Major progress on the Lemniskate - I rode it under power for the first time! 

It didn't catch fire! - but parts did get warm quickly, it needs more power and less rolling resistance, but I have a few ideas to address those things.  I need to get the second motor connected, and use a better speed controller for the motors.  The long wires are adding too much resistance, and they are too small in diameter.

Like yesterday, all day today was spent machining parts.  The complicated to machine spacers and the main connecting frame, along with the pulley shaft, pulley modifications and a pulley spacer were machined.

The last parts that are missing is the linkage that makes both ends steer equal and opposite directions - even so, I did take it out in the driveway (at midnight) and drove it about 50' a couple times.

Spent all day machining parts, but completed the two foot plates and the four supports for them.  The side plates were done at night over the last week.  It's great to see the parts fitting together for the first time!

Much of this long weekend was spent replacing the clutch in the Camaro.  That required jacking the car up, removing the exhaust, drive shaft, clutch slave cylinder, transmission and clutch assembly.  While we were at it, we replaced the rear main seal of the engine and the pilot bearing in the end of the camshaft.  The weird discolorations on the shiny faces are bad - it shows that the parts were badly overheated.

I did find material to make the 1/8" thick aluminum plates for the Lemniskate, and got all of the pieces rough cut for machining.  My father is helping me design a special fixture plate (purple plate in lower image) that will be used to hold all of these unusual shapes.  The plan is to drill all of the holes in the blank plates first, then use those holes to bolt the blanks down to the purple fixture plate so the other features can be machined, such as the outer profile and inner windows.

The ten guide wheels took a LONG TIME to machine (about an hour each), and I made a spare just in case.  They need to be made from a tough steel to hold my weight and ride on the steel wheel flanges that I made on May 3rd similar to how a train rides on tracks (but in this case they are circular).  Lots of time working on the lathe to reduce the outer diameter, add the undercuts at each end to ride on the flanges, along with a large through hole and a tight tolerance bearing pocket in each end.  I also sanded and painted the wheel flanges and assembled the wheels.

I was busy machining the 10 axle shafts needed for the ten guide wheels, five for each large wheel.  All of the work was done on the 9" South Bend manual lathe.  They consist of ten Ø5mm shafts with locknuts on each end, some end spacers, two bearings for each guide wheel and a central spacer between the bearings.  Next up is to machine the ten guide wheels that go over the bearings.

I finished painting the steel wheel flanges that were laser cut.  Unfortunately, the inner diameter wasn't as smooth as I wanted, so we used the rotary table on the milling machine to clamp all four parts together, carefully center them with an indicator, then use a carbide endmill to recut the inner diameter smooth (see picture).  If I were to leave the bumpy surface here, the small rollers would magnify the small bumps.  Because this is such tough steel, we had to use a carbide endmill and rotate the part slowly, taking very light cuts.

The machining of the gearbox housings is complete, with just a couple minor mistakes on one of them.  Hopefully it will still work well enough.  It should be able to be ridden with only one motor, otherwise, I will have to make another from scratch.

Working on programming the next parts to machine, including finishing the two gearbox housings, the sheet metal side panels and the 10(!) inner guide wheel assemblies.  I am also placing the orders for some steel that I can machine and then harden to ride on the inside of the laser cut metal rims.  The blue lines are the path that the end mill cutter will take to carve away the unnecessary metal, turning it to chips.  This must be done for each part, which can take anywhere from a few minutes each to program to more than an hour each depending on how complex the part is.

Today I presented at the 2025 Innovations In Learning Conference in Concord, and discussed this ELO project and my experience at the week long ARMI class I took in February.

Finally received the laser cut AR-500 steel plates from SendCutSend, and assembled the first wheel!  This is a special wear resistant steel that is used for excavator buckets and heavy duty applications.  It arrived with black flaky scale on the surfaces that had to be sanded off.

I was able to assemble the first wheel, which fit together great!

My father helped me complete the installation of the new subwoofer in the Camaro, using the fourth revision of the 3D printed speaker mount.  We also learned how to 3D print TPU, a rubbery Thermoplastic Polyurethane that we have never printed before.  We used it to make custom gaskets for the speaker and adapter (shown in the 3D printing slicing software in the picture).  We had to remove half of the dashboard and several other interior panels of the car to run the 8 gauge wire from the battery all the way to the amplifier in the trunk, using the full 20 feet of cable.  The difference in sound quality is excellent!

