In order to resolve the disagreement between the wet steam hyposesis and the water spill-though hypothesis it's reasonable to ask how much energy it takes to break water into droplets and lift these a few inches before sending them out the exit of the rossi device.
The energy requires to increase the water surface area is given by the surface tension in the equation dW = gamma dA. W is the energy input , A is the area of a droplet and gamma is the surface tension. The surface tension of water is 59 mN/m (wikipedia on surface tension of water at 100 C). For a spherical droplet of radius r, W = 59 * 4 * pi * 10^-3 r^2. W is in Joules, and r is in meters. A good value to pick for the volume of liquid required over an hour is a little under 7 liters, or 6.75 liters. The remaining 0.25 liters leave the device as vapor. 7 liters/hr has been one value quoted. Each droplet carries off a volume, (4/3) pi r^3. The most error prone part of this exercise is determining the nominal water droplet size that will be lifted off the surface to exit the chimney. We may be able to establish and upper bound on droplet radius, r_u, where half the droplets of the radius r_u will exit the device, and half will drop back to the surface. It should be noted that smaller droplets carry more surface tension energy per unit mass than larger droplets. If an upper bound on r_u can be established, then a lower bound on the required energy can be established. The mean time it takes for a droplet to leave the surface and find its way to the exit is dependent upon the mean path it takes from the surface to the exit. This is dependent upon the height, h of the exit from the liquid surface, so any results obtained will also depend upon h. The time it takes a droplet to fall is a function of the radius of the droplet and the dynamic viscosity of steam, over microscopic dimensions in the order of the droplet radius. --------------- The gorrilla in the room that is hard to ignore is the energy efficiency. How efficient is the process of taking water from a surface, breaking off tiny bits of it, and suspending it long enough to leave through an exit at height h. This process is initiated by the energy imparted to vaporized water rising through the liquid, surfacing, and breaking off small pieces in the process. This part of the analysis at first sight seems intractable without emperical evidence. However, steam bubbles will have a terminal velocity in rising through a liquid. If their nominal size could be known, an upper bound on their energy would be known. This would place an upper bound on available surface tension energy. Area increase is proportional to surface tension energy. But we care about the volume of water generated, not the area generated. So the upper bound on the suspendable volume of water also depends upon nominal droplet size.