Making hardware a little less hard
A technical look at the heart of Electroloom — the pump
This is the second piece in a two part technical series about Electroloom.
Check out part one about Electroloom’s switch to a GUI system.
The prequel — doing it by hand
Unlike traditional 3D printers that use rolls of plastic filament, Electroloom’s future fabrics begin life as a liquid. We mix up concoctions of solvents, polymers, and cotton constituents on used lab equipment reminiscent of Professor Snape’s potions class. Once our solution is ready, we slowly pump it into the Electroloom, where the actual fabric is born.
Initially, our entire process was done by hand. We filled syringes with our solutions, connected some circuitry, and tried our best to expel the syringe at a constant rate. This approach made for a good proof-of-concept, but it was incredibly inaccurate, not to mention a little exhausting.
The pump’s early days
To gain precision, and take out some of the manual work, we decided to build a syringe pump. Syringe pumps are commercially available, but expensive, which made them out of the question for an early stage prototype.
Our first couple iterations were about as simple as a syringe pump can get — a few feet of aluminum extrusion, a stepper motor, and a pulley. While better than a manual approach, they were cumbersome, difficult to position, and user-unfriendly. To alleviate some of these concerns, we decided to mount the bulk of the pump (along with the supporting electronics and controls) inside a housing dubbed “The Magic Box.” With that setup we only had to swap out the syringe each time we ran an experiment. This was the state of things in late June 2014, when we demoed our machine to Boost VC (we were accepted and completed the program at the end of October).
Micro steps
At this point we realized we needed to streamline our syringe pump setup. While our previous pumps had been standalone units that only supported a single nozzle, we wanted the next iteration to be fully contained within the Electroloom, and allow us to provide fluid to multiple nozzles at once (to improve coverage and speed). Connecting a single syringe to a manifold with multiple nozzles proved difficult, since a clogged nozzle just resulted in fluid diverting to the remaining open nozzles (our fluid operates at a low pressure, and the path of least resistance was always through a different open nozzle instead of building up enough pressure to push past the clog). Because of this, our pump had to support multiple syringes, so that each syringe could connect directly to a nozzle, and maintain enough pressure to keep the system clog-free.
The pump above is what we ended up with. It still uses the same basic components, but this time we upgraded to a high torque stepper motor to give us the power to push three syringes simultaneously. When mounted inside the bottom section of the next Electroloom model we had just finished, it made for a self-contained machine that could sit on a desk and run from a single power cord (a big step up from our previous mess of wiring and power strips).
Our first major redesign
As the Electroloom continued to grow in size and complexity, it became clear that our syringe pump needed a big upgrade. This time we overhauled it from the ground up, and came up with a design that allowed us to run eight larger syringes simultaneously. We used the same high torque stepper motor, but this time connected it to a leadscrew-driven platform. Overall, this pump worked great. It allowed us to quickly produce much larger pieces of fabric than anything our old pumps could have handled. There were, however, a few drawbacks. The biggest hassle came from having to load/unload eight syringes at once. We knew we needed to draw our fluid from a single reservoir to make the Electroloom user-friendly, but our previous trials with a syringe-manifold system had only led to problems.
Our second major redesign
After a lot of design work, we came up with a peristaltic-style pump that uses rollers to pump fluid through individual tubes. By shifting the fluid driving force from the back end (the reservoir) to the middle (the individual lines between the reservoir and the nozzles), we are able to use multiple nozzles with a single reservoir (and the system scales a lot better).
We are currently running experiments with this pump, and are excited by the early results. As with each iteration, it will allow us to further expand the range of nozzles we can use, and bring down the overall job time. Once this pump is refined a little further, we may even offer a low-cost or open-source version on our website.
Moving forward
“Good design, when it’s done well, becomes invisible. It’s only when it’s done poorly that we notice it.” - Jared Spool
We hope that applies to our pump someday. Our users should have as seamless (pun intended) an experience as possible. Right now we envision that as loading a “pod” of solution as you would an ink cartridge, and forgetting about it. We still have a long way to go to get to that point, but we are constantly making progress. After a few more iterations, we hope you won’t even notice our pump quietly working away in the depths of the Electroloom.