Pop-Pop Boats are notoriously inefficient for many reasons but the main problem is the action by which it moves forward.
When the engine pops it squirts water out of the exhaust. This water travels almost straight out with little divergence.
When the engine sucks in fresh water it sucks from all around the inlet. The water sucked in from the sides should be symmetrical so they sum to zero but approximately a third of the water in taken comes from the same vector as the exhaust squirted which cancels out say a third or more of the forward thrust when the net product is found.
Put valves on the engine so that it sucks from the front and blows out the back potentially quadrupling the forward net thrust!
Step 1: The Plan
With the concept set the principles of operation are simple. All that remains is to design an appropriate way of implementation.
These only have to be one way and should require no return so should take little force to open. I went straight for the shuttle valve. I found the smallest reasonably sized ball bearings I could and designed the valves around them so they would close from one side and free flow the other. Then, they would be mounted to allow free sucking from the front, no blow and free blow on the back, no suck. The ball bearings where 6mm steel.
Exhaust and Intake
I know through my research that the number of tubes is essentially irrelevant. One tube can work as effectively as two or more. The aspect to be considered is surface friction. As the output of the engine has an optimum cross sectional area the way to attain the least losses is to have one tube of the required diameter. Since I want to split the input and output I needed two tubes, one in the front for intake, one out the back for exhaust. I decided to have them both of the same diameter copper pipe of around 6mm internal diameter (1/4") To fit my valves.
calculated from working engines boiler size to tube cross sectional area ratios.
Here I have two choices, the simple coil boat or the difficult to seal pop pop boat.
There is a minimum radius to which you can bend a pipe without flattening or crimping it. My quarter inch copper pipe dictated that the coil would be rather large in diameter and hence hard to package in a boat hull of the size I was envisioning. I also know through research that pop-pop boats don't work well at larger volumes as the inertia of the water displaced with each cycle limits the frequency of operation to a point where there is little if any or the motion can be rather sporadic. This meant the internal volume of the coil that would form my boiler may be too large.
Classic Pop-pop Engine
These are simplest when constructed out of steel beverage cans. Instructions on how to do this are on the internet and I dare say here on Instructables. If you have trouble finding them, send me a message or e-mail and I'll either add a link here or send you the link. This would be good but the boiler needs to be at a fairly horizontal angle to catch the heat from the heat source. A vertical boiler would receive little to no heat as heat rises. This would mean that complex bends would be required to get the engine in the middle of these two pipes going on opposite directions. Fluid dynamics say that corners are bad as far as resistance and efficiency goes so this would not be optimal.
This left a completely fabricated boiler of my own design which was in the same size ratio as the smaller commercial pop-pop boats that you can buy but with an in line intake/exhaust orientation which made sense to me for the intake/exhaust placement I was envisioning.
Steel shuttle valves with a 1/4" copper pipe intake/exhaust and a conventional boiler with slight modifications.