Since it’s Sunday and I love guestimation: To put this in perspective compared to the new 1.2 V, 3500 mAh NiMH batteries (available for ~$1 each on eBay), let’

Since it’s Sunday and I love guestimation: To put this in perspective compared to the new 1.2 V, 3500 mAh NiMH batteries (available for ~$1 each on eBay), let’s say we need 5 of them to go from 6 V to 5 V as they discharge. That’s a total of 17500 mAh (17.5 Ah). 5 V at 4 W is a current of 0.8 A by P = IV, so the batteries could run for just under 17.5/0.8 = 22 hours (slightly longer, since V > 5 which means I < 0.8). They don’t say how much thermal wattage the stove puts out, but 1000 W electric camp stoves are readily available so I will go with that. 2-4 W of electric power translates to about 2-4% thermocouple efficiency which is pretty good. I think perhaps in the near future we’ll see some innovations in heat engines like Stirlings which easily reach 50% of Carnot efficiency. They have a reputation for low power to size ratios, but that’s due to lack of compression (easily remedied if anyone, well, tried): https://www.youtube.com/watch?v=H_Vnxapd5fs https://www.youtube.com/watch?v=H_Vnxapd5fs The Carnot formula is: Carnot efficiency = (T high - T low)/(T high) So to boil water at 212 F (100 C or 373 K) at an ambient temperature of 72 F (22 C or 295 K): Carnot efficiency = (373 - 295)/(373) ~= 21% So a Stirling engine would achieve 10.5% at the heating surface, or about 2.6 times better than a thermocouple. However, if we put the Stirling engine within the flame with modern ceramics and, say, graphite lubricant which can go to 842 F (450 C or 723 K), we could run at say 500 F (260 C or 533 K): Carnot efficiency = (533 - 295)/(533) ~= 45% So a Stirling engine would achieve 22.5% in the flame, or about 5.6 times better than a thermocouple. Of course then you are dealing with moving parts and it would probably be unreliable. But the whole engine could be made airtight with the generator enclosed inside, so it might be possible to build one with a service interval of tens of thousands of hours or more. Also I’m probably being too conservative on Stirling engine efficiency and too optimistic about thermocouple efficiency, so a Stirling engine is probably more like 5 to 10 times more efficient than a thermocouple. I’ve often wondered if it’s possible to build a solid state device that raises the frequency of infrared radiation to the visible range so that it could be converted to electricity by a high efficiency photovoltaic cell. It can’t be done by concentration alone because the laws of thermodynamics state that concentrators can’t reach a temperature hotter than the source (5300 C for the surface of the sun), although using multiple cells optimized for specific frequency ranges seems promising: https://en.wikipedia.org/wiki/Concentrated_solar_power#Efficiency https://en.wikipedia.org/wiki/Concentrated_solar_power#Effic... I see that as perhaps one of the great engineering challenges of our time, because low temperature gradient free energy is all around us (orders of magnitude more than we could ever use) but we have no way to gather and concentrate it cheaply. So far this is the most promising method I’ve found for visible light: https://en.wikipedia.org/wiki/Luminescent_solar_concentrator https://en.wikipedia.org/wiki/Luminescent_solar_concentrator