This is the final completed assembly, where the spindle is able to spin easily by hand. We verified that friction torque and the deflection allowed at the end of the spindle due to 100N force are lower than expected, and our spindle lifetime is calculated as higher than expected. 

Making the X-Slide

After careful discussion, the X-slide assembly consolidates to the following design. First, the motor and the handwheel are placed on opposite sides so that the assembly is more balanced when the user is operating it. A flex coupling connects the motor shaft to the motor, and shall be disconnected during manual mode so that the motor shaft is not damaged.

Second, a tapered brass gib is fabrictated that sits between the dovetail ways (the sliding surfaces of the male and female dovetail) for the following reasons:

• Over time, wear can cause a loose fit, leading to backlash or wobbling of the sliding components. The gib can be adjusted with set screws to eliminate this looseness, maintain accurate alignment, and compensate for wear. This helps extend the life of the machine parts and maintains the precision of their operation.

• A properly adjusted gib allows for the right amount of friction — enough to keep the parts secure and aligned, but not so much that it hampers smooth movement.

Given the softness of brass, warpage is a common issue during fabrication. I learned that for a raw stock, facing off all surfaces before milling it down to size can release internal stresses and reduce warpage.

The finished X slide assembly works as follows. The no-load repeatability, backlash, and X direction deflection under load are measured to be lower than required.

Both flexures are extremely challenging to fabricate on the waterjet, because both consists of thin wall sections that could easily break if the waterjet is not tuned well. After waterjet, there are also milling work that needs to be done, including drilling and tapping small holes. It took us 5 attempts to fabricate a functional dancing queen flexure.

After fabrication, potting is applied to 4 places:

• secure the rail bushings to the carriage skirts and rail flexure. With potting, we are able to reduce irregularties in our fabrication and make sure the X-slide sits level on the rail.

• secure the drive nut to inside the dancing queen flexure. With potting, we make sure that components must not move relative to each other, except along the desired degrees of freedom.

Potting is challenging in that we need to apply enough to securely fix two components, but not too much that it would flow onto the rails.

Inspired by the mandrel, which is a long piece that is held by the chuck and the tailstock that holds wooden pieces, we decided to incorporate a steady rest that would hold a metal piece of up to 10 inches. In this way, we could fabricate a pen (which is about 6 inches) in one go, and the steady rest could reduce deflection at the end of the workpiece.

Our initial plan was to purchase an off-the-shelf steady rest, which turned out to be too big and heavy for our lathe. Then, we decided to make our own.

The Working Lathe

Here’s the video of the lathe cutting stainless steel.

Good threads :)

Single point threading is an art by itself (2x speed)

• Here’s another example of bad threads. This is an internal thread of the nut that is potted by epoxy to our dancing queen flexure, which then connects to the entire X-slide platform. The nut goes onto the Z-axis leadscrew. Ideally, when the leadscrew turns, the nut and the dancing queen will move forward and backward along the leadscrew axis. In the following video, the internal threads are too shallow. It worked for the first 1-2 times when we ran the machine, but after multiple testing, the threads were destroyed under weight and I can even slide it back and forth! The most dreadful part is that in order to remake the nut internal threads, we need to also remake the dancing queen, which took countless hours of waterjet and milling time.

• The impact of bad threads are not obvious at first but detrimental after repeated testings. We learned that it is essential to practice thread making and making sure we are using the right tool!

• Sufficient potting is critical :))) First of all each potting requires 6-12 hours of cure time, so re-doing the potting put a lot of stress on our schedule. Secondly, we tried potting the drive nut to the dancing queen twice, but we were too conservative each time.

First attempt of potting. After curing there is so much gap between the drive nut and the dancing queen flexure.

• Waterjetting flexures :))) Because flexures flex, we learned to add additional support to the design, so that we can perform the waterjet and milling operations before cutting away the support. We also learned to consider so many details such as waterjet path, garnet level, and tab design. See below for 5 attempts to making the functional dancing queen.

The Finale: Lathe Death Test Videos

At the end of our project there are two death tests: the drop test and the hammer test.

As mentioned before in the potting section, the drop test disabled our Z-slide motion and we could hear clear “clanking” sounds when trying to turn the Z handwheel.

After the hammer hit our spindle assembly three times, they still turn well, indicating enough strength in the Z direction.

Huge thanks to my teammates: M. Patrick Serbent, Yasin Hamed, Olivia McGrath, Michael Lu, Carolina Warneryd, and Ruben Castro Ornelas. Huge thanks to Professor Marty Culpepper, Wade Warman, Laura Rosado and Alejandro Martinez for their help throughout this course.