But can we combine the advantages of 3D printing with the disadvantages of CNC milling today in order to produce a good outcome? Trying this out was a long-time goal of mine; I’m delighted I finally got around to it. 3D printing was utilized to create an enormous object, which was later cut down to size using CNC milling, a subtractive manufacturing process. Is it worth it? Please check out the test project for further information.
The necessity to design and produce your own multi-part wheels was one of the largest adjustments, though. Over the years, several groups have experimented with various approaches, including 3d printing. Aside from the obvious layer lines, they weren’t quite correct. Another team tried machining the wheels from 15-millimeter thick acrylic, but they gave up before the project was finished. Acrylic warped and became unusable due to its high temperature and subsequent distortion.
We needed to spend some time on the lathe to get them precisely where we wanted them, but we couldn’t get them there. However, wheel inaccuracy was a persistent problem despite our best efforts to complete and race the vehicles. However, we could see layer lines on the 3d printed wheels, and their accuracy deteriorated as a result. On the other hand, the CNC machined wheels were quite exact, but they wasted a lot of the very costly 15mm acrylic, and their precision suffered as a result. A 3D-printed product will still be quick and efficient with materials as long as heat and warping are properly managed, but if we build this component oversize and then grind it down to specification, we should have an extremely high degree of precision. On paper, this seems great as long as we can control the temperature and prevent warping, but in practice, the procedure will be more complicated due to the two steps involved.
That’s what’s going on right now. However, it’s a good fit in the grand scheme of things. So, what’s the goal here? This can’t be, right? For example, when e3d debuted its tool changer, they talked about additive, subtractive manufacturing per layer, where a compact spindle was used to manufacture components of the print mid-way through the process.
3D printing would be made possible, but the resulting pieces would have extremely tight tolerances, making them ideal for assembly with external parts. To be honest, we never had anything like that. Using a standard 3D printer and CNC machine, was our strategy. With a 3D printer, we would begin with a wheel blank that is larger than the final product. It has a part on top for locating the end mill and zeroing the machine, and it might be connected to a threaded mounting plate acting as a spoil board.
Machining away the superfluous plastic would produce a smooth and precise surface suited for a competition like this. First and foremost, we must locate a 3D printing filament that can withstand the high temperatures and resulting warping associated with CNC machining. Constructed a small test component with high infill and increased perimeters all around to provide meaty outside walls for the test item I created. To start, I used PLA as a control filament, and I fully expect the next PETG print to be a colossal failure. As a common, easy-to-find filament that has a higher thermal resistance, Apollo X (based on asa), which is based on abs, and nylon, which warped off the bed and came loose, which is why I’m hoping the other filaments provide an improved solution.
In contrast to pla, which occasionally melts when drilled through with apollo x, I was able to drill through it a second time with no problems. Despite the fact that I sanded around the edges, the plastic did not melt and remained in its original shape. Using my samples on the disc, sander, and liner, I’ll see how much material can be removed while also heating the pad. It’ll be a basic test, but I’m hoping to learn something. Finally, I’ll exert a lot of pressure on the substance to see if it melts and then examine the resulting commotion. Finally, here are some examples of the work. Pla turned into a molten mass and completely lost its form. After melting, Petg turned out to be a pleasant surprise.
That was about all we got. My earlier testing with this material shocked me, and I was startled to see how similar it appeared to apollo x at the end of the 3D print and CNC test. This time around, I’ve decided to continue working with petg and Apollo X. Therefore, petg has emerged as the darkhorse. So here is my wheel, and it’s a lot simpler than the school’s design of an f1 car. However, this test requires it.
A bearing may be found in this blue region. These have a three-millimeter hole and are used in wheels to reduce rolling resistance. During the machining process, a bolt will be used to keep the wheel in place while I verify how snugly the bearing fits. The pocket is seven millimeters wide. For this procedure to work, we need bigger geometry, which you can see here in blue. The center bore should be just big enough for our bolt to pass through, and then we have this tiny lip to center the CNC end mill and help us zero. That’s what you can see here in blue, too. In the section view of the model, it is hoped that everything would be retained in place. Next, we’ll need to print our bigger wheels in PETG and apollo x to complete the project.
In order to get a perfectly solid wheel, I chose to increase the infill % rather than decrease it. I added a few more perimeters to the design. In the end, you’ll have a sturdy model with a pleasing infill pattern. Right-side petg has some fine stringing, and both models share a noticeable zed seam, the kind of wall artifact that has caused us problems in the past. Using the printed lip on top to locate the six-millimeter end mill was a success. However, the space on the left was too small for the bolt to pass through, so I had to grind down a bolt to make it fit.
In order to pull the threaded insert into the bottom of the plywood without damaging it, I used a long bolt and washer to pull it into the hole I drilled before applying super glue. The bolt was centered on the spoil board, but it didn’t matter because everything on the underside was flush and ready to go. Clamping the spoil board in place, I nailed down the first piece of printed material to the middle of the work area. To aid in the final alignment, I used 0.1-millimeter increments to lower the cutting bit to just above the printed component after putting it in the rough position. Once I was certain I had zeroed the machine, I would manually crank the spindle to better line the end mill with the correct homing position. Things got off to a bad start, but I was able to work things out.
