Designing an Automated Subsystem While Managing Constraints

Designing an Automated Subsystem While Managing Constraints

Russian composer Igor Stravinsky, whose compositions sparked riots, once touted the benefits of working within a narrow framework. “The more constraints one imposes,” he said, “the more one frees one’s self.” Constraints, he continued, can lead a designer to precision in execution.

Sure, Stravinsky was talking music, but his point has broad appeal – even in our field of medical device design. Clients often bring an array of constraints, and they’re not a bad thing. They can even inspire us to make something better.

A few years ago, Key Tech was approached by a global medical diagnostics company that was modifying an existing instrument to analyze human tissue for disease markers. The company had hit a roadblock. They needed to automate removal of waste liquid from the patient sample after centrifugation, and the subsystem had to fit within the existing instrument like a piece in a puzzle. What’s more? They were on a tight schedule and needed a safe, efficient module as quickly as possible. The subsystem had to be able to grab an existing specimen container, invert it to decant the waste liquid while retaining the sample pellet, and prevent spills at every step of the way.

Such were our constraints: Build a tissue intake subsystem, fit it into an already existing instrument and do it on a tight turnaround. Our experience with that subsystem – and other similarly constrained projects – has taught us some crucial lessons and we’re happy to share five of those, here.

1.    Be nimble, and ready to pivot. Our client’s automated system was following a sample preparation protocol that was already approved by the FDA. That procedure involved decanting waste liquid from sample tubes after centrifugation, and then blotting the edge of the tubes on an absorbent pad to ensure no fluid remained on the lip. The client’s original requirements called for a self-contained module with integrated electronics and firmware that replicated these manual steps, and we quickly developed a successful prototype. During the design process, however, we realized that we could eliminate the need for blotting by automatically tapping the tube lips and demonstrating that simple procedure performs as effectively as the FDA-approved blotting process.   This idea led us to invent a decanting routine using the robotic pipettor heads already used for other tasks in the larger instrument.  Custom rotating grippers were designed to interface with the pipettors, allowing the pipettors to perform the decanting action within the existing workflow.  In the end, it was simpler, better, and saved our client money.

2.    Communication is key. Plan weekly check-ins.  Active collaboration is important on every project, and the benefits of constant contact can’t be underestimated when the clock is ticking. Building a module for an existing system required us to work hand in hand with the client, the third party instrument supplier and other vendors, to satisfy subsystem requirements and to make sure our design would work at its interfaces with the instrument. As with every project, we met weekly to make sure no details, no matter how small, fell through the cracks.

3.    Collaborate with manufacturers as soon as possible. Some of our clients have their own in-house manufacturing capabilities, and others work with contract manufacturers.  In either case, and especially when you’re designing within an existing instrument, the manufacturer should weigh in early. Specific feedback on manufacturing processes helps to tailor the design to ease transfer down the road. In addition, a knowledgeable manufacturing partner can begin to scale –   up in parallel with final design work, getting to the finish line faster.

4.    Get prototypes early, and test, test, test. This one is obvious but needs repeating! The sooner you have a prototype in hand, the sooner you can see if it works within the overall system. In our case, a functional prototype of the initial design intent inspired a pivot to simplicity.   Quick turn fabrication vendors are your best friend, as getting parts back in days instead of weeks allows for more testing and iteration cycles. In our case, starting from scratch, we were able to design, test, and iterate each of two different subsystem designs within the larger automated system twice over a four month period.

5.    Never skimp on the check and review process. I may have saved the most important for last here. With the clock ticking and the deadline looming, you might feel pressure to skip or accelerate the review process. Resist that temptation!  This is especially true when working on a complex integration project like this where multiple parties are working on different parts of the system.  Does our subsystem footprint allow for the pipettor paths programmed by our client’s automation engineers without collisions?  Has the instrument vendor confirmed that their pipettors can withstand the loads imparted by our custom – designed grippers?  If you cut corners, you’ll usually end up missing something, and the schedule will need more time, not less, to fix that thing you missed.

Creating a subsystem with no time to spare sounds daunting, but these challenges cause us to recognize and benefit from newfound resourcefulness and nimbleness. Tackling tricky projects might even be your sweet spot. We like to think it’s ours. Knowing our strengths, we can quickly assemble the right mix of expertise and experience to turn constraints into focus and precision. Reach out and orchestrate a win with Key Tech. Like Mr. Stravinsky, we welcome working within a narrow framework.

Eric Schneider

Eric Schneider

Eric’s particular areas of interest include injection molding, sheet-metal, and precision machined part design, UL, CE, & IEC code compliance, finite element analysis, and electromechanical system design. He received a BS in Mechanical Engineering from UMBC (Magna Cum Laude). Eric is the currently the vice chair of the ASME Baltimore section and is working toward obtaining his PE license.
Eric Schneider


Every challenge is different – Tell us about yours.