I finalized the design of the wheel rings.  While looking through our "stock" I realized we only have mild steel which is too soft for a wear surface for the rollers.  A friend has used SendCutSend for his BattleBots robots in the past, so I had these parts quoted through them over the weekend and I just ordered these today.  I wanted to get into the shop this weekend, but I'm not feeling that well so this was something I could do to move the project along.

More machining.  The two rectangular gear box parts at the top of the image are about half done, and are looking great.  To see what this part will look like, see the picture of the design from Dec 28-29.  The motor and die cast gearbox from the circular saw fit perfectly.  This part was a bit simpler to machine and I was able to bolt it down solidly to the table, so the machining went much better on these parts.

I presented my ARMI and Engineering ELO to the Timberlane School Board.  Video can be found here: https://vimeo.com/1063351781

I started machining the first parts for the Leminskate over the weekend.  These are much more complicated parts than I have made in the past, and my dad has been teaching me CAM - Computer Aided Manufacturing.  This is the process of converting the 3D model in SolidWorks to create G Code, which is the type of code used to control the CNC Bridgeport milling machine in the garage used to cut the parts.  This also involves fixturing and clamping, so the part is securely held down while the cutting is taking place.  The lower shaft support part has been a bit harder to get right.  The three on the left are machining failures, while the one circled in green is a good part.  Hopefully the fifth try will also work for this part, as I need one for each wheel.  The finished part is only about 10% of the block of material that I started with.  Each try took between 2 and 3 hours.

The old amplifier was mounted directly inside the subwoofer box, but the new one is much larger.  I realized that it could be mounted to the lower rear section of the box with a simple metal plate. I used heat stake inserts, which are small brass cylinders with grooves on the outside that can be heated with a soldering iron and pushed into plastic to provide mounting points.  5 of them were installed into existing holes in the back section of the subwoofer.  A white metal plate was then made to fit the size of the amplifier, and 9 holes were drilled in the plate.  5 line up with the new brass inserts and 4 are for the corners of the amplifier.  The white plate is visible under the amplifier in the lower green box on the left side of the picture.  As you can see in the green circle, the fourth version of the 3D printed speaker adapter finally fits!  It'll be a bit more work to get all of the wiring done and the panels reinstalled.

I ordered a new subwoofer speaker and amplifier from Crutchfield, and started figuring out how these new parts would be mounted and connected.  The subwoofer speaker is the same size speaker cone and the same mounting, but the magnet on the back is HUGE and doesn't allow it to fit into the subwoofer box.  I designed and 3D printed a simple spacer ring to shift the speaker forward, but it could only be about 0.6" or the speaker would hit the inner wheel well of the car.  The simple option didn't work, because the large magnet hit objects inside the box.  I tried again and shifted it and rotated it.  It still interfered.  Another try and it was still hitting.  Finally, the fourth try seems to work.

With spring coming up soon, I am anxious to get the Camaro out of storage and back on the road.  It's much more fun than the Prius.  The original 31 year old speakers just weren't sounding too good, so I am working to replace the subwoofer in the rear hatch.  The old paper speaker cone was warped and damaged, and the original amplifier no longer worked.  I started by checking the original manual for the car that has all of the wiring neatly organized to see if there are any wires that can be reused when upgrading the subwoofer and amplifier.  There were a few different options for the audio back then, but I found the diagram that matches the wires that we have in the car.

This weekend, the recent breakthrough was expanded to cut another expensive and difficult to machine part whilst also simplifying the device. Originally there would be 5 "V" shaped wheels that ride on the inside of the massive metal gear. With a new modification that was changed. Instead of "V" wheels that the weight of a person is supported on, they are now stepped shaped wheels (the dark gray part with the ball bearings) that ride on the two metal plates that clamp the solid tires.  To not crush the plastic, aluminum hexagonal standoffs with threads are inserted through the 3D printed plastic internal gear. This new design uses the outside plates that are necessary anyway to support the new roller wheels and simplify their design. While both types pf roller wheels need to be custom machined, the new ones are simpler as they don't have the "V" angles.

Today there was a big breakthrough for this project. Originally, the wheel sub-assembly (screenshot to the left), required a massive, custom, internal steel gear to mesh with the belt drive. This gear is what the belt, that the motor drives, sits in and allows the wheel to rotate. This gear was also intended to take the load of a person's mass while riding the device, requiring it to be both precise and very durable. When working today I asked myself why does it have to be this way? Why can't the massive gear only take the tension of the belt and not have to support the weight of a person? So I did exactly that, in the new design the gear is now 3D Printed plastics, as it only has to take the power of the motors. The weight of the person will now ride on the interior of the metal rims of the tires. This modification saved a massive amount of time and money, as machining a giant steel gear from a large steel plate takes a long time, and steel stock that big is expensive.

