High Speed Fan Testing on the Noctua L12 Ghost Edition
After earlier questions about possible improved cooling performance with the Phanteks T30 in advanced mode at maximum RPM, I decided to go back and retest my Noctua L12 Ghost Edition (after removing the bottom 92mm fan) with every 120mm fan I have available that exceeds the maximum 2000 RPM of the Noctua A12x25.
Fans used:
Noctua A12x25
Be Quiet! Silent Wings Pro 4 – in ultra high speed mode
Be Quiet! Silent Wings 4 High Speed
Phanteks T30 – in advanced mode
Scythe Wonder Snail
The Noctua A12 x25 was tested at 100% PWM with a maximum noted RPM of 2039.
The other four fan were tested both at 100% PWM as well as the PWM % that matched as closely as possible, the highest RPM of the installed Noctua fan.
The DB meter was 30cm from front center of the chassis facing directly towards the front of the chassis, and 19cm high from the desk top. The motherboard chipset fan did not activate during testing.
The Evga 3060Ti XC Graphics Card was disconnected but left in PCI Slot to restrict some of the airflow along with the upright mounted PSU on the opposite side of the motherboard.
The Ryzen 7 4750G’s integrated graphics were used to increase the wattage consumed and heat generated by the CPU.
The L12 Ghost Edition was mounted with the cooler fins running across the motherboard from the GPU to PSU side. This placed the cooler fins running top to bottom when used on a vertically placed motherboard, along with the cooler’s heat pipes running horizontally. Some cooling advantage might be achieved by rotating the cooler 90 degrees – while ensuring the heat pipe tips face upwards as per Noctua advice. This was a non-issue with this horizontal motherboard placement used in this testing.
Reported HWMonitor sensors were updated to reflect APU usage instead of graphics card.
Test system used:
MakerBeam open chassis test bench / proof of concept
Aorus X570 ITX MB; Ryzen 7 4750G using integrated graphics; 32 GB Corsair LPX 3200 RAM;
(Evga 3060Ti XC Graphics Card – Disconnected but left in PCI Slot)
Noctua L12 Ghost Edition
Noctua NT-H1 Thermal Paste
CPU-Z Stress Test & Unigine Valley 1080P Ultra Settings for 10 minutes
Room Temp 21C;
Fan noise characteristics noted:
The Noctua A12x25 was quiet & smooth sounding. Not surprising on a Noctua cooler.
The Be Quiet! Silent Wings Pro 4 – was loud at maximum RPM but much easier on the ears than the T30. It was the same noise level as the Noctua at 67% PWM.
The Be Quiet! Silent Wings 4 High Speed – has the same noise characteristics as the SW Pro 4. It did answer the question – are the larger radiator corners of the SW Pro 4 any benefit on an air cooler – no they are not – it was fractionally quieter at 79% PWM than either the Noctua A12x25 or the SW Pro 4.
The Phanteks T30 was quite loud with a harsh edge to the sound at maximum RPM. At a reduced 70% PWM it was still appreciably louder than the Noctua. At high speeds it can further reduce CPU and other component temperatures. From earlier testing (without a DB meter) I believe this is caused at least in part by the T30 being designed for radiator use with tighter fin spacing. I noted the T30 to be quieter at the same fan speed on the Alpenfoehn Blackridge cooler than the Noctua A12x25, while the reverse was true when used on the L12GE.
The Scythe Wonder Snail harmonizes much better on the L12GE cooler than it did on the Thermalright Silver Soul 135. There was only one resonance range noted at 75-76% PWM – 1901-1923 RPM.
At 83% PWM it was almost at the sound level of the A12x25.
CPUID HWMonitor & Tech Power Up GPU-Z Readings:
After earlier questions about possible improved cooling performance with the Phanteks T30 in advanced mode at maximum RPM, I decided to go back and retest my Noctua L12 Ghost Edition (after removing the bottom 92mm fan) with every 120mm fan I have available that exceeds the maximum 2000 RPM of the Noctua A12x25.
Fans used:
Noctua A12x25
Be Quiet! Silent Wings Pro 4 – in ultra high speed mode
Be Quiet! Silent Wings 4 High Speed
Phanteks T30 – in advanced mode
Scythe Wonder Snail
The Noctua A12 x25 was tested at 100% PWM with a maximum noted RPM of 2039.
The other four fan were tested both at 100% PWM as well as the PWM % that matched as closely as possible, the highest RPM of the installed Noctua fan.
The DB meter was 30cm from front center of the chassis facing directly towards the front of the chassis, and 19cm high from the desk top. The motherboard chipset fan did not activate during testing.
The Evga 3060Ti XC Graphics Card was disconnected but left in PCI Slot to restrict some of the airflow along with the upright mounted PSU on the opposite side of the motherboard.
The Ryzen 7 4750G’s integrated graphics were used to increase the wattage consumed and heat generated by the CPU.
The L12 Ghost Edition was mounted with the cooler fins running across the motherboard from the GPU to PSU side. This placed the cooler fins running top to bottom when used on a vertically placed motherboard, along with the cooler’s heat pipes running horizontally. Some cooling advantage might be achieved by rotating the cooler 90 degrees – while ensuring the heat pipe tips face upwards as per Noctua advice. This was a non-issue with this horizontal motherboard placement used in this testing.
Reported HWMonitor sensors were updated to reflect APU usage instead of graphics card.
Test system used:
MakerBeam open chassis test bench / proof of concept
Aorus X570 ITX MB; Ryzen 7 4750G using integrated graphics; 32 GB Corsair LPX 3200 RAM;
(Evga 3060Ti XC Graphics Card – Disconnected but left in PCI Slot)
Noctua L12 Ghost Edition
Noctua NT-H1 Thermal Paste
CPU-Z Stress Test & Unigine Valley 1080P Ultra Settings for 10 minutes
Room Temp 21C;
Fan noise characteristics noted:
The Noctua A12x25 was quiet & smooth sounding. Not surprising on a Noctua cooler.
The Be Quiet! Silent Wings Pro 4 – was loud at maximum RPM but much easier on the ears than the T30. It was the same noise level as the Noctua at 67% PWM.
The Be Quiet! Silent Wings 4 High Speed – has the same noise characteristics as the SW Pro 4. It did answer the question – are the larger radiator corners of the SW Pro 4 any benefit on an air cooler – no they are not – it was fractionally quieter at 79% PWM than either the Noctua A12x25 or the SW Pro 4.
The Phanteks T30 was quite loud with a harsh edge to the sound at maximum RPM. At a reduced 70% PWM it was still appreciably louder than the Noctua. At high speeds it can further reduce CPU and other component temperatures. From earlier testing (without a DB meter) I believe this is caused at least in part by the T30 being designed for radiator use with tighter fin spacing. I noted the T30 to be quieter at the same fan speed on the Alpenfoehn Blackridge cooler than the Noctua A12x25, while the reverse was true when used on the L12GE.
The Scythe Wonder Snail harmonizes much better on the L12GE cooler than it did on the Thermalright Silver Soul 135. There was only one resonance range noted at 75-76% PWM – 1901-1923 RPM.
At 83% PWM it was almost at the sound level of the A12x25.
CPUID HWMonitor & Tech Power Up GPU-Z Readings:
CPU Cooler Fan | Noctua A12x25 | Phanteks T30 | Phanteks T30 70% | BQ! SW Pro 4 | BQ! SW Pro 4 67% | BQ SW 4 HS | BQ SW 4 HS 79% | Scythe WS | Scythe WS 83% |
TZ 10 Temp Sensor Max | 16.8 | 16.8 | 16.8 | 16.8 | 16.8 | 16.8 | 16.8 | 16.8 | 16.8 |
System 1 Max | 44 | 39 | 41 | 41 | 44 | 41 | 44 | 41 | 42 |
CPU Max | 74 | 70 | 72 | 72 | 75 | 72 | 74 | 72 | 74 |
PCIEX16 Max | 21 | 21 | 21 | 21 | 21 | 21 | 21 | 21 | 21 |
VRM Max | 46 | 40 | 42 | 42 | 46 | 43 | 47 | 44 | 46 |
CPU Fan Max RPM | 2039 | 2960 | 70% 2039 | 2766 | 67% 2039 | 2472 | 79% 2051 | 2327 | 83% 2057 |
Chipset Fan Max RPM | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
CPU Pkg Max Temp | 74 | 70.6 | 72 | 72.4 | 75 | 72 | 74.3 | 72.1 | 74.3 |
CPU Pkg Max Wattage | 86.8 | 87.38 | 87.43 | 87.24 | 86.93 | 86.75 | 86.77 | 86.9 | 88.16 |
NVMe Top 980 Pro Temp | 45 | 41 | 43 | 42 | 46 | 42 | 45 | 43 | 45 |
APU Max Temp | 55 | 51 | 54 | 54 | 55 | 53 | 55 | 53 | 55 |
SoC Max Temp | 58 | 54 | 55 | 56 | 59 | 55 | 58 | 55 | 57 |
DBA Sound Level | 43.4 | 57.7 | 46.5 | 51.7 | 42.9 | 48.5 | 42.8 | 47.3 | 43.5 |
