Dechorionator

  Before delving into what was built, some quick context is probably needed. A dechorionator is a device that removes chorions from embryos. Chorions are the outer membranes of an embryo which provide protection, and a permeable membrane which can allow gasses and nutrients to reach the developing animal inside. As SARL is a toxicology lab, and its experiments need to be deterministic, this protective layer can drastically skew tests results, and even worse, can can variances embryo to embryo, or egg batch to egg batch. To remove this, a special protein is added to a petri-dish full of embryos, and then the dish is gently swirled with jerking start and stop motions. The goal is to provide light agitation between the embryo, the dish, and their neighbors, helping the protein eat away the chorion and sluff off into the dish. This can, and has been done by hand, but when I joined SARL they already had two machines which which could automatically perform this task. However, they were incredibly expensive and massively overcomplicated, requiring a whole table's worth of custom shaker units, networked peristaltic pumps, and servos. The engineering team was tasked with simplifying this setup while reducing both their size and cost.

  We started with a COTS shaker unit from the company ELMI, which had a stepper-motor-based drive system, making it a perfect candidate for easy retrofit. After gutting the existing electronics, and taking some measurements, I started on a custom PCB design. Basic requirements were that the board needed to be able to control the stepper motor, control the speed of a liquid pump, provide controls to users, allow for config editing from those controls, and provide a screen for cycle progress and editing those config values. Since this was one of the first PCBs I'd ever designed and hand-assembled, I started with a basic proof-of-concept which was for bench use only (revision: 1.1.0). While I worked on the electronics, my co-worker and good friend Dylan Thrush was busy designing a top-plate for the shaker to hold the dishes, shower them with water, and drain the pumped-in liquid.

  First tests showed that the overall concept was going to work, just needing signal conditioning for the rotary encoder to avoid ghosted or missing inputs. A larger problem we found was that the brushed-dc-motor driven peristaltic pump was not going to be able to supply the flowrate needed to properly shower the four dishes. We'd already chosen one of the highest-flowrate pumps which could fit inside the shaker housing, and ended up having to pivot to a much more expensive one from TCS Micropumps. Luckily, not only did it solve our flowrate problem, but also held up much better to the saltwater solution being pumped through it than our initial choice.

  With the proof-of-concept design functional, I began a redesign of the control PCB to replace the existing control panel and drive circuitry from the ELMI shaker (revision: 3.0.0). The existing control circuitry had a unique assembly design that I'd not encountered before, using a PCB with solderable standoffs as the front panel, and soldered copper strips as retention tabs. I was so fascinated by the design that I decided to emulate it. Check out the images at the end of the reel above to see how this unique assembly was put together! Around the time that the PCBs were ready, we'd hired a new engineer, Aaron Rito, who I tasked with writing the firmware while providing input and guidance.

  Over the next few months, many revisions were made to the firmware, as well as mechanical designs for the showerheads and water manifold. Dylan also had a final design for the top-plate milled out that looked beautiful. Once those changes were complete, we provided the prototype unit to the researchers, along with documentation on how to use the tuning values. They then spent a few weeks running the new dechorionator alongside the old ones, while tweaking these parameters until the performance matched. We then built up four more units, and pre-flashed them with this configuration. Three of these went into the lab, where a total of four of our new dechorionators sat on the same table where just two prior-generation ones used to live. It even had additional space for pre and post prep work on the petri-dishes! The last one I installed in a partner lab on the east coast after flying there with the head of the lab, Robyn Tanguay and deputy director, Lisa Truong.