I've delved a bit further and the answer, it seems is to collect all the gas and condense it. This could still come from the top, but feed directly into a condenser pipe, maybe set in a 20 litre tin with cool water in/hot water out pipes. I did think of using a hot water coil, but think it would be better to have it zig zagging in a single downhill plane through the 20 L tin, and and no point where liquid could collect. I suppose you could still have the second retort pipe running back underneath with an in line pressure gauge and a gate valve to burn off to control the heat levels better/reduce excessive pressure in the barrel. I wonder what the failure/bursting point would be in psi of a 200 litre oil drum would be? I wouldn't have thought it much greater than 50 psi? A second vent could be good as well as an external burner if the temperature needed reducing along with the pressure. This European Commission PDF for Norway has some more good information suggesting that the optimal temperature from pyrolysis by destructive distillation in a retort would be about 400 degrees C. An industrial thermometer may be useful too.
My thought on the no return valve is to KIS. A hinged flap that hangs vertically over the outlet in a shut position until liquid or gas pushes it open would suffice, as any drawing for air back up through the outlet would merely shut it tight. Once again this is where a pressure gauge would be good, although I strongly doubt that there would ever be a catastrophic implosion by lowering of pressure in the kiln. I think the no-return would only be useful for the beginning of the firing as there would soon be a build up of pressure in the kiln.