I've not mentioned the hardware yet, let's start with the list:
Case: Lazer3D LZ7 First Edition, with "open" GPU panel option
Motherboard: ASRock Z87E-ITX with the latest firmware
CPU: Intel Pentium G3258, default: 3.2GHz, clocked at 4.5GHz with 1.00V to 1.20V adaptive voltage
CPU cooler: Noctua NH-L9i with custom fan profile and Arctic MX-4 thermal paste applied
RAM: 2x 4GB Crucial Ballistix VLP DDR3-1600 at default clocks
Storage: Samsung SSD 830 256GB with an updated Windows 7 install
GPU: MSI Radeon RX 460 4G OC with fan + shroud removed and Arctic MX-4 thermal paste applied
PSU: Silverstone ST45SF-G (450W SFX 80+Gold) with a NoiseBlocker "fan-mod" and custom cables
Cooling: 2x Prolimatech Ultra Sleek Vortex 14 with custom fan profiles, one as chassis fan, one attached to GPU (w/ adapter)
Most of the components I already had on-hand for this build, except for the case fans and GPU. I choose the fans because it was recommended by K888D as one of the better performing ones and it gets universal praise from SFF afficionados alike.
The GPU was choosen based on the usage: for the LAN parties I visit, there is rarely more power needed than what a random mainstream card can deliver. This card was cheap, has a flat heatsink and doesn't require extra PCIe power connectors: check, check and check !
I wanted to quantify the noise measurements and have atleast a somewhat scientific aspect to it:
That is a Samson Meteor, a mic suspension (with a custom solution for the Meteor) and a boom arm. This decouples the microphone from surface vibrations and allows me to position it more neutrally.
The microphone signal feeds into Audacity over USB at 100% sensitivity, I take note of the value at idle, which is all the devices, except my own PC running Audacity, turned off. My own PC at about 1 meter distance from the microphone has all Noctua NF-F12 fans at silent settings (300-500rpm) at which point I can't make out the noise over my own breathing. This is not a lab test, but it will do until I find a better way.
Audacity showing the microphone input with noise in the background. Note the negative value to indicate the range.
Since decibels are relative in the sense that there needs to be a point of reference, I think it's best to just look at the idle value, compare it to the load values and take my subjective opinion into the equation. Decibel is also a logarithmic measurement, so from -40 to -30 isn't 33% louder but more in the range of 10 times as "powerful". Many have often frustrated about the difficulty to put noise into the correct perspective, but this is an extremely personal aspect. Some people have better or worse hearing, some people live in very noise poluted or "clean" areas, some play with closed headphones and don't care. So be sure to not read too much into sound measurements.
I also make use of Prime95 to test maximum synthetic CPU load, Heaven 4.0 Benchmark for synthetic GPU load and HWinfo to measure the temperature and fan speeds.
Idle readings showing 32°C on the CPU, 1270rpm on the CPU fan (600-2500rpm range) and from looking it up in the BIOS, the chassis fan is turning at arounf 500rpm. My microphone registers around -42dB and to me this is silent. I could put my head close to the case and make out sounds, but that is not a real use case for most.
Prime95 has been running for 6 minutes and the temperature has stabilized. I did test this a few more times for about an hour in the past month and the results were generally the same. At this load the CPU are at 75°C, the NH-L9i's fan spinning at 2070rpm and I calculate the case fan at about 850rpm. My microphone measures around -39dB and to me this is clearly audible. But low volume music or the soothing voices of LowVolu.me podcast can quickly drown that out. I'm amazed how fast the little fan on the Noctua heatsink can spin without turning into a wailing banshee.
Heaven Benchmark after about 30 minutes, the temperature of the GPU stabilizes around 75°C and as you can see in the graph it consumes about 50W. The simple all-aluminium heatsink on this card isn't very hefty and the temperatures sure can show it, with the second Prolimatech fan at around 1050rpm pushing buckets of air at it. The Meteor mic outputs about -36dB of noise at this point, I'd say loud enough to ruin a quiet office day, but not loud enough to bother someone playing an action-intensive game with the sound at reasonable levels. I'll be amazed if it would bother me at a LAN party.
