TBBT: Portable Shelters

The third challenge of The Big Brain Theory required the teams to build a man-portable (i.e., something that could be carried by one person in a small bag) shelter that could be set up in 5 minutes to ‘survive the elements’ – namely, 200 mile-an-hour winds, flames from a flamethrower, and a burst of water from a giant water cannon. Check out the full episode here.


The opening of this episode took place in a fire department training ground. They had a portable jet engine, a gimbaled flamethrower, and a gigantic, fire-hose fed, pneumatic water cannon pointed at a pre-fab shed they built on a concrete platform. The first element they showed us was ‘wind’; the jet engine spooling up to top speed. Almost nothing happened to the shed, other than the plastic chairs and standing umbrella scooting out of the way. The second element they showed us was ‘fire’, where they shot a relatively cold (the more red and smoky a flame, the colder it is) flame into the shed for 10-15 seconds. This scorched the shed, and started a small fire, but otherwise had no effect. They then demonstrated ‘water’, and brought out one of WET Design’s custom waterjets pointed sideways. This blew both the front and the back off of the shed, much to our surprise. We then proceeded to the Blueprint challenge.

At this point, they had simply told us the shelter needed to be moveable and deployable by a single person, and hadn’t elaborated further. I then proceeded to design a modified, armored hand truck that could be quickly wheeled somewhere and flipped over, and would form a hardened carapace for the wearer when on the ground. This, I reasoned, could be extremely well-engineered and relatively light-weight, and would offer rigid protection against the most damaging element by far – the water cannon.

Episode 3 Blueprint

Unfortunately, the judges chose to inform us that we needed to both a) fit in a small duffel bag and b) offer equal 360 degree protection after the Blueprint challenge was over. These extra requirements excluded everyone’s designs except for Amy’s and Alison’s, and was a point of significant frustration for most of the rest of the cast. In engineering practice, you need to know your project requirements upfront to be able to design a good product – in this particular case, we weren’t granted that.


Based on her frustrating experience with my leadership in Challenge 2, Amy picked me last. I admired her ability to put that experience behind her and pick me for her team again, and used this as an opportunity to switch into a strong design role.


Amy started the ideation and brainstorming process in her characteristic way, by having everyone draw their ideas for shelters on the boards. In this particular instance, it really felt like a second ‘Blueprint challenge’, because the requirement of the duffel bag was so limiting that nobody felt comfortable with their original idea. Interestingly, Amy almost immediately discarded her own turtle shell idea as infeasible due to our inability to work with inflatable devices in a timely fashion, and put the onus on the team to come up with a different solution.

Truth be told, I didn’t really participate in this whiteboard process, as I was thinking in overdrive about how the challenge worked with new restrictions. In this case, we were designing one component – a deployable structure that had to collapse to the volume of a bag. Instead of drawing on a whiteboard, I created to-scale models of a human being in fetal position (the smallest position configuration) and the bag itself. I then attacked the problem by creating cardboard sketch models, using the physical prototype volumes as a basis for the design. My rule was simple – if it didn’t work in cardboard, and didn’t fit in a to-scale bag, it wasn’t a reasonable design, period. I ignored the whiteboard as a medium, by and large, because I feel like drawings on whiteboards aren’t to-scale and aren’t good representations of what is physically possible – both of which are extremely dangerous pitfalls to succumb to in a 3-day design challenge. This process is how I arrived at the design of the folding geometric structure that was adopted as the team’s design.

Episode 3 Cardboard Sketch Model


The final design required a person in fetal position to hang from a lid on this conical figure (thus significantly increasing the mass of the device, which was severely limited by the portability requirement), covered in a fireproof and heat resistant shroud.

Episode 3 Design

The biggest challenge with this particular design was that the hinges had to be able to a) fold more than 180 degrees, b) stack as flat as the panels themselves, and c) weigh almost nothing – as we were going to need upwards of 50 linear feet worth of hinged surface. Conventional pin-based door and piano hinges were immediately discarded between those three requirements, and we had to improvise. I proposed a design for rolling band hinges (where two rounded edges roll off of each other, and are prevented from separating by steel bands – offering a full 360 degree range of motion for no weight), and we built up a full hinge out of plywood and regular printer paper for testing. It’s worth noting that even when the bands were made of paper and stuck together with spray adhesive, it was impossible to rip the panels apart by hand – how cool is that?

Rolling Band Prototype

In the final design, the panels themselves were made of 3/8″ thick Lexan – an incredibly tough, thermally insulating, transparent polycarbonate material frequently used to make bulletproof glass, shield Battlebots from weapons, and hold back the weight of water in aquariums. It was actually our second choice for panel material – Corey and I attempted to design honeycomb panels (featuring an aluminum honeycomb glued to top and bottom sheets), but couldn’t find raw materials that were both cheap enough and light enough to substitute for the plastic in the time we had. The panel thickness was designed to stand up to the waterjet cannon hitting dead center, which we estimated to be a continuous load of up to 1,600 pounds of force.

The panels were covered by high-temperature silica cloth, the most-insulating fabric material I could find on McMaster-Carr, our industrial supplier of choice due to their same-day turnaround. The cloth was so effective that I could hold a MAPP gas torch against one side of the fabric, and not feel anything more than general warmth on the other side with my bare hand. The only significant downside was that it was almost impossible to sew, as the fibers didn’t respond well to thin thread – in the end, Amy devised a method of installing large grommets in the fabric in order to keep it cohesive. It was ugly, but very effective.

One final addition to the design was made in the second half of Day 2; Dan suggested we add an underlying metal frame that would be responsible for keeping the panels a certain distance apart, so that the assembly wouldn’t collapse in on itself when hit with water. This was deemed to be a good addition to the design, and Dan spent the last day and a half making this system with Amy’s help.


This is the first challenge where we started to make significant use of WET Design’s automated production capabilities – a resource we had access to that hasn’t been publicized much on the show. I designed the structure to be cut out by CNC waterjet cutters and metal-cutting lasercutters, so we essentially just shipped them full 4’x8′ sheets of Lexan and stainless steel and they shipped us cut-out panels and hinges some time later. While we were waiting for these parts to be cut for us, we decided to build a full-scale wooden mockup using the wood panels, paper hinges and spray adhesive.

Prototype Shelter

The mockup gave us some early test time with the design, allowed us to make sure it could be packed into the duffel bag, and verified that we could fit a human being inside once deployed. On top of all that, it won us major points with the judges. Soon afterwards, we received our raw panels and bands, and began the production process in earnest.

Episode 3 Bare Panels

Each panel needed to have both sides rounded (by a specialized rounding end mill), needed significant sanding, cleaning and scoring at each hinge site (as we were both screwing them down and gluing them together with JB Weld, per Corey’s suggestion), needed 30 waterjet-cut holes countersunk, and needed 30 self-tapping screws drilled into them on two sides. With 16 panels total, this came out to a huge amount of mind-numbing work that had to be done in a very particular order. We formed an assembly line and got to work, doing little else on Day 2. We were done with panels and their assembly by the middle of Day 3, and spent the rest of Day 3 working on a lid, creating an underlying metal structure, making final adjustments to the overlying fabric, and arguing.


At some point in Day 3, Dan got it in his head that we must not be allowed to slip on the concrete while being hit by a giant cannon. Corey argued (and backed up with simple math) that we would be hit with far more force by trying to stay in one place, than if we simply allowed ourselves to slide. Dan wouldn’t stop screaming and yelling (even uttering the ridiculous absolute statement ‘if we don’t add neoprene, it will fail’ – since when is anything ever so certain?) until the four of us gave up and installed the neoprene strips on the underside of the metal frame. This is the one point in the competition where I really felt Amy could’ve done a better job stepping up and making the correct engineering decision that was desired by most of her team; otherwise, she had been a fine leader throughout the challenge. I will also say that Corey was unwilling to let this issue go until I told him that we would either win, or we would kick Dan off the team for being so illogical and needlessly combative if we lost.

Episode 3 Needless Confrontation

I do want to take a moment to say that up to this point, this was the most inappropriate behavior I’d ever experienced in any engineering team setting, in my entire life. For a team member to get defensive about a design feature to the point of yelling, screaming, and utterly refusing to compromise is inexcusable. I had been giving Dan the benefit of the doubt up to this point, but this started me on a path of seriously doubting his capabilities as an engineer.


All in all, we did fairly well. The shelter assembled quickly without a hitch. We survived the wind without a second thought (though the silica fabric took a little more of a beating than we expected). We actually got blasted with the flamethrower 4 times, from 2 different directions, because wind kept interfering with the blasts – regardless, the highest temperature inside our shelter was under 100 degrees Fahrenheit. Then, of course, we got hit by the water cannon.

Episode 3 Water Blast

The hinges failed between two of the panels, where they got hit by the cannon. It appeared that both the JB Weld failed, and the screws we were using ripped through the edges of the steel sheets.

Never forget – 2 out of 3 dummies prefer Red Team.


At the end of the episode, the announcer suggests that our screws were small and we had no washers on them. This is true – we used countersunk screws, which have small heads by default and don’t generally allow for washers. I imagine that by sinking the countersunk screws into the flat steel bands, we created stress concentrations that would allow for tears to propagate fairly easily to the edges of the material. I’m surprised the glue didn’t hold better, however – to be honest, we saw the glue as the primary attachment mechanism, and the screws as a backup. Mark suggests in his Blueprint breakdown that we could have spaced the screws further away from the edge of the bands, but we already had 3/4″ or more material there; I’m unconvinced adding more would’ve significantly improved our strength.

If we had to do it again, I think I would’ve wanted to switch from steel sheets to nylon webbing with built-in grommets for the bands. Nylon is generally much more resilient than steel, and much less prone to tearing in that form factor. I don’t think we could have easily ‘added washers’, as the announcer implied, because the panels would then not fold flat and the design wouldn’t have fit in the bag. I would’ve also wanted to switch from the neoprene rubber to a slick plastic that would’ve encouraged sliding (or perhaps a combination of slick plastic and neoprene, to encourage pivoting out of the way of the stream of water), so that we could get out of the way of the water faster and absorb much less force.