Core B Design

Core B is a ridiculously simple stopper-style design that turns any soda bottle into a water rocket. This is a discussion of its origins and design decisions.

Design goals

The oversimplified story of Core B

It is tempting to write a narrative of linear invention. Something like:

The reality is far from this. Below is a slightly simplified version leaving out some of the boring parts.


Picture: Still frame from validation testing of Core B, October 2021

The just-enough simplified story of Core B

Simplifying assembly by removing PVC joints from Core A

Problem: the Launch Tube is hard to assemble. PVC cement is finicky and not everyone knows how to use it. It’s got nasty chemicals and exposure isn’t great. I bet assembly of the Launch Tube is a major barrier to this design being easy to make.

Idea: Maybe I can have fewer PVC joints. I can take off the top tube entirely. You lose the stabilizing effect of launching off a tube, but it also takes a lot of strain off the Tube Core.

Test: SN013 and SN014, designs without upper tubes. They work fine. It does mean that the water goes right into the Tube Core… but we’ll get to that later.

Removing the bolt?

Idea: Now without the upper tube levering the Tube Core, can I remove the bolt entirely? Two fewer pieces to buy. And wouldn’t need wrenches.

Test: SNO015 and SN016. Yes. Works fine. Do I really need the lower tube? Its job is to interface the tire valve to the O-ring. Could I do that directly and not need PVC cement?

Put the O-ring on the tire valve stem?

Idea: Put an O-ring onto the valve directly and interface it to the bottle.

Test: Ugh. My existing O-ring (7/8" OD, 3/32" thick, 11/16" ID) doesn't fit. And double-ugh: I'm using a funky imperial dimension almost certainly not available everywhere. And it's better not to have multiple O-rings anyway. Can I make one work with both Core A and Core B? Time for research.

Found that a 15 x 22 x 3.5mm O-ring fits perfectly in the groove on a tire valve and interfaces to the inner bore of a soda bottle. Great!

Core A will need some redesigning though, since that uses a different O-ring....

Now for a 3D part that holds things in the right place?

Idea: Make an 3D printed interface that holds the new valve/o-ring assembly and fits into the existing Base

Test: Designed, printed, and fitted a bunch of iterations. The extra length on the stem is useful. Decided to use the TR414 for both Core A and Core B. It does seem require a hose extension because there's no space for the pump to fit. Are hose extensions widely available? It looks like they are in one form or another. Good.

What about backflow into the bike pump?

Idea: I need a snorkel of some kind now that I don't have a launch tube. Seems like the hole in the wide end of the tire valve is a standard 7mm. I need a tube that's a bit bigger. Here's a stainless-steel IKEA straw... let me just cut this with a pipe cutter and sand it down and... wait. How about use a plastic reusable straw?

Test: An 8mm plastic straw fits snugly and doesn’t fall out. Great!

Does it work?

First test: Is it even pressurizing? It's eerily quiet and low-resistance and there are no bubbles. Maybe it's leaking somewhere?

Oh wait. There’s no leaking. At all. It's quiet because air goes straight up the snorkel to the top of the bottle. It's pressurizing the air pocket directly, and everything is airtight. 

Let's stress-test it then. Is it holding?

Picture: 110psi validation stress-testing of Core B, October 2021

How to depressurize if there are no leaks at all?

Now we have a new problem. Core A had pinpoint leaks that would slowly depressurize the bottle over several minutes if left alone. This was a useful safety mechanism for launch aborts.

Core B doesn't leak at all. How to depressurize?

Perhaps the internal spring valve core can be removed from the hose extension so when you disconnect the pump it will let out the pressurized air. The snorkel keeps the water from spraying out.

How to remove the spring valve core without specialized tools? 2 years later: oh, let's design a small printable removal tool.

Design tradeoffs

On launch, the stopper-style Core B immediately releases water from the bottle. It's therefore intrinsically lower-performing than launch tube configurations like Core A which conserve water for the initial phase of launch. But it's universal, remarkably robust, and assembles in less than a minute. It's a great addition to the launcher, but doesn't replace Core A.

Core A, however, can be daunting to assemble, requiring three off-the-shelf parts, three 3D printed parts (one of which needs supports and different settings), and two sections of PVC pipe that must be cut and sanded. Assembly requires wrenches, PVC tube cutter and PVC cement, a knife, gloves, and a well-ventilated workspace. Is it worth it?

What if a launch tube configuration could be built that requires no adhesives, just a pipe and the parts from Core B? I would have to figure out airtightness on at least some aspect of a 3D print. These ideas developed into Core C.