3 ms·

As a matter of fact it wouldn't overheat immediately but some 15-20 hours later. Very strange. As soon as it wasn't sealed, it would stop overheating and would

As a matter of fact it wouldn't overheat immediately but some 15-20 hours later. Very strange. As soon as it wasn't sealed, it would stop overheating and would transfer the heat steadily for weeks. Keep in mind that the whole tank was stainless steel with some 50 liters of oil inside, hence the oil-to-air transfer was more than adequate.

542354234235 2y ago
That makes sense, because it took 15-20 hours for all that oil to become fully heated and reach its carrying capacity. Your stainless steel tank and passive conduction was obviously not adequate, or you wouldn’t have overheating problems. When it was unsealed, the air convection across the top of the tank carried enough heat that it wouldn’t overheat. In another comment you say you use a passive cooling Noctua heatsink. If you seal up your PC with shrink wrap, your CPU will still overheat due to the lack of that same air convection. In a sealed system, you have to actively move liquid through radiators. You can move as much heat as you want into a heatsink, if there isn’t a way to move heat quickly out of the heatsink, then you aren't doing anything.
- 542354234235 2y ago
  Additionally: Sealing the tank would turn the airspace on top into an insulator, much the same concept as a double pane window. It would convert the entire top surface from a primary way heat is moved out of the system, through that convective circulation of room air, to a primary way to trap heat in the system.
  - jojobas 2y ago
    Losing just the top surface would leave the steel sides largely in place for oil-to-air heat transfer. I'd think, given the odour problem, evaporation actually played a role in his cooling.
  - atmanactive 2y ago
    Could be. But, as soon as I would drill a tiny hole, just to release the pressure, the temps would go back to nice and cool, even though the top gasket/lid/cables is still in place. So my bet stays with pressure build-up. It's not that I tried the whole thing once. I had several prototypes and spent at least two years trying various things before I finally gave up and switched to passive air cooling exclusively.

542354234235 2y ago

That makes sense, because it took 15-20 hours for all that oil to become fully heated and reach its carrying capacity. Your stainless steel tank and passive conduction was obviously not adequate, or you wouldn’t have overheating problems. When it was unsealed, the air convection across the top of the tank carried enough heat that it wouldn’t overheat. In another comment you say you use a passive cooling Noctua heatsink. If you seal up your PC with shrink wrap, your CPU will still overheat due to the lack of that same air convection. In a sealed system, you have to actively move liquid through radiators. You can move as much heat as you want into a heatsink, if there isn’t a way to move heat quickly out of the heatsink, then you aren't doing anything.

542354234235 2y ago
Additionally: Sealing the tank would turn the airspace on top into an insulator, much the same concept as a double pane window. It would convert the entire top surface from a primary way heat is moved out of the system, through that convective circulation of room air, to a primary way to trap heat in the system.
- jojobas 2y ago
  Losing just the top surface would leave the steel sides largely in place for oil-to-air heat transfer. I'd think, given the odour problem, evaporation actually played a role in his cooling.
- atmanactive 2y ago
  Could be. But, as soon as I would drill a tiny hole, just to release the pressure, the temps would go back to nice and cool, even though the top gasket/lid/cables is still in place. So my bet stays with pressure build-up. It's not that I tried the whole thing once. I had several prototypes and spent at least two years trying various things before I finally gave up and switched to passive air cooling exclusively.