Arctic Freezer 50 TR Review: Can It Cool Threadripper and Ice the Competition?


As something new to our reviews, I recorded an installation video for the ARCTIC Freezer 50 TR. Being a rough first video, the audio quality is one thing we need to improve, but it gives a good sense of how the Freezer 50 TR is installed. The only thing we really had trouble with was locking the clips for the center fan back in place after the heatsink was mounted.

The center fan’s shroud locks in place using a clip mechanism

Our chosen TR thermal pattern manages to get a nice, complete spread of thermal paste across the whole base of the heatsink. The heatsink itself makes good contact across the critical areas of the CPU’s heat spreader with extra margin, however it doesn’t reach to the edge of the spreader to maximize contact surface area like the Noctua’s with their coldplate, so it may have one step forward and a minor step backward for ARCTIC’s direct heap pipe method. The coverage is large enough already that the difference due to surface area should be minimal.

Test Setup

AMD Threadripper 3960X Test Bench
CPUAMD Threadripper 3960X
CPU Cooler<Item Being Tested>
Thermal PasteARCTIC MX-4 (included)
GPUGigabyte AORUS Radeon RX 570 4GB
MotherboardGIGABYTE TRX40 AORUS Xtreme
MemoryCorsair Vengeance LPX 3200MHz CL16 4x8GB (32GB total)
StorageCorsair Force MP600 1TB (PCIe 4.0) NVMe
CaseRosewill RISE Glow ATX Full Tower
PSUFSP Group Aurum PT 1200W

A few things of note in our test bench is we’re now using the Rosewill RISE full tower rather than the Lian Li O11 Dynamic. You may have noticed the bright bank of three blue fans in the front of the case as well as a black vent fan from the installation video. To maintain parity with our previous tests in the fanless Lian Li case, the Rosewill case fans were all unplugged during testing, as well as leaving the side of the case open, resulting in the same pseudo open-air test setup.

We also managed to lower our noise floor to 38.6 dBA, as measured by our BAFX3370 that we use to capture all our noise level readings. This was due to moving testing to an isolated room with soft furnishings and large, soft sound absorbers on large sections the walls (drapes and decorative blankets basically). This gave us a good balance between thermal regulation, for maintaining a constant 21°C room temperature, and a low noise floor.

In This Review

Liked it? Take a second to support Kirk Johnson on Patreon!