Case: Lazer3D LZ7 First Edition, with "open" GPU panel option
Motherboard: ASRock Z87E-ITX with the latest firmware
CPU: Intel Pentium G3258, default: 3.2GHz, clocked at 4.5GHz with 1.00V to 1.20V adaptive voltage
CPU cooler: Noctua NH-L9i with custom fan profile and Arctic MX-4 thermal paste applied
RAM: 2x 4GB Crucial Ballistix VLP DDR3-1600 at default clocks
Storage: Samsung SSD 830 256GB with an updated Windows 7 install
GPU: MSI Radeon RX 460 4G OC with fan + shroud removed and Arctic MX-4 thermal paste applied
PSU: Silverstone ST45SF-G (450W SFX 80+Gold) with a NoiseBlocker "fan-mod" and custom cables
Cooling: 2x Prolimatech Ultra Sleek Vortex 14 with custom fan profiles, one as chassis fan, one attached to GPU (w/ adapter)
Most of the components I already had on-hand for this build, except for the case fans and GPU. I choose the fans because it was recommended by K888D as one of the better performing ones and it gets universal praise from SFF afficionados alike.
The GPU was choosen based on the usage: for the LAN parties I visit, there is rarely more power needed than what a random mainstream card can deliver. This card was cheap, has a flat heatsink and doesn't require extra PCIe power connectors: check, check and check !
I wanted to quantify the noise measurements and have atleast a somewhat scientific aspect to it:
That is a Samson Meteor, a mic suspension (with a custom solution for the Meteor) and a boom arm. This decouples the microphone from surface vibrations and allows me to position it more neutrally.
The microphone signal feeds into Audacity over USB at 100% sensitivity, I take note of the value at idle, which is all the devices, except my own PC running Audacity, turned off. My own PC at about 1 meter distance from the microphone has all Noctua NF-F12 fans at silent settings (300-500rpm) at which point I can't make out the noise over my own breathing. This is not a lab test, but it will do until I find a better way.
Audacity showing the microphone input with noise in the background. Note the negative value to indicate the range.
Since decibels are relative in the sense that there needs to be a point of reference, I think it's best to just look at the idle value, compare it to the load values and take my subjective opinion into the equation. Decibel is also a logarithmic measurement, so from -40 to -30 isn't 33% louder but more in the range of 10 times as "powerful". Many have often frustrated about the difficulty to put noise into the correct perspective, but this is an extremely personal aspect. Some people have better or worse hearing, some people live in very noise poluted or "clean" areas, some play with closed headphones and don't care. So be sure to not read too much into sound measurements.
I also make use of Prime95 to test maximum synthetic CPU load, Heaven 4.0 Benchmark for synthetic GPU load and HWinfo to measure the temperature and fan speeds.
Idle readings showing 32°C on the CPU, 1270rpm on the CPU fan (600-2500rpm range) and from looking it up in the BIOS, the chassis fan is turning at arounf 500rpm. My microphone registers around -42dB and to me this is silent. I could put my head close to the case and make out sounds, but that is not a real use case for most.
Prime95 has been running for 6 minutes and the temperature has stabilized. I did test this a few more times for about an hour in the past month and the results were generally the same. At this load the CPU are at 75°C, the NH-L9i's fan spinning at 2070rpm and I calculate the case fan at about 850rpm. My microphone measures around -39dB and to me this is clearly audible. But low volume music or the soothing voices of LowVolu.me podcast can quickly drown that out. I'm amazed how fast the little fan on the Noctua heatsink can spin without turning into a wailing banshee.
Heaven Benchmark after about 30 minutes, the temperature of the GPU stabilizes around 75°C and as you can see in the graph it consumes about 50W. The simple all-aluminium heatsink on this card isn't very hefty and the temperatures sure can show it, with the second Prolimatech fan at around 1050rpm pushing buckets of air at it. The Meteor mic outputs about -36dB of noise at this point, I'd say loud enough to ruin a quiet office day, but not loud enough to bother someone playing an action-intensive game with the sound at reasonable levels. I'll be amazed if it would bother me at a LAN party.
Last edited:
