VoltGround
GPU · VBIOS · Benchmarks · Thermals
← Back to Articles Guide

NVMe SSD Thermal Throttling: How to Detect It, What Triggers It, and Whether a Heatsink Helps

SSD throttling is invisible during normal use but measurable during sustained transfers. The controller temperature rises during sequential writes, hits a firmware-defined threshold, and the drive drops to a reduced performance mode to protect the NAND. This behavior is normal, but its frequency depends on airflow and heatsink presence.

How to Detect Thermal Throttling on an NVMe Drive

The clearest indicator of thermal throttling is a sustained-write performance curve that starts high and drops sharply after 10–60 seconds. Open-ended sequential write benchmarks using CrystalDiskMark’s 64 GB test or AS SSD’s 1 GB benchmark will not produce long-enough transfers to trigger throttling on most modern controllers. Use the following procedure instead:

  1. Download CrystalDiskInfo (free, portable) and note the drive’s current temperature at idle. This is the baseline.
  2. Open HWiNFO64 in sensors mode and pin the NVMe temperature sensor to a visible widget. Many drives report both composite temperature (the highest of all sensors) and the dedicated controller sensor separately.
  3. Start a large file copy—ideally a sequential write of at least 20 GB from RAM to the NVMe (use a RAM disk as source, or copy from a fast source drive). A 20 GB transfer at 5 GB/s should take around 4 seconds; at 10 GB/s, under 2 seconds. Use a 50 GB or larger payload to produce sustained thermal loading.
  4. Alternatively, run DiskSpd from an elevated command prompt: diskspd -d120 -b128K -w100 -o4 -Sh -c50G testfile.bin runs a 120-second sequential write workload using 128K blocks, which is sufficient to fully thermally load any current controller.
  5. Watch the HWiNFO64 temperature graph during the transfer. If the drive temperature rises to 70–80 C and you simultaneously see the write throughput drop in Task Manager or CrystalDiskMark, throttling has occurred. The performance drop is typically 40–70% of peak rated speed.
  6. Check the SMART attribute “Thermal Throttle Status” in CrystalDiskInfo after the test. Not all drives expose this attribute, but Samsung, WD, and Crucial drives commonly do. A non-zero value after the transfer confirms throttle events occurred.

Controller Temperature Thresholds by Drive Type

Firmware throttle thresholds are not always published, but consistent community testing and some manufacturer documents have established approximate ranges:

Drive / Controller Interface Throttle Onset Temp Emergency Shutdown Temp
Samsung 990 Pro (Blade) PCIe 4.0 70 °C 85 °C
WD Black SN850X (SanDisk) PCIe 4.0 72 °C 88 °C
Crucial T700 (Phison E26) PCIe 5.0 68 °C 85 °C
Seagate FireCuda 530 (Phison E18) PCIe 4.0 70 °C 85 °C
SK Hynix Platinum P41 (Hynix) PCIe 4.0 75 °C 90 °C
Generic QLC budget drives (SM2267) PCIe 3.0 65 °C 80 °C

PCIe 5.0 controllers dissipate significantly more power than Gen 4 equivalents, often running 6–10W at peak sequential load versus 3–5W for Gen 4 parts. This makes Gen 5 drives substantially more likely to throttle without a heatsink in systems with moderate airflow. The Crucial T700, for instance, ships without a heatsink in the standard package and routinely throttles within 20 seconds on sustained Gen 5 sequential writes in unventilated M.2 slots located under the GPU.

Heatsink Options and Their Effectiveness

Three categories of thermal solution address NVMe throttling with different tradeoffs:

Motherboard M.2 Heatsinks

Most mid-range and high-end motherboards include stamped aluminum heatsinks for M.2 slots. These provide passive conduction cooling through a thermal pad contacting the controller and NAND packages. Effectiveness varies: a well-designed motherboard heatsink with good pad contact and slot airflow can reduce steady-state controller temperature by 15–25 degrees Celsius, which is often enough to prevent throttling on Gen 4 drives. The primary limitation is that the heatsink has no airflow of its own—it relies on case airflow over the motherboard surface.

Third-Party Low-Profile Heatsinks

Products like the EKWB M.2 heatsink, Thermalright M.2 2280, and similar aluminum fin-stack coolers clip onto the drive and add fin surface area. When positioned in a slot with direct airflow from a case fan, these can hold Gen 5 controllers at 60–65 degrees Celsius even under sustained load. Fin density matters: higher fin counts help only if airflow reaches the fins; in stagnant areas of the case, a wider low-fin heatsink conducts better than a tall dense-fin unit.

Active Cooling (Fan-on-heatsink)

Small blower-style coolers designed for M.2 drives exist from brands like ICY DOCK and Sabrent. These are rarely necessary for Gen 4 drives but can be justified for Gen 5 drives in poorly ventilated cases, or in workstation builds that run extended sequential write operations (video editing, large file transcoding). Noise and the added complexity of fan power routing make them uncommon in typical gaming builds.

For drives installed in the M.2 slot under the GPU, no passive heatsink compensates for the GPU’s radiated heat. GPU exhaust temperatures in that zone can reach 50–55 degrees Celsius at ambient, establishing a high baseline before any drive load. Use the M.2 slot furthest from the GPU on boards that offer multiple slots, or route a dedicated case fan to that zone.