Sometimes You Need a Thing

Necessity is the mother of invention. -Proverb

This is a story about solving a simple problem with a lot of technology. What could have been a makeshift workaround became something different. I don't know if we've coined a phrase for it yet, but we probably should. 

That being said, this story is for entertainment purposes only, and designs associated with this story are to be used only at your own risk. 

Unforeseen Problems

The story goes like this. I inherited a vehicle rooftop storage box from my father. For weeks after getting it, we'd been planning a trip where we intended to use it, so we'd already been planning all of our packing around having it. 

Being foolish, I automatically assumed it'd fit the stock roof aero cross bars on my wife's car. Being an engineer, I vowed to test this well before we attempted to make a trip with it. Being a father, I didn't get to the test phase until the night before the trip.

Immediately upon getting the box on the roof I discovered that I had a big problem. The clamps for the roof box, at their tightest, weren't even close to keeping the thing secure. There was a visible gap on both the top and the bottom. The clamps would close tight enough to keep the thing from falling off the roof, but it would slide back and forth the whole trip. Completely unsafe. We would have to ditch the thing and somehow jam everything in the car or come up with a safe solution. 

Time Pressure

We're about to go on a week-long trip to the beach with two children under five. If you have two kids under five, you know the support apparatus required to go on such a trip is extreme. From absolute necessities to "required" items by the kids, to convenience items to make the parents life easier, it's a lot of stuff. 

When you're banking on having an extra couple hundred liters or over a dozen extra cubic feet of space, suddenly discovering that is not an option hours before a trip is a major issue. This is a problem that has to be solved. Now. 

T-20 hrs to departure. 

My first thought was jamming a rag in there, clamp it down real tight, and hope for the best. It'd probably work, too, said the improviser in me. It was the only thing I could come up with from materials in my house, and all of ChatGPTs suggestions involved items I didn't have (strips of rubber, foam spacers, etc.) But why use a rag when you have a 3D printer and TPU? 

Planning

I moved the box into the garage and set about coming up with a solution. From my wife's perspective, I suddenly began rushing about with a set of digital calipers. 

I was rewarded with a pleasantly tolerant yet subtly skeptical look. She knew what I was up to. She knew how close we were to leaving. 

The subtle skepticism was absolutely warranted. She was also off to pick the kids up from school, but I had a problem to solve!  

On my part, thus ensued phone camera shots of all the various dimensions measured by the digital caliper. 

Width of the clamp. Gaps between spaces opposite the clamp. Length of the space where the aero bar intersects. All along I'm building a mental map of design requirements. 

Whatever it is, it needs to fill this gap and keep the box from moving around, ideally be easy and quick to design and print, and definitely not be worse than a rag wrapped around a bar. Wind is a factor. Mitigating that will make it better than the rag. 

Solutioning

I could 3D scan the bar and the clamps, do some work with meshes in Fusion360, extract an exact fit... no. No time for that. Plus, I'd need the exact dimensions of that bar... Just one thing missing, the width of the aero bar. 

I opened the garage door to find she was gone. Off to get the kids. Oh yeah. Can't measure that. Guess that changes my design requirements. I'll just have to design based on what I know, and factor in some flexibility. 

I decide that best I can do is a pad that somehow fits into the grooves in the clamp. I don't know how thick it has to be, but I can ballpark it, and I can adjust easily if my model is inherently parametric. 

Fortunately for me and my limited time and even more limited design skill, everything seems to be achievable with a series of extruded rectangles. 

T-19 hrs to departure. 

Boot up Fusion360. Step one, load in a bunch of parameters. Pad length, width, and height. I figure I need two prongs to fit into these grooves. Prong length, width, and height. Fill out those values with my measurements and guesses, and I have a starting point. 

Side note, does anyone else sit around with a set of digital calipers in their hands and just measure random stuff? I do, and it's probably weird, but it does help with ballparking measurements. Anyway, back to the design. 

The Design

I want this to be parametric so the next person with this problem doesn't have to design it themselves, so I start with a center point rectangle on the origin, and give pad width and length as the dimensions. I create a construction line from origin with a value of (pad_width / 2 + prong_width / 2). This should make the center of my prong rectangle consistent.

Extrude the pad to pad height, extrude the prongs from the same sketch to a value of (pad_height + prong_height) and use join, next thing you know, I have a possible thing. 

Note, as a complete amateur, I've learned that using fewer sketches and getting as much of your two-dimensional data in that first sketch as is possible seems to make fixing it later a lot easier. 

T- 18hrs to departure. 
T- 1.5hrs to dinner. 

Ready for a test print. My main worries are my prongs are either too thick or not thick enough. My goal is, it slides in easily but doesn't fall out. 

I, for once, take my own advice and fire off a test print by cutting the full design down to a size that has sufficient depth to prove it'll fit, without wasting filament (or, in this case, time) on a full print. 

I base this cut entirely on gut feeling and no math whatsoever, and land on about 5mm overall height when the original design was 50mm.

Note the highlighted filament selection in the image. More on that later. 

Test Print

Time to prep the printer. I've been running red PLA+ for a bunch of test prints since I have an abundance of it, but it's time to change to TPU. I set the extruder temp in Moonraker while fielding a call from my brother. 

While catching up with him I unload the red PLA+ filament I had in there and load the TPU. Clock is ticking, I still have to order pizza for dinner, drive to get it, return, and y'know, save the day with my hobby. For fun! 

