I am by no means an expert in air circulation. To me, the main (APU) fan has the purpose to push air onto the radiator fins. The front fan pulls extra fresh air into the case, while the back fan pulls hot air out of the case. Of course the air flow is altered inside the case, but that does not seem to matter much from a practical POV: I get 7-8ºC lower temperatures than with just the original Noctua.
I think both the original Noctua and your aftermarket fan are “static pressure” fans, so it’s not a problem, I was replying the above comment by convexproblem.
“Airflow” fans typically have thin blades with high angle of attack, creating faster airflow at a given RPM, but the blades will stall at lower speeds, reducing efficiency. “Static pressure” fans have thick, “curvy” fan blades where the angle is shallow at the beginning and steep at the end, to force air from a very low velocity to areas with air resistance such as radiator fins. According to the photo you posted at OP, your fan is already “static pressure” so there’s no problem.
The shape of the heatsink is very mismatched to a 140mm fan… That’s what I meant.
How’s that? Can you explain? And would it be a perfect match for a 120?
I got similar delta temps when I was using the 140 mm fan as I get with the current 120 Phanteks. If you were right, that would not have happened.
I swapped things again to prepare for a bigger case for my nvidia V100 addition, took out the 25mm tall shroud and put in my 12mm 3d printed stator vanes and bought a new 120mm fan. This one has a big 50mm center hub diameter so I’ve probably lost most of the airflow through the center of the heatsink. The heatsink temp sensor with this fan on full blast reads noticeably higher than any of my 92mm fan configurations, but the noise is down considerably and the CPU temp is still well below 80C so it’s fine.
This setup would probably beat out the 92mm fans if I put in another spacer but it’s hard to justify the effort.
I recommend getting a Coolermaster MasterFan A120, its the brand new 120x30mm all metal fan. The static pressure from it is so insane that a front fan restricts the air flow. Before I started dialing up the TDP, with the stock fan curve and TDP, my 395 64GB variant was sub 60C. I’ve since dialed the fan speed up and TDP so that I can run 5.1GHz CPU and 2.9GHz GPU sustained at ~135W at around 65-72C average.
Pressure is mainly about blade speed. The motor hub has a dead spot behind it where air doesn’t flow, that’s the issue here. The 92mm fan idea was to allow a lower total height because the smaller hub shadow needs less shroud height to fill. The cooling performance was good and the “whoosh” noise was low but the higher frequency sounds were slightly annoying.
This bigger fan pushes air well since the root of the blades is further out and travelling fast, also reducing the total area the air sees to reduce backflow, it just needs more shroud height to help the air diffuse into the heatsink evenly. The stator vanes recover the swirling momentum that normally is done by the heatsink fins when the fan is mounted closer.
This is way too simple of a statement.
With the A120 fan, I was planning on doing testing between the shroud mounted in front of the fan and behind the fan, since to get it within clearance, I took Noctua’s shroud and reduced the ‘cone’ by 5mm (since its 30mm thick).
Either way, if you haven’t tested the MasterFan A120 in this setup, I highly recommend doing so, they’re readily available and like $40. I have several 92mm and 120mm fans, including one of the old Fractal Design 140mm fans with 120mm spacing, and its really hard to describe how much airflow comes out of the front and rear with that fan. The performance difference was about 10C less than the ‘standard’ Noctua 120x25mm fan that you can buy with the rig.
the CM A120 specs 77cfm / 60Pa. what is interesting is that according to the site it is rated at only 2watt (which looks like a whooping 3.6W/AW efficiency ! which is a really good number).
as a side note, if anybody tried to use claude/chatgpt to help you find the highest efficiency fans, you could see numbers down to like 2AW/W, but nobody i think will like the fans, they run at high rpm
(AW = airwatts power)
In practice, the fan dramatically out performs every other fan I’ve tried in this system. Its designed to provide static pressure for radiators, and it definitely does the job here.
I experimented a bit with it too recently, mounting my shroud behind the fan instead of infront of it, and seemingly based on the fin design of the cooler, it performed dramatically better direct mounted to the heatsink rather than spaced off by a few millimeters. I’d highly recommend anyone with a Framework Desktop pick one up and give it a try, its a surprisingly inexpensive part considering the construction.