holes along the flat parts of the ring where the rivets would go. I made the rivets using the Elipsoid, Cylinder, and Boolean Union commands.

~ oooh magical floating ribbon retainer ~

The rest of the ball bearing was fairly easy to model. I got a lot of the measurements from the schematic diagrams I found online, but I also double-checked them with my caliper when possible.

One part that I struggled with (which looks simple in retrospect) is bending the inscription on the bearing shield. At first, I couldn't figure out how to place the spine to create an arc. I kept starting the spine on one end and ending it at the other end of the text. This would bend the shape, not in an even arc, but in a slope-like way. I ended up figuring out that I could create the form I wanted by starting the spine in the center of the text to bend one side. This feels like a bit of a cheat, and I bet there's a better, more precise way to arc text. However, this method did work fairly well.