Oroborus; Did you get a chance to open the link to the Fine Homebuilding article? I built that stove under a greenhouse/atrium space attached to a residence in Missoula, Mont. in the 80's. It had approximately 40 ft. of horizontal flue run in five runs with four 180 -degree turns, and it drew like the space shuttle taking off. Draft in these devices is rarely a problem, and it is a problem easily alleviated. Notice the
cast iron clean-out door at the chimney base, above the floor, in the left side illustration? That's critical for cold weather operation. If the stove is not kept warm and the inner bricks are allowed to 'get cold', the stove will draft backwards upon firing, with quite unpleasant results. The clean-out door at the chimney base allows you to establish a positive draft in the stove BEFORE lighting a fire in the firebox. Once draft is established in the right direction, the main fire can be lit and will draw properly until the fuel load is consumed.
Cobb is a cheap alternative to high-quality refractory materials. In the primary and secondary combustion zones, I would suggest spending the extra bucks for top notch high temp materials. Downstream, I would be more apt to switch to cheaper materials, as anticipated temperatures should not require expensive refractory to guarantee component longevity.
Secondary combustion air is certainly a design requirement, and has been handled by other designers in a variety of ways. Pre-heated, secondary air, applied strategically to the flue gas stream where it passes from the firebox proper into the secondary combustion chamber is the ideal location. This delivery system might be constructed of stainless pipe exposed to the fire and connected to and pulling its air through a hollow door frame.
One design element missing in most of the talk I've read about rocket stoves is the relationship of excess air control to maximizing combustion. What I found in operating the k'ang in Missoula was that, prior to having an air-tight door in place, the stove would roar away with itself, taking in all the air it wanted and sending a great yellow flame all the way down the first flue run, and presumably, beyond for who knows how far? Yet, there was always some smoke associated with this type burn.
When the door/ air control was installed, turbulence was minimized, the yellow 'thrusters' turned blue and shortened, with the result being much more heat was dumped into the floor over the expansion chamber/first flue run than had been the case previously. Essentially, their is an optimum balance to be achieved through carburetion and an air-tight door, with controls, allows one to experiment with settings until a satisfactory balance is attained.