The Force Is Strong with [TORMACH]

There's a lot to make. And what needs to be made would take a really long time, if they had to be made manually. So I'm always looking for force multipliers that will make the jobs faster or easier, or both. In the Lowell Makes machine shop, that force multiplier is the Tormach!

Hyperdrive frame on the Tormach CNC mill

The Tormach CNC mill is a computer-controlled machine that carves parts from blocks of metal. To produce a part, I have to design and model the part in 3D. Then I create a CAM program to tell the Tormach how fast and deep to cut, what tooling to use, and what paths to take. All that takes hours before any IRL metal is cut.

The latest part I worked on is the frame for the main drive train of the Falcon, which I dubbed the Hyperdrive (haha). Here's the design.

Here are just two CAM operations that make the frame.

The frame is too wide for the Tormach's reach, so I had to create two setups that each worked on half of the frame. I had to move the plate over to do this and somehow keep the plate aligned. The short side of the frame was aligned using the back, stationary jaw on the Tormach but the long side still needed a common starting point between the two setups. The trick was to machine an edge into the first half of the frame as part of the first setup, then use that edge as the origin for the second setup.

Surface used as origin of second CAM setup

After many changes, our machine shop captain blessed my CAM programs. I got started making the real frame. I spent the first day just setting up the stock on the machine. 

First, I had to adjust my CAM setup until all the operations fit inside the Tormach's milling envelope. It can realistically mill inside a 14" x 6.5" area, but that also includes the diameter of the tool used, so effectively it's 13.7" x 6.2" using a 3/8" diameter tool.  It was easy to tell when it fit because the machine refused to continue if it didn't. Next, I had to mount the stock metal. This was the hardest part of the setup.

Because of the Tormach's dimensional constraints, I had to slim the design of the frame by almost an inch and reduced the size of the pulley mounting holes to reduce the length. The final frame dimensions became 5.5" wide x ~25.5" long. I also milled down the stock from 7" wide to just under 6 inches. It was quicker to do that on the manual mill. 

The last part of the setup was securing and supporting the stock material. The center of the milling area was secured by a vice on the machine, but metal stuck out from both ends. If I tried to machine it like that, the metal would start to vibrate whenever the tool danced over the ends. I needed to jack up the ends to increase the material's stiffness. 

Instead of ordering machinist jacks and waiting, I took our machine shop captain's advice and made jacks out of some spare aluminum blocks and a few bolts. Then I used those jacks to hold up the ends of the stock. 

All that was the first day. 

I then spent the next two days running the first CAM program, moving the stock, resetting the X origin, and running the second program. And cleaning. The sheer volume of aluminum flakes was enough to fill half of a 30 gallon trash bag. That takes a while to vacuum up.

The final result came out great! I had a few mishaps that removed some "extra" material from the surface of the frame, but nothing that would ruin the part. And now it has some character.

I would have never been able to do something this intricate manually. I could've made something, probably rectangular shaped, that worked but didn't look as good and probably weighed more. And I think it would have taken at least as long. Now, thanks to the Tormach, it's ART.

Oh, and multiplying my machining skills is not the only advantage of using the Tormach. Once a program is proofed (i.e., pretty sure you won't ruin the machine, the part, or your tool), I can do other things while a job is running. For example, I can prepare the stock for the next CAM job while the current one is running. But I'm always listening for the high-pitch whine of the tool happily cutting through material.

The frame was too complicated to leave unattended. But while I worked on the CAM program for the frame, I used the Tormach to make 14 bearing cups that will mount and weld to the frame and 4 tensioner blocks that will also mount to the frame and will allow me to tighten the belts around the pulleys. 

The bearing cups were a particular win. Making those on a lathe would have taken me days. With the Tormach, I was able to bang those out in 2 hours.

I now have all the pieces to put together the Hyperdrive, and it's thanks to the force of the Tormach!

Rudy July 06, 2023

Read more

I Made Alot Of Special Modifications Myself

I dry fit everything we've made so far on the main truss. Everything mostly went together. The rear supports took a little convincing, but went in relatively easily. The front supports needed some light hammering to mount. Then. The issues. Since I assumed there were going to be issues, I wasn't surprised when I found some. 