Haste Makes Waste - Proverb

My filament unload macro has it pull the filament from the nozzle slowly at first then rapidly unload the Bowden tube. It does the same in reverse for a load, which, thanks to the excellent engineering of the Voron M4 extruder, works flawlessly until I try to jam some stringy TPU in it at those speeds. 

The extruder gets into its loading groove with the TPU and starts clicking loudly. Suddenly, I remember everything I just said about loading TPU. I hit the emergency kill in Mobiraker and realize I'm T-30 from desired dinner, T- 25 from expected pizza readiness, and T-15 from actually getting said pizza. Time to leave this mess for later. 

T-16hrs to departure
Dinner+.5hrs
Kids Bedtime -. 5hrs 

Ok, dinner is done, kids are occupied, time to figure this out. I go down to the printer and, just as I'd expected, it managed to load an entire Bowden tube worth of filament into some interdimensional portal in the back of the Voron M4. 

This is not the first time I've done this with TPU. I don't know where it puts it. I built the damn thing, there's not a lot of space inside it, certainly not enough to hold well over 10cm of TPU but sure enough that's what I pull out of the thing. 

I get it extracted, cut the mangled end, and load the filament down the Bowden tube, close up the extruder, and fire off the test piece at last. 

First Test Print

I return 20 mins later after I get a notification the print is done, and I pull it off the bed. Well, I pull most of it off the bed. The test print turns partially to spaghetti. If you remember the highlighted filament selection from the picture of the slicer, you'll notice I didn't change my filament from PLA to TPU so it printed too cold which resulted in poor layer adhesion. 

Fortunately, the test piece was at least coherent enough to check the fit. The fit was good, for moderately adhered spaghetti. Time to roll the dice. I checked my measurements again, decided on a whim to drop the overall length by 2mm, and sent it to the slicer.

Second Test Print (Or first of four successful prints, if I'm lucky)

Print time was two hours. OK, children will be up at 0600 if I'm lucky, which I'm not. It's currently 1800. Test print will be done by 2000. If things go well, I can fire off 3 more before I go to sleep. If not, hopefully this print is close enough where it will work, and I can make final adjustments to other three. Off to the printer. 

T-14hrs to departure 

Print is done. Time to test. Correct temperature and filament profile means it actually comes off the print bed without problems. Well, without structural problems. TPU on my build plate adheres fantastically. I use a PEI plate, nontextured, sanded with 1500 grit sandpaper. I did this to get proper bed adhesion for ABS, but it's done great in general. 

Peeling TPU off the bed is a slow and steady process with this setup. In the future, I may consider using the textured side, would love to hear input on that, but for now slow and steady. I like the flat finish. Anyway, after it's off the bed it's off to test. 

Moment of Truth

IT WORKS! I slip it into the grooves on one clamp and it fits in perfectly without falling out. Tighten the clamp down. I see good squish on both sides. Time to see if it is stable. 

I try as hard as I can, perched on the door and the seat, to shift the box and feel no movement. This is good for tonight, but a more rigorous safety test will need to be conducted before driving to the beach with my family. 

That being said, I head inside to share my quick thinking, design skills, and 3D printing victory with my wife. But it isn't victory until it doesn't fly off the roof. Either way, I need three more to finish the job. 

I drop three more on the build plate, hit print, and pray I don't have a print failure in the middle of the night.  

T-4hrs to departure 

I wake up in the morning to find that, miraculously, I did not have a print failure. After rushing around and getting the kids ready, I finally have time to install them all and do a proper drive test. 

I decide on a quick loop with some back roads, highway speed roads, and bumps. Windows open to listen for noise, sharper than necessary turns, hard stops, all good. Time to load it up and see how we fare on this trip. 

T+3hrs from departure

Trip was a success! No shifting, no issues, couldn't even hear the thing up there. I decided to leave it on the roof during our stay, despite some concerns about the heat, and suffer no ill effects from that. The picture on the right is how it looks after a week of being clamped down, driven to and from the beach, and left exposed to the elements. 

Overall, the few hours or so of design work, printing, and testing, has resulted in a successful trip and, hopefully, something that will help someone else with the same problem down the road. 

One of the things I really love about this hobby is the ability to solve specific problems like this in a clean way and on short notice. I wasn't able to find anything that I could purchase to do this, other than some rubber strips or something. 

Yakima didn't have a replacement that I could find. It's an older roof box so it is entirely possible replacement parts are out of production. Now, something that would be waste is now useful again. 

Epilogue

Two days after our trip, we're home safe, roof box off the roof, no issues with the pads at all. My wife reminds me that I said I'd print a replacement buffer ring for our deck table. She makes the excellent suggestion of using TPU. The umbrella hole ring has long since disintegrated, and the umbrella is starting to chip the glass. 

I take some measurements, sit down to design a part, and quickly realize I already have one. The pad design can be adapted to be a single long strip that I can put in the hole. 

Since I designed this model to be simple and parametric, I just load in new parameters, the model adjusts, and I print it. No design work at all, it's ready to print in 5 minutes. I never would have considered this use for something I designed to be a gripping pad for a roof box, but there it is. 

For now, this ends the probably way too long story of a simple design going from a single need to multiple functional uses. If you've made it this far, I'd love to hear some feedback. 

What are some designs you've found multiple uses for? What other uses could this design be adapted to? Comment below with your ideas, thoughts, or suggestions, and thanks for reading!