This was the final tool path I created using the cam program deskproto. This was, in reality, the sixth edition. Consider how we arrived at this point: visualizing g-code. Meandering toolpaths are a combination of conventional and climb-cutting spindles running at 15000 revolutions per minute and feeding at 500 millimeters per minute, which I found to be too sluggish. Because of this, I increased the feed rate manually to 1000 mm per minute using the machine controls.
Cutting an outside portion and then raising it was a rather inefficient tool path at this point. As a result, it might go back to the model’s interior to cut the tool route. This would, I reasoned, give the plastic a chance to cool and therefore avoid melting. It was a bummer to see the screw come loose and then the model remove itself on the final pass of the job and, of course, ruin the exterior of the wheel, so on to iteration. First I used some Loctite and applied more torque to secure the retaining bolt to the spoil board. However, I still had some issues with the screw coming loose and the model removing itself on the final pass. A thousand millimeters per minute feed rate was permanently added to the g-code, and another alteration was made.
As a means of expediency, I increased the step-down from one to two millimeters, but this proved to be a mistake. A second dramatic collapse occurred as a result of the increased stress on the 3D print’s base. It was thus necessary to reinstate the one-millimeter drop in order to prevent this disaster, and this proved to be a successful combination with my first wheel, which was successfully completed without any faults. A little lip on the bottom of the model was left uncut because of an error I made in des proto’s job setup. The fourth sense is like an elephant’s foot: you can’t see it but can feel it.
I reduced the final cut’s depth. However, the spoil board was a mess. As a result, the wheel’s diameter remained consistent thanks to this sacrifice, along the whole length of the truss. Version 5 has the same appearance but is more effective. As a result, the cutter can focus on one region at a time instead of having to switch back and forth between the inner and outside areas of the job.
Aside from the switch to just climb milling, you may have noted that the surface quality should be improved. One adjustment was made for the final version to battle the vertical lines left on the outer wheel’s surface, which I assumed were the result of these recurrent vertical motions. For each layer shift, I implemented a ramping method rather than a vertical movement. Finally, I was able to machine the wheels in a very efficient manner because of the zig-zag movement exhibited above. Despite the extra effort, I arrived at my destination. The first step is to verify the correctness of the dimensions, and there are three critical aspects to this.
A test fit of the seven-millimeter bearing in the internal ball was performed first, and I’m happy to report that it was snug and an excellent fit on both the print and machine versions. On the other hand, how does this compare to a completely 3D-printed equivalent done with a very narrow nozzle? In order to get the proper clearance, the model or slicing procedure would need to be modified because the bearing ball isn’t quite perfect. The minimal exterior diameter is 26 millimeters. Thus we usually strive for 26.2 or 26 millimeters.
While the fully printed version is quite precise, there is some fluctuation, resulting in an object that is not perfectly round. The diameter is actually closer to 27 millimeters in the elephant’s foot area. The printed and then machined wheels were really pretty uniform in terms of their quality. Inches long.
Consistently across their circumference, and best of all, the circumference was rather uniform across the four wheels’ circumference. Having ball screws instead of belts on the CNC I used probably improves the quality of the finished product. Wheel width is our last measurement, and the minimum is 15, so we strive for a little more than this. But my first three machine wheels consistently measured 15.6, which is too broad for what we were trying to achieve while zeroing out the machine.
I had a difficult time determining when the cutter had descended far enough to contact the last wheel piece of the model. To get a more accurate final measurement of 15.2 millimeters, I first zeroed the z-axis on the surface of the wood, then moved the end mill into place to zero x and y before manually moving the z height to 15.3. Our final comparison is the quality of the surface, which has an impact on the performance of the interior. ‘
The machining markings on the left-hand wheel are nice, but the most significant element of the machined wheel is the wheel surface. The printed version is smoother and more consistent, although there are pits and imperfections that detract from the aesthetic. For testing purposes, I removed the salvage board mounting hole and drilled a new one. This way, the long bolt will be centered above the bearing, which means it will align perfectly with the outside surface of the wheel. So, if it’s true that the wheel should rotate in that direction, then it should be possible to wet sand and polish the wheel’s outside, but as you can see, I’m not very patient with sanding. There wasn’t much of a difference despite my efforts. However, additional experimenting with feeds and speeds is likely to yield a superior surface polish, so I’ll keep trying.
Because petg is more widely available, I decided to stick with it instead of apollo x, which I didn’t find to be much better. In order to avoid gaps between the extrusions and provide maximum strength, the only changes I would make to the printing process are an increase in the flow rate and temperature. For the vast majority of individuals, I’d say this isn’t worth their time, but for those with a specialized use like this, it could be just what they need.