somehow, "normal idle" CPU temp keeps rising.
That change will be correlated with something, but you don't give a lot of detail to make solid guesses with. I hope you will be content with vague guesses.
Today's operating temp is 62°C
Key fact: semiconductor operating temperature is always relative to ambient temperature. The first prerequisite to a steady operating temperature, therefore, is a steady ambient temperature. Is that the case?
If everything around the router is getting hotter (e.g. higher load on nearby equipment, higher air conditioner set-point, global warming, etc.) the router will naturally get hotter even if nothing else changes within it.
Instead of saying 62°C, it would be more useful in a situation like this to say instead something like "38°C over ambient." That requires making two temperature measurements, the other being out in the room, well away from the major heat generators. Doing this allows you to make comparisons against this variable over time. If you find that your historical 22°C rise correlates with a 20°C rise in ambient temperature, you know 2°C at most are inherent to the router itself.
after I re-attached the stock black heatsink with TWO 80mm Noctua fans blowing directly onto heatsink
Did you refresh the heat sink compound, or did you reuse what was there before?
If you refreshed it — as you should have done — then did you clean the old stuff off first? If so, how? How much fresh compound did you use?
currently, chassis is open
That isn't necessarily best for cooling. Fans work by developing
𝚫P to move air. If you dissipate their force in all directions outward from the PCB, they will necessarily move less air over some areas of the PCB. There's only so much pressure to go around.
I've used equipment that ran perfectly fine when the covers were closed, but would shut down due to overheating if you opened it up while it was running.
It's only possible for what you've done to help if the router's designers did a bad job of modeling the internal airflow, so that opening the whole thing up to ambient air gives a better overall result. This is by no means certain. A lot of hardware design shops buy and use expensive thermal design software for this purpose. You might've wrecked months of careful airflow design with that move.
Since you speak of opening the router up, can I assume you've thoroughly dusted the router's innards? If not, that's an insulating blanket that will accumulate, contributing to rising operating temperatures.
Stock fans (2x 40mm 10k rpm) were just TOOO noisy, even at lower rpm (around 4-5k rpm).
2×80 may not be equivalent in 𝚫P if the new ones run slow enough. That doesn't explain the steadily rising temperature you report, but it would explain an initial step change. What did you observe?
Today, this is NOT possible.
Aside from the above guesses, what else has changed between then and now?
Have you upgraded RouterOS over that span? Different software can change operating temperature by putting more background processes into play to deliver new functionality. New software can also take more CPU power to deliver the same functionality in a better manner; ROS 7 is known to do that, though you'd see a one-time step change here, not the steady rise you report.
You've spoken of "idle" conditions, but is the router truly idle? If you have a background data flow through the router, and that background load has changed over time, it could explain this. For example, you might have security cameras constantly streaming to an NVR somewhere, and you've learned to disregard them as part of the system's "noise floor," but you've neglected to consider that all of the old 720p cameras have been replaced first with 1080p cams, then 4k, doubling bandwidth requirements each time.
Ya canna move packets without moving electrons, and ya canna move electrons without generating heat.