In my 100-300k OP, I ended with this...Tom Booth wrote: ↑Sat May 18, 2024 2:41 am According to Tesla's theory...
If we first ran the Peltier module to cool the bottom plate. (Keeping it insulated of course, to preserve the cold).
The engine could then run on the surrounding ambient heat, and with perfect insulation and a 100% efficient engine, it could run indefinitely.
He recognized however, that with the technology current in his day, those two conditions could not be met, so the "cold hole" would need to be re-cooled or refrigerated from time to time.
But, be reasoned. Because the engine is actively converting the ambient heat entering into the engine to other forms of energy, less heat would reach the sink than goes in at the top.
He further reasoned, that when heat did need to be removed from the "cold hole", it could be taken out "by its own energy".
"So, for expansion at 300k, we need 3Qin for 3Wout while for compression at 100k we need 1Win for 1Qout. In this manner, eff=.66 where (300-100)/300 is inline Carnot. Unfortunately, the compression process taxes the system twice: (1) via mechanical backwork=1Win (2) via heat sink=1Qout, thus doubling losses whereby ideal eff=.33 and Carnot laughs."
Here's a graphic described as an Otto cycle, but could also represent a Stirling cycle. If Stirling, process 3-4 will be isothermal expansion with ambient input, but then process 1-2 would be isothermal compression with heat sink filling cold hole. It really doesn't matter if ambient pressure supplies compression force, since as Fool has pointed out several times, the 'spread' between 2 isothermals or adiabats requires a heat transfer. The Stirling does this via regen, but requires heat sink during compression vs Otto does this via source and sink, but no heat transfer during compression. Without this temperature spread, there's no Wnet area.
Tom - even with your fantasy backstroke, your suck cycle would require an input temperature spread, so if pt 3 was T ambient then pt 2 would be T sub ambient, and this raises the backwork ratio.
This is all fantasy. In my OP, I mentioned cold hole issues as "(1) via mechanical backwork=1Win (2) via heat sink=1Qout" but ignored that compensating for "sink=1Qout" would only tax the system .33 when maintaining cold hole was achieved at 1.0 eff. What your suggesting is transforming any sink heat above 100k into greater heat (300k ?) AND at 1.0 eff !!!Tom Booth wrote: ↑Sat May 18, 2024 2:41 am To explain, let's say, we have a heat pipe that takes the heat generated by the Peltier unit while operating and carries that heat back up to the ambient side on top.
In this way, the heat taken out of the "cold hole" will assist in its own removal.
While heat is being removed, the temperature difference will increase and the engine will have more power to make the cold hole colder using the heat taken out of the cold hole.
After that, the engine can run steadily on the ambient heat once again.
He believed this scheme did not require a 100% efficient engine.