Cross-Spool Distributor for Fiber Spinning
Background:
The fibers that are produced during melt-spinning are distributed along the takeup spool as they’re spun. This keeps the take up rate relatively even and prevents thermal buildup because the fibers are spread laterally rather than winding in one portion of the spool. Now that we can achieve consistent fiber takeup in our lab-scale spinline thanks to the updated spool and winder and custom hopper, the next step is to add a module for distributing the fiber along the spool.
This project was challenging and time consuming for multiple reasons. This was my first project including a microcontroller, my first project reverse engineering unknown electronic components, and also the components I planned to use were damaged at some point and I had to design extra components.
Criteria:
A fiber distributor module for our lab-scale spin line should:
Guide the fiber back and forth along the spool as the fiber is wound at a consistent and adjustable rate (pitch)
Have adjustable stroke so that spools of various sizes can be installed
Have a modular design so that it can be mounted wherever the winder is located
Mechanical Design (Take 1)
The core of this design is the linear motion. From rails to lead screws to electronic actuators there are many options for driving and constraining the side to side motion of the fiber guide. I found that most custom built solutions would run into the hundreds of dollars after all was said and done, so I chose to re-purpose a defunct inkjet printer in the lab. The printhead of the printer is constrained to a stamped sheet metal rail and moves back and forth on a belt driven by a DC motor. Thus, I had a ready means to make the back-and-forth motion required for this project.
The printer, an Epson Xpression Home XP-440, was previously used for experimenting with printed circuits and was no longer useful, so it made for a great opportunity to reuse some hardware that would otherwise go in the trash. I tore down the printer and extracted the rail with carriage, belt, and motor. The printhead’s location along the carriage is indicated by an encoder, but rather than reverse engineer the encoder ribbon I removed it and its associated sensor in favor of a pair of limit switches.
The rail with carriage, belt and motor as removed from the Epson Xpression 440 Printer
To keep the design simple, I retained the carriage and designed an armature to be 3D printed that would mount into the existing carriage where the ink cartridges would attach. Finally, I designed mounts for the limit switches and mounts for attaching the rail to the aluminum extrusions of the spinline frame, all to be produced via FFF (aka FDM). The rail mounts and limit switch mounts were designed to be a single part each that would slide onto the end of the rail and lock down with soft tipped set screws.
Circuit Design
I designed the circuit with a momentary switch to start and stop the motion, a trim pot for speed control, and microswitches to signal the end of the carriage’s travel. An Arduino Uno coordinates all of these inputs and outputs and the entire system is powered by the printer’s original power supply. I used the printer’s power supply because it outputs the necessary voltage (40V) for the DC motor, and I used a step-down breakout board from Pololu, D36V6F5, to provide 5V to the Arduino. An Adafruit DRV8871 motor driver breakout board is employed to control the motor from the Arduino. I connected all of the dots on a breadboard for simplicity and modularity. You can find my Arduino code here on Github.
Mechanical Design (Part 2):
Busted! A new carriage was required.
Sadly, I came into the office one day to find that the carriage had been broken. It had seemingly been knocked off the shelf, and the features that keep the carriage constrained to the rail had busted off. Irreparably damaged, I designed a new carriage and took the opportunity to make new limit switch mounts as well. The set screw fastening method did more deforming of the rail than holding the switch mount in place, and I learned that the first microswitches that I purchased were too stiff to take advantage of the lower end of the carriage speed range. I found different microswitches from Pololu with half the spring rate, and designed new mounts that used an aluminum strap across the back of the rail to clamp on the broad side and distribute the load. The new carriage was designed to use cheap plastic rollers with bearings to engage with the OEM rail. I incorporated a flat stage for mounting the armature that had already been printed for the original carriage. I had to recreate the belt pinching feature on the OEM carriage, so I designed a block for testing different bend diameters and belt gaps. In my experience, it’s always best to make a test part to get the critical gaps right because additive manufacturing tolerances vary. The final part incorporated pinches of three different radii so that I could change the belt tension on the fly by winding the belt through another or a combination of pinches.
Results
A spool of recycled nylon fibers produced on the Li Lab spin line. This was one of my best spools (aesthetically) and shows the even distribution thanks to the fiber distributor. 2g of fiber spun into a continuous 0.03mm diameter fiber.
Losing the original printer carriage turned out to be a blessing in disguise; The new carriage slides back and forth more smoothly than the OEM carriage and the new microswitch carriers are sturdier without deforming the rail. Now, I can spin fibers continuously and the fibers are spread evenly along the spool, which allows for more even cooling and a more consistent takeup rate. That’s not to say that some things couldn’t be better: I was conservative on the spacing of the rollers on the carriage, so above a certain speed the carriage can work its way out of the rail. On the other end, at too slow a speed the motor doesn’t have enough torque to move the carriage back and forth at a constant rate, so the usable speed range is small. The armature that guides the fiber could use a redesign as well. In a second iteration I would replace the roller with a smooth V-feature, perhaps made of bent wire, because the fiber can move off the wide, shallow roller if the extrusion flow changes. Nonetheless, the improvement to our melt spinning capabilities is immediate and immense. Next for the spin line will be improving the winder with a DC motor to replace the stepper and a custom control scheme so that we’re not limited to the 6000 turn max of the off-the-shelf unit.
ACKNOWLEDGEMENTS
This project was supported by the U.S. Department of Energy (DOE), Vehicle Technologies Office under the contract number of DE-EE0008195.