APOLLO II in field temperature performance.
The case design is tight. The metal case is critical in thermal dissipation. Airflow is also critical. When the fan increases speed it means the temp sensors need more airflow to compensate higher temperatures. IF fan speed is lower there is sufficient airflow to cool the cpu and miner.
With the miner in normal position, the design as you say is to increase flow through the top vents. THis is correct that the airflow through the top vents does help cooling of the power supply. But there are only 3 fine vents on the top. It looks like the design was for most airflow through the bottom. The problem with this design is air is pulled through the top and bottom and then out the side. The side exhaust creates a feedback airflow loop back to the bottom of the unit. Thus, a percentage of hot exhaust air gets pulled back into the bottom increasing temps entering unit. Next, in the real world where there is dust, if you watch with the correct light you will see dust pulled into the top vents and dust landing on the surface also gets pulled into the bottom at a higher rate than fans than intakes located on the side. The top vents accumulate dust much quicker causing the same airflow problem that increases heat into the power supply. Most computer designs I've seen and used have the intake and exhausts on the side. I would think a future design with the fan vent pointed up and intakes on the side and bottom would be better as thermal heat wants to rise.
Now, for this SITU: The power supply seems to have a heat sink connected to the case looking through the vents. WIth an infrared temp sensor I measured the metal sides and top.
In normal position the APOLLO II reported temps 2-3deg. C. higher and fan speed 200-300 RPM higher than turning my unit on its side. This tells me the unit is not getting enough airflow based on measured temperatures. The exterior side where the power supply measured 37-38 Deg. C. When turning the unit with the HDMI pointed up. miner temp dropped and fan speed dropped. The temp on exterior case was 39-40 Deg C. This all while running in turbo mode. This should not be hot enough to create problems. I would think running the miner cooler is more important to extend life of miner. Case temperaure next to power supply was similar to apollo reported system temperaure. System temperature was actually warmer than case temp and did not change much based on position. Maybe there needs to be a temp sensor in the power supply for future design. With intake on the sides there is no exhaust feedback to the bottom part of the case that is now on the side. The volume of air pushed out the exhaust actually creates airflow over the sides helping fresh airflow into the vents and natural cooling of the metal case
Regarding the Switch. Either its a design flaw in excess current through the small switch or its a faulty switch. Temp measured at switch is same at case temp 38-40 Deg. C. It's also next to fan. I've had no problems with my switch from a unit shipped in october. 5A through a 5A switch will fail more often than 4A through a 5A switch and 3A through a 5A switch with fail much less if at all.
Running the miner in normal mode or eco mode should automatically pull much less power reducing stress on the power supply, no matter what position the miner is placed.
In eco Mode the miner runs very well in the normal position. is very quiet and uses little power.
Overall this is a very nice design and I'm sure was not easy to fit everything into this small package and slick look. The exhaust heat sink and heat pipes are very efficient to pull heat away from the cpu and miner. The other reason I like the unit sideways is so I can look inside the bottom and see all the LED lights hidden from view which tell me how my miner is operating. The LED light on my NVME lets me know when data to memory is active or inactive.