The Back Porch
Embedded Files
Know When to Fold ‘Em
By
Johnathan W. Carson
Originally published July, 2026
One thing I learned very early in my career was that sometimes the worst thing you can inherit isn’t a difficult problem; it’s someone else’s bad solution. My first real job after joining the Mechanical Design Engineering Section at JPL was to be the THz Structure Cognizant Engineer (Cog E for short). I was responsible for the design, build, test, and finally delivery of the structure for the THz instrument. This structure needed to support a scan mirror, a telescope, sensors, and electronics. It also had a minimum first mode requirement of something like 40 or 50 Hz.
Before I took on the role, someone else had already started the design. This existing design had already been presented and accepted by the project as the baseline structure design and I was told to “make it work”! The design I just inherited was quite simple. It consisted of a vertical plate, which looked like a tombstone, bolted to a horizontal plate with two struts bracing between the two plates on the sides. I spent the next few weeks working with a structural analyst trying to make that structure work, but nothing could get it above 3 Hz. We tried increasing the thickness of the plates, no real impact (unless we made them comically thick). We tried changing the material, barely helped. Adding more braces, nope, still nowhere close! No reasonable combination of materials, thickness, or braces would make much more than a dent in this flawed design. Adding material would increase the stiffness, but the increased mass would bring that frequency almost right back to where we were before. The problem was the design approach itself. It was clear that we weren't fighting a weak structure—we were fighting the wrong idea. What we had was an open box section that would rack and twist like a potato chip. Anyone who's ever twisted the ends of an open shoebox has seen exactly what happens. No amount of additional material or brackets was going to solve that problem. The structure was dominated by local bending and nothing except a new design approach would fix it.
I did my best to reason with my project manager to let me redesign the structure, to throw out that flawed design and start from scratch. He was very resistant. The fact that I was young and only had a couple years of experience also worked against me. Nobody knew me or what I could do. Looking back, I can understand his hesitation. I was a young engineer asking the team to abandon a design that had already been accepted and begin again from square one. By that point, the project had already accepted the baseline, interfaces had begun to solidify, and changing direction carried real cost. In hindsight, the resistance wasn't surprising. When others finally realized there was no way we were going to get where we needed to be with that initial design, the project manager relented.
As soon as I got approval, I went off and thought about the problem for a couple of days. Then one night, while trying to fall asleep, the design came to me; a space frame! My new design concept had two triangular aluminum bulkheads, connected with carbon fiber tubes forming a truss structure. The only complication was that one side required the use of a K brace to accommodate the optical path. Now I had a space frame, and it was considerably stiffer by design. Every structural element was now loaded in either tension or compression. No more bending! No more racking into a potato chip!
When the first analysis run of this new design came back, we were very pleasantly surprised. We had spent weeks fighting just to get the original structure up to 3 Hz. Now, virtually overnight, it jumped to just over 80 hertz, and it was about half the mass of that 3 hertz Tombstone! Doubling the frequency requires roughly four times the stiffness, so jumping from 3 Hz to over 80 Hz wasn't an incremental improvement—it was a completely different class of structure. As you may imagine, my boss was quite happy, and I made some good impressions with my management. One of those managers still talks about that structure today. This structure launched on July 15, 2004 as part of the Microwave Limb Sounder instrument. As of this writing, it’s still functioning and continues to return data about the Earth’s upper atmosphere.
The lesson that hit me hardest at the time was that local bending is absolutely evil. Avoid it like the plague! That stuck with me for the rest of my career and I’ve always been conscious of that. Looking back, the biggest lesson wasn't about structural dynamics at all. It was about recognizing when you're focused on solving the wrong problem. We spent weeks trying to improve a fundamentally flawed concept because that's what we had inherited. Once we stepped back and questioned the design itself instead of endlessly refining it, the solution came remarkably quickly. Sometimes the fastest way forward is having the courage to throw away the baseline and start over. I didn’t realize it at the time, but that experience helped to form and cement my current design philosophy.
As for how I'd approach the project today...that's a story for another time.
What I Learned:
Sometimes the most valuable engineering decision isn't finding a better solution—it's recognizing that the current one cannot succeed.
Knowing when to start over isn't giving up; it's good engineering.
Takeaways
Question the design, not just its dimensions.
Don't let sunk costs keep you on the wrong path.
Always be mindful of your load paths; local bending kills!
Sometimes the fastest solution is a fresh start.
Explore Further
Before you go, here are a few related conversations.
Page updated
Report abuse