This weekend I 3D modeled the mounts for the cordless drill batteries. My Dad owns a lot of Milwaukee cordless batteries, and they have enough power in order to run the two motors that were taken from the cordless circular saws. These batteries will be easy to use and also reliable. An integrated battery management system (BMS) is planned for the next version of the Lemniscate once the proof of concept has been established. A BMS would allow a battery to be completely integrated into the device and safely charged while being on the device. This system will also be able to be customized in order to maximize power, runtime, and weight.

The orange parts are the motor and die cast housing from the Ryobi saw, along with the two gears.  The transparent housing is the piece I designed last weekend.  I realized that the output shaft and drive pulley of the new gearbox was unsupported on the far side. If this was not fixed, the shaft and drive pulley could bend out of place under the tension in the belt. The fix to this problem is to support the other end of the output shaft. This required some strange workarounds as there is nothing else to mount to besides the face of the new gearbox half.  This is because the whole motor and gearbox has to move upward as one piece in order to tension the belt. The design ended up being this weird "n" shaped bridge with two long screws.  This creates a housing for a bearing to support the the shaft without any major changes to the design.  It also allows clearance for the belt around this unusual shape.  This bridge will also be machined from aluminum.

To effectively use these motors and gearboxes, I had to create the other half of the gearbox as in the original design which is in the plastic housing of the saw. While I could have chopped up the old plastic case, it would have been difficult to mount and integrate into the design, so I modeled the new half housing displayed in the screenshot in SolidWorks, which will be CNC machined out of aluminum.  This is a fairly complicated part and took a while to design, but it shouldn't take too long to CNC machine on the Bridgeport milling machine in our garage.

For the first version of this project, I did not want to spend money on major parts that may not be utilized in the next more refined version. For this reason, I started by tearing apart two old cordless drills for motors and variable speed triggers. Through testing, the speed of the motors was determined using a Speed Indicator (100:1 worm gear drive) and a stopwatch. Using a spreadsheet for rough calculations of the motors under load, using the drill motors and the attached gearboxes they could not output the needed RPM for the desired speed of the wheel.  Next, a couple of old Ryobi battery powered circular saws with similar motors (donated by my grandfather) were torn apart and these similar brush motors and gearboxes had the required torque and speed to match the intended top speed of the device. The end result is a combination of drill variable speed triggers and the motor and gearbox displayed in the photo.

This weekend I have done some further work on the Lemniscate CAD model. I changed and tweaked some parts of the project to better have an accurate design of the first prototype. This is because it's almost time for me to start 3D printing and machining parts for the first actual prototype. In the past, some of the designs have been for the future and possibly better versions, but for now the most simple basic, design is necessary to get real-world information on how the project responds to real-world challenges. 

I recently printed the new fidget tool model that I designed from scratch, but it doesn’t have the right dimensions in my hands, and I have an idea for a new wheel for the design.  The old original clicky wheel has also broken multiple times as the plastic spring does not hold up.  The two that I downloaded and printed are at the top of the image to the left.  My designs are the all black ones below, with some iterations.  The next design I finished is more ergonomic, and it incorporates a new wheel. Instead of a regular pattern on the wheel similar to a gear, this one has an asymmetrical indent. Shaped like half of a yin-yang symbol to get a different feeling when the wheel is turned in each direction.  Once the design is comfortable to hold and feels how I like, my father will show me how to program our CNC machine using G-Code in SolidWorks, the fully machine the parts from solid blocks of aluminum and stainless steel.

I wanted to get more practice with CAD and the process of slicing models for the new 3D printer, so I thought a small side project would be a good idea.  You can think of slicing similar to the slicer at a deli - the 3D model is cut into thin slices and the software coordinates the motor movements while the nozzle melts and applies plastic layer by layer like a miniature hot glue gun.  I have wanted a fidget tool to help me focus in class, and it’s a great way to learn what is possible to do with the new 3D printer.  I’ve always found most fidget things to just be too distracting, but I’ve recently found one that has worked well for me. I just printed it out, I found a file someone else made online. It's a small box with a simple mechanism, it's a wheel that clicks as you roll your thumb over it. I also Printed out the variations the other made as a way to get more comfortable with the new 3D printer. I also started modeling a new design utilizing a ball bearing instead of a plastic spring in the original design.

I updated the CAD model for the inside-out belt idea.  Then I exported an STP file from SolidWorks and opened that model in the Bambu software for the new 3D printer.  I 3D printed a couple of versions, with a minor adjustment made between printings, to make sure the tooth spacing was perfect.  It took a couple of iterations, but the second version seems to work well.  Now that we have confirmed the proper tooth spacing, we can be confident that it will work before machining one from a steel plate.