On the front right side, I managed to install the gussets that we added to the front suspension upside down. smh. I now have to widen that hole in order to connect that CV axle to the differential axle. 

On the left side, the front suspension frame warped a little during welding. This pushed the top joint down a little over an 1/8" lower than designed. The knuckle binds when it rotates. This directly affects steering, so it cannot be ignored.

There are holes at the top of the knuckle brackets that the frame tubes fits into. I thought I would have to mill that down a little deeper, but after staring at it for a while, I realized there's an easier fix. 

The CV plate sits on ledges that are cut into the sides of the knuckle brackets. I think I can just remove another 1/16" from each ledge--a much easier operation. The knuckles should then fit in the space available and solve the binding issue. 🤞

For the rear supports, I used the new rear wheel hub to check the fit. At first, the hub was difficult to put in place. I realized that the welding beads on the back of the rear brackets were pushing against the truss' cross-members and moving the supports inward. Sliding the brackets down until the beads cleared, solved the problem.

Because each side of the truss' cross-members go in opposite directions, the axle bolt will need to be installed from the side after the wheel is dropped in place. At least we know the rear wheel won't fall off. 😅 

After all of that, I broke out my grinder to cut out the truss cross-members that are in the way of mounting the rear wheel.

The structure is now ready for the rear wheel. Of course, there is still one more step. There's always one more step. The new blue hub has to be mounted on the existing rear tire. This will require detaching all the spokes on the 26" tire, rebuilding the wheel with the new hub, and re-balancing the tire. 

That seems easy in comparison to other things we've accomplished, but I've never done it before and I have been stalling (i.e. doing other things first) 

I should be fine. I've watched this video, like, two times.

Rudy June 26, 2023

Read more

Would It Help If I Got Out And Pushed?

I spent most of last weekend in the machine shop. Progress has been slow but steady. Wheels have not been mounted yet, but hopefully they will be soon. We worked on preparing the rear wheel supports for welding. Our team welder, Jonathan, tacked those up and is working on them this week.

I tried to assemble the front suspension, but ran into a real world vs. virtual world discoveries. There was too much play where the tubes attach to the truss, making it very easy to pinch the thrust bearings at the knuckles. The knuckles (which support the front tires) need to spin freely or turning the sculpture will be impossible.

So we had to hold the attachment points at a specific distance and maintain the alignment between the top and bottom tubes.I was going to add tubular structure, but realized that the front axles run right through the area where the reinforcement would be located. We needed a brace that had a clearance hole in the middle of it. Jonathan suggested using plates instead of tubes. Great idea! Now, we just needed to make them. This sounded like a job for the Tormach!

I took some measurements, modeled the part in Fusion 360, and then created the CAM program to carve them out.

The fit up went well. Jonathan is also working to weld those pieces on this week. I also worked on the lawnmower differential.

This was the riskiest activity of the weekend. In order to get this to work with bike chain, I had to remove the large ring gear (the giant gear in the picture) and replace it with a sprocket. This would also remove some dead weight from the front. And I had to disassemble the differential in order to get measurements for mounting the sprocket. 

I thought the mounting hole locations on the sprocket might be a problem, but they mated up perfectly with the holes in the differential housing. Awesome! Problem one solved. The next problem was that the hole at the center of the sprocket was much bigger than the diameter of the axle rod. However, it had to be a loose fit hole in order to keep the miter gear properly located inside the differential housing. Another job for the Tormach! 

I measured up the dimensions on both the housing and the sprocket and modeled an adapter plate. I designed it with a hub to center the sprocket and a hole through it for the axle rod. I also added a matching bolt pattern so the adapter, sprocket, and housing could all be bolted together through the existing mounting holes.

Then I wrote the CAM and made the part on the Tormach.

Upon inspection of the housing, I noticed that the mounting flange had tabs sticking out on its edges. The original ring gear had a relief milled out to accept the housing flange with slots where the tabs stick out. When the differential spins, the ring gear turns the housing by pushing from the tabs transfer the turning forces instead of the pins/bolts that hold everything together. 

