Yeah was looking at that and the Peerless Spirit but concerned about RAM clearance as I'll be using all 4 slots....
Try just using the middle fan. I built 5800X3D rig for someone else and used the spirit with only the middle fan for asthetic reasons. (Because the other fan only clears the RAM when it's mounted high and sticks up over the heat sink.)
Compared to the D15S setup on my primary rig, the single fan Spirit games the same, and all-core stress testing (R23) reaches my 87C temp limit sooner. (Neither cooler can stop the X3D part from hitting the temp limit indefinitely)
It used to bother me that the X3D parts were capable of hitting the temp limit even with strong cooling, but once AMD allowed setting a custom temp limit in later AGESA versions, I just set it to 87C and let the chip downclock in the off chance that I do something that demanding. (Just gaming has never got the chip that hot.)
I realize "just let it downclock" is probably blasphemy on an OC forum, but the 5800X3D's gaming performance comes from its massive cache rather than clock frequency. Also, gaming is it's primary job and temps are fine there anyway. When I occasionally use it for rendering video, I might get the video rendered a *little* quicker if I had a 360mm AIO, but not enough for me to care.
*edit* just retested and the single-fan spirit stayed just under 85C for a full benchmark run. I think I either looped for a longer period of timeand / or had a higher room temp last time. (23C this time)
I also decided to retest my D15S rig, and it actually runs *hotter* than the single-fan Spirit setup. (86.4C)
Only thing I can think of is the two rigs have different SOC voltages. The Spirit rig is running 3200MT RAM and an SOC of 1v. My primary rig is running 3600MT with an SOC of 1.1v.
That and just general silicon quality. They both score within 100 points of each other.