I noticed that my windscreen was losing heat out the bottom, so I reconstructed it a little wider and more channels.
I bought Al flashing from big box store. 7" x 10 feet – $10.
I cut a piece 22" x 2.5". Folded 1/2 inch on both ends to interlock. Made 17 channels. Carefully marked lines spaced at 1/2 inch and then acordian folded over sharp edge. 2 inches at both ends without channels to keep handle from getting hot.
Then I put a #18 galvanized wire around it to keep it from falling off pot and to make good thermal contact. 1.3 ounces total:
I tried without wire and I could get it to stay on pot, until I turned on flame, aluminum expanded, and quickly dropped off pot. That's always been a problem with this design.
I made a lid out of aluminum with hole for thermometer. I measured temperature before. Ran stove until it hit 90 C and then turned it off. Then waited until the temperature increased a little more and recorded temp when it maxed out.
Weighed canister before and after.
I heated 15.5 ounces of water (what I need to make tea).
I calculated g/L/80C – how many grams of fuel to heat 1 liter of water a temperature difference of 80 C:
5 cases:
no 1 – no windscreen, Soto stove adjusted 1 half turn above simmer position
no 1/4 – no windscreen, 1/4 turn above simmer
yes 1 – with windscreen, 1 half turn
yes 1/2 – with windscreen, 1/2 turn
yes 1/4 – with, 1/4 turn
The "simmer position" is exactly 2 half turns. The stove will barely stay lit at this point.
When I turn the Soto 1/4 half turn past simmer, it takes almost twice as long to boil. The idea is to determine if turning flame down improves efficiency.
If you compare "no 1" to "no 1/4", it takes 3 minutes vs 5.3 minutes, and fuel use goes down from 14 to 12.1 g/L/80C. Turning flame down saves 15% in fuel usage.
If you compare "no 1" to "yes 1", it takes a tiny amount less time, and fuel use goes down from 14 to 12.4 g/L/80 C. Using windscreen saves 12% in fuel usage.
If you compare "yes 1" to "yes 1/4", it takes 2.8 vs 4.8 minutes, and fuel usage goes down from 12.4 to 11.2 g/L/80C. With windscreen, turning flame down saves 10%.
Comparing no windscreen, full flame, to with windscreen and turning flame down saves 20%.
With the Soto, there isn't much difference between the "1/2" position and the "1" position. I arbitrarily defined the simmer position to be exactly 2 half turns. If you turn it 1/2 half turn more, it's about as high as it will go. Turning it further to the "1" position only increases the flame a little.
I don't have good repeatibility data. But, if you compare the "yes 1", "yes 1/2", the time to boild goes up from 2.8 to 3.2 minutes, which is a small difference, and the fuel use goes down from 12.4 to 12.3 g/L/80 C, which is also a small difference, and going in the expected direction, which is an indication of repeatibility. Also, if you compare the "no 1" and "no 1/4" difference, it's about the same as the "yes 1" and "yes 1/4" difference, another indication of repeatibility.