I decided I would also cut a relief out of the new sprocket to do the same. I measured and modeled again. Once I had the pocket I wanted, I wrote the CAM to make it. Since I only had one sprocket, I decided to do a test run on 1/4 inch aluminum plate.

Making a test piece was great. The housing didn't actually fit the first time.If I had run the job on the sprocket, I would've had issues. I adjusted all the dimensions to be slightly larger and tried again. This time it was a perfect fit!

Once I knew the relief would work, I then had to carve it into the aluminum sprocket. You might ask, "How did you hold the sprocket on the mill?" That was a question that perplexed me for a while. Then I realized I already had the solution.

The adapter I had just made had square sides and centered the sprocket. The only issue is that I couldn't use the existing bolt pattern to secure the sprocket because I was cutting through them to make the relief. So, I drilled and tapped a hole inside of that bolt pattern and used a shortened bolt (short enough to avoid the cutting head above it) to secure the sprocket to the adapter.

Milling out the relief was a lot harder than the test runs I did. The sprocket is black anodized, which gives it a tough surface finish that is apparently very, very hard. I had to slow down the feed to 2% (!) of the original, but I was able to cut the relief into the sprocket. The roughly 2 minute job on untreated aluminum plate took almost a half hour to complete on the sprocket.

I have to wait until at least Wednesday before I can work on the front/rear suspension. But in the meantime, I now have all the parts I need to mount the differential on the main truss! The parts all fit individually. Now we will see if it all goes together.

Rudy June 13, 2023

Read more

You Must Unlearn Everything You've Learned

After some early success welding the inside joints, our team welder ran into trouble joining all the pieces of the front wheel supports together. Aluminum is very difficult to weld. 

The surface of aluminum is made up of aluminum oxide, a tough surface that protects the interior aluminum from corrosion. It melts at around 3700 degrees, hotter than the melting point of steel. The aluminum underneath melts at around 1200 degrees. So welding aluminum requires punching through the skin at high temperature and somehow not completely melting everything underneath. The analogy I was taught is that it's like welding a banana. 

At the same time, keeping extraneous gases out of molten metal is critical to the strength of the weld. Any contaminants that mix in while the aluminum is being welded will weaken the joint at best or make it impossible to weld at worst. He dealt with the latter for weeks. 

His experience welding steel helped a little, but it took him a lot of thought and experimentation to figure out his process. It basically boils down to obsessively cleaning all the surfaces with acetone between each prep step. 

After a lot of work, he completed the front wheel supports, but he might swear off working with aluminum forever.

We are almost ready to assemble the front suspension!

Rudy June 07, 2023

Read more

This Deal Is Getting Worse All The Time

While our welder has been working on joining all the front wheel support tubes, I turned my attention to the supports that will hold the front wheel differential in place. I last worked on these back in January. The original idea was to make a bunch of aluminum plates and weld them together. But then we started having some trouble welding the aluminum tubes. Since I had the time and I had started to think maybe we shouldn't try to weld everything, I decided to design a weldless version of the support. Yes. I altered the plan further.

The new, NEW design

The old, new design

Instead of using ball bearings, I decided to reuse the original bronze bushings that came with the differential. They are much smaller so they don't need as much material to hold them. I first made a bed with caps to hold the bushings, but that turned out to be too hard to fabricate. The inside edges of the bushing pockets cannot be cut square, so I would never be able to finish making this part as designed.

Our machine shop captain suggested just making blocks and pressing the bushings into them. Then I could just bolt the blocks to a plate and bolt the plate to a support. That simplified things even more. 

Now I "just" have to make the parts. I already finished one of the plates.

I tried to make the support brackets last weekend, but I had trouble with the CAM operations I programmed. The cutting operations IRL were a lot deeper than the virtual ones seemed and I ended up breaking my end mill bit. End mills are not cheap. The worst sound in the world is the cutting going from a high pitch hum of a good cut to the low frequency drone of a bad cut.

