I haven't started my project yet - I'm still in the research phase. I'm leaning away from putting a coil around the barrel, and in fact leaning away from a hot radiating barrel at all and am going to build something that stores more into mass:
Given the potential dangers involved in heating water, Erica and Ernie have shied away from discussing this subject in these forums, which is why threads like these don't go far.
I found another forum where it's talk of RMH all the time (including heating water):
I'm thinking about putting a (probably unpressurized) storage tank inside the rocket mass heater. Thoughts range from a 30 or 55 gallon barrel with a copper coil submerged in it, to a 40 gallon water heater (with the jacket and insulation peeled off), all the way up to the insane and ridiculous idea of using an old 250 gallon oil tank.
My goal with in designing this heater would be to burn for a few hours in the evening and have enough heat stored to heat my 1800' house overnight. Once or twice a year it will get to -30F here, but usually winter nights are around 0F in February. So how much heat do I need? That is the question.
Running some quick heat loss calculations on the house, I come up with 14,000 BTU heat loss per hour. If I want to go for 10 hours overnight, I'll need to be able to store 140,000 BTUs.
Theoretically we will only be able to extract 80 BTUs per pound of water (water boils at 212 obviously, but the radiators require water in the 130-140 range. 212-130=82. I'm going to use 70 in my calcs, because water doesn't like to stay at 212.
140,000/70=2000 pounds of water. water is 8.3 pounds per gallon 2000/8.3 = 240 gallons.
So it looks like I really may need a 250 gallon tank. Not sure if that's practical, and how long it would take to actually heat it up. It may be necessary/simpler to use a 250 gallon tank remotely, and put a smaller tank in the RMH, and feed the bigger tank with a pump or thermosiphon.