In the picture, notice the inside-out belt with the round silver pulley at the bottom.  The motor will drive the small pulley and the belt will engage the teeth around the inside of the wheel rim.  There will be rollers attaching the rim to the frame.

While drag racing the Camaro at the New England Dragway in Epping, the BRAKE light came on and the pedal was feeling a bit soft.

The replacement master cylinder for the brakes in the Camaro finally arrived, so it was time to get it installed.  The first replacement cylinder that arrived was used and leaking old brake fluid, so my dad had to order another.  It's been a while since I helped my dad with a brake job, and this one was more involved than usual as I wasn't only changing the brake fluid, but changing the flexible rubber hoses and replacing the master cylinder (Which is depressed when the brake pedal is pressed on.)  First, the old master cylinder was removed before the new one had to be bench bled. All this means is that while the master cylinder was being held in the vice on a workbench, new brake fluid was manually pumped through the cylinder to ensure that it was properly bled and there was no air inside.  This is a critical step as if the master cylinder is not bench bled it can malfunction through not having enough fluid inside when actually in the car. After the new master cylinder was installed in the Camaro, the next step was replacing the old flexible brake line with new braided stainless steel ones. After the lines were installed the final step was to introduce new brake fluid to the brake system and systematically bleed each brake, at each wheel, starting with the right rear, the farthest from the master cylinder.  The BRAKE light went out, confirming that the problem was the old master cylinder, and the feel of the brake pedal is MUCH better.  The track is now closed for the season, but we will go back in the spring for sure.  I also hope to try our autocross racing (time trials on a road course style track) in the spring.  Someone my father works with is a professional driving instructor, so we will be taking some classes in the spring.

My family just got a new 3D printer for the side business and during downtime I can use it for things I would like to make. It's a much newer printer, 10 years newer, and most things are different with it. For example, the entire machine has its own completely isolated controlled environment, to regulate moisture and temperature. It also has a removable build plate to allow easy removal of printed parts. The most difficult to get used to is the new software it requires to run it. It has a completely different layout than the out one along with new features and options this printer provides. I’ve spent this weekend learning about the new printer and trying out some basic prints with it. 

I reviewed the brushless motor idea to see what it would take to make that work in the future, but this will only be done if I can ride the prototype!  The magnets are easy to purchase and epoxy into place, but the trick is finding and prototyping a custom motor winding.  Two parts of this are tricky to source, one is calculating the proper diameter of copper wire to wind into coils and winding it properly.  The second is finding a source for “electric steel” or “transformer steel”.  This material is thin steel plates that are stamped or laser cut and laminated together, then a few small welds are added to hold them together.  The shape of the steel plates is unique for each motor, as the “magnetic circuit” is different for each motor, so I think if we can find a source for the steel we can get them cut.

This weekend we worked on a few little things on the Camaro.  The driver’s side power mirror won't tilt up or down anymore, but I could hear the motor running.  My father showed me how to disassemble the driver's side door and remove the mirror.  When we got the mirror apart, we found a broken part.  There is a 6mm screw with a weird head that is spherical, with a pin driven through it.  We figured out it was molded plastic, and being 30 years old, it’s not a surprise that it broke.  I checked the rack of hardware in the basement and found that we had some stainless steel bolts of the correct size thread.  I used the spherical turning attachment on the metal lathe to form a ball shape on the end of the screw where the head of the screw had been.  I then mounted it in a vise in the milling machine and drilled a hole through the center of the ball feature to insert a pin.  The parts turned out great, and the power mirror works again!

The other thing that hasn’t been working as well as it should is the remote to unlock the car.  It only works when you are standing right next to it.  My father had fixed this before, and it was due to a corroded connection at the lock/unlock module.  He showed me where it was located, and how to take it apart and clean it.  Now it will work from about 20’ away - not great, but much better!

We looked through the parts we have on hand, and also at a couple of cordless saws that my grandfather gave to us, as we may be able to use the matching motors and gears from the saws to power the wheels.  We also have two cordless drill motors that could also work, but we will choose the best and easiest option for the first prototype just to see if it can be ridden. I also found the belts, pullies, and solid rubber tires that I had ordered previously. These tires are especially important because they are airless they have no stem to put in the air and therefore nothing to interfere with the Lemniscate design.

This weekend, I mainly worked on updates and modifications to the 3D model of the Lemniskate. It’s been a while since I last looked at the CAD models, and I opened them and updated them to the latest version of SolidWorks.