I will revisit that next weekend. In the meantime, I will make the bushing blocks on the CNC mill. These parts are simpler so the CNC operations are simpler. 

We'll see what happens. 🤞

Rudy May 31, 2023

Read more

Pray I Do Not Alter It Any Further

I've been vacillating on the design of the knuckles that support the front wheel axles for a while now. First, it was a series of flat aluminum plates that would be welded together into a bracket. Then, I redesigned it to use rectangular tubing to avoid having to weld plates together. I knew I'd have to machine all 4 sides of the rectangle, but it seemed easier than the first scheme. I was confident enough that I ran some stress analysis on the rectangular bracket. Everything seemed doable. 

But as we've gotten closer to actually making these parts I've changed my mind again. The way it was designed, the parts would require both CAM (computer-aided machining) and considerable manual machining to complete, including a lot of tricky alignments between the faces. Making the knuckles suddenly seemed harder than I originally thought. So I altered the plan one more time.

Instead of making it out of rectangular tubing, I went back to the idea of multiple pieces but with a twist. The piece that holds the wheel axle is again a flat plate, but, instead of welding, it will be bolted to two brackets.

The brackets are thicker than the plates that preceded them, so they can better support the forces going through them without needing an inside wall. I also added ledges to the brackets to reduce the forces going through the bolts that hold the face plate in place. And I designed them so 90% of the machining can be done by computer. I still have to drill and tap mounting holes on each of the brackets, but that should be easy to do manually after the CAM jobs are complete.

So yes, it still requires both automated and manual machining, but I think it's much less work to complete than before.

It took all day Sunday and Monday evening, but I was able to run the CAM jobs on the Tormach with the help and supervision of our machine shop Captain. The parts came out great and now they are made, so I can't alter this any further. 

I just have the manual bits to complete. I have to flip the plates over and remove the extra material in a facing operation (looks like the plates are in Carbonite! 🤣) and I have to drill and thread mounting holes on the brackets.

Rudy May 10, 2023

Read more

Angle the Deflectors

We started making parts! After months of designing, planning, analyzing, sourcing parts and buying materials, we finally started making parts. The front end wheel supports are involved, so we started there. The front structure has multiple tubes at multiple angles that all have to fit together as snugly as possible. Most of the sculpture and pilots' weight will be supported by these tubes, so getting them right is critical.

To make it harder, the cuts on either side of each tube are perpendicular to each other. There's even one tube that has cuts in two different directions on the same side. I always say, "You're only as good as your tools," and this was no different. There's no way I could do this without specialized tools.

I had already bought a tube notcher that would allow me to cut the 30° and 60° notches that I needed, and 1-1/2" and 2-1/4" hole saws. I already owned a digital angle finder that I could use to set angles accurately on the tube notcher. To align the two ends correctly, I'd get to use my newest tool--a digital level.

I started with the front two tubes, since they had the most occurences of the same cut. I set the notcher at 90° using the digital angle finder and cut the large arcs.

Then I taped 2-1/4" tube pieces snugly into the notches, so I could balance the digital level on it (Thanks to our machine shop captain for the great tip). The other side of these tubes needed 1-1/2" notches that were perpendicular to the first cuts. After I measured the angle of the drill shaft, I rotated the "T" until it was 90° to the drill shaft and cut the outside notches.

Aligning the tube to finish the long notch on the lower, rear tube

Cutting a 1-1/2" notch 

I repeated the steps until all the profiles were cut. It took several hours over some weeknights and a weekend.

 

Fit check - lower tubes

I would have been done sooner, but I screwed up the 60° cut on the lower rear tube. I convinced myself the alignment was correct, but it actually was backwards. And that's how it goes sometimes. I had to buy more tubing to make the correct part a few nights later. But they're all cut now. 😎

Next, we need to make the brackets and plates that make up the front wheel knuckles, formerly known as the CV axle brackets. The knuckles will support the wheels while allowing them to pivot around an axis. These parts will be made on the CNC mill, which is a topic for the next post.

May the Fourth be with you!

Rudy May 04, 2023

Read more