There is a lot of material on the web about building a propane poofer.  I won't go into a lot of detail about the plumbing aspects, assuming the reader can find this elsewhere.  I also leave this to the reader so that this project is not seen as a simple step by step task.  Due to the elements involved, a lot of knowledge is required to build and use this safely.  Instead, I will only provide an brief overview and notes about specific decisions I made note of.

A propane poofer uses LP(Propane, as is used commonly on outdoor grills and some home heating), evaporated and reduced in pressure through a regulator, and stored in a tank referred to as an accumulator.  The purpose is to allow a sizable volume of gaseous propane which can be vented quickly.  The LP in the tank is naturally liquid, and can only evaporate so quickly.  Attempting to evaporate faster can cause the system to literally freeze up.  In electrical terms, the LP tank can be thought of as a battery, and the accumulator can be thought of as a capacitor.(Ironic that capacitors used to be called accumulators?)

The gas in the accumulator is released through a solenoid valve, and is ignited, commonly with an open pilot light.

A variety of parts are involved in this, all of which must be correctly rated for propane in liquid or gas forms.  Many materials cannot tolerate propane exposure.  Further, it must be a leak free system.

For ignition, I will go with a hot surface igniter.  This will prevent the tell-tale pilot from being visible to the general audience prior to it's use.

I really desire a somewhat slow burn, not loud.  The intent will be to run at a lower pressure than many others, with maximum flow.  Reducing the pressure will reduce the jet like appearance of some examples out there, as well as the roar.

I am working on the theory that creating the slow burn is best done by retarding the mix of oxygen and propane.  I was unable to find any resources which discussed how to implement this, only various build descriptions and videos on the web.  In the end, I will have to build it, test it, and adjust from there.

Two ideas that I have not seen implemented come to mind to assist in this.  The first is to use a laminar flow nozzle.  I was unable to find data indicating if this concept works with gas, and would be happy to hear others thoughts.  This method only affects the output barrel, so will be postponed until after initial testing to see if it is worth the effort.

The second is to pre-launch unignited propane to allow a larger volume to be in the air just above the cannon for ignition.  Doing this requires a second release barrel which is shielded from ignition, and a timing controller which can sequence the two.  Though I could not find resources discussing the safety of this, the primary concern I can think of is the general release of propane which is unburned.  Testing will explore this, which will be done initially on a slightly windy day, in wide open air to allow quick dissipation of anything unburned.  This idea is to be built into the system from the start as it requires additional parts to be ordered at the same time.