Question Quiz - - - what is your fan RPM ?

[donline]

Hi everyone,

I was wondering if you could share what your CPU fan RPM and temperature is under moderate usage (e.g. full HD/1080p video running on YT)?

My fan is usually at 1400 RPM (and CPUs around 60C temperature) when idle/minimal usage, and then kicks up to 1800-2000 RPM (and around 60-65C) when playing a full HD video on YT. I tested this with Xsensors (in the Linux Software Manager).

Is it normal for the fan to run at the RPM above? And what would you consider acceptable and noisy in terms of RPMs?

Thanks very much

 

[Lutfij]

You nee to consider some factors. You would need to include your processor's make and model as well the cooler you're using as well as the case and the number and orientation of fans in said case, not to mention the ambient room air temps.

I could say my fan Noctua NH-L9i is set to 85% fan speed which is roughly 2000 rpm's but that only is part of the picture.

Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:

 

[SyCoREAPER]

I have everything maxed out. It's a fairly open case so loudness isn't awful. I'd rather have everything run cool and potentially have to replace fans a bit earlier than have a toaster.

 

[Cellie]

That depends so much on your ambient temperature, case, CPU, cooler type and last but not least important: fan settings -- you can't really compare with data from random people. Mine for example:

~1700RPM (75%), CPU 32º, ambient temperature 26ºC. Because it's a WC, the only fan in the case and I've set an aggressive curve to keep the mobo under 60ºC since my computer got gremilins.

 

[zinkles]

I have everything maxed out. It's a fairly open case so loudness isn't awful. I'd rather have everything run cool and potentially have to replace fans a bit earlier than have a toaster.

FYI: there wouldn't be any noticeable difference in lifespan when using the fans at max speed and when using them at a slower speed/auto. Wear and tear is a thing, but it will affect both and the difference between it is very negligible.

 

[CountMike]

Hi everyone,

I was wondering if you could share what your CPU fan RPM and temperature is under moderate usage (e.g. full HD/1080p video running on YT)?

My fan is usually at 1400 RPM (and CPUs around 60C temperature) when idle/minimal usage, and then kicks up to 1800-2000 RPM (and around 60-65C) when playing a full HD video on YT. I tested this with Xsensors (in the Linux Software Manager).

Is it normal for the fan to run at the RPM above? And what would you consider acceptable and noisy in terms of RPMs?

Thanks very much

That's not an answerable question because of so many variables involved.

 

[donline]

That's not an answerable question because of so many variables involved.

Thank you CountMike.
What would you suggest, how can I limit the variables?
Ideally I would like folks to test running a full HD (1080p) video on YT (without other apps open) and see what the CPU RPM and temp is.

 

[CountMike]

Thank you CountMike.
What would you suggest, how can I limit the variables?
Ideally I would like folks to test running a full HD (1080p) video on YT (without other apps open) and see what the CPU RPM and temp is.

Start with this: What is the function of fans ? The answer: To move air over/thru coolers to help heat dissipate. So heat/temperature is indication how good job they do. Follow temperatures and you will find answer for your setup.
My way is to have enough of adequate fans to keep temps tolerable. Fan speed is of course adjustable but there's no use of fussing too much about speed curves. I just set minimal rpm for temperature at idle and than straight line to top rpm at just before maximum allowed/tolerable temperatures. If it's not enough ad more and if too loud, find quieter ones.

 

[SyCoREAPER]

FYI: there wouldn't be any noticeable difference in lifespan when using the fans at max speed and when using them at a slower speed/auto. Wear and tear is a thing, but it will affect both and the difference between it is very negligible.

Generally I agree with your statement but bearing design plays a role as well as the fan curve.

I've seen people unknowingly set it to, or auto going too low, where the lowest point doesn't supply enough voltage to even spin it. Just like good old HDD, a full stop and go is hardest on them.

I know that's the opposite but figured I'd add to what you said.

 

[Paperdoc]

I like CountMike's approach. It aims for low fan speed at minimal workload and max fan speed just below the max CPU internal temperature. In fact, most default "fan curves" in BIOS Setup will do something very close to that. However, IF you want to keep your CPU temperature LOWER that that at all times, OP, you CAN set up your own custom "fan curve" and force the fans to run faster for all temperatures. This will result is slightly faster wear of the fan bearings and thus slightly shorter fan lifetime, and it MAY also result if longer lifetime for the CPU. But be aware that MOST CPU's will actually last longer than their "performance lifetime" - that is, at some time you will WANT to upgrade your hardware for several reasons and NOT because the CPU has started to fail. So the gain in CPU lifetime may never be used!

OP, you also should be aware that the conditions you cite for testing do NOT represent a max workload for your CPU. A lot of the work in that scenario is being done by the graphics card. You need to look up stress tests for CPU's particularly to force it into max workload conditions for testing.

There are so MANY variables in this field, but fan SPEED so not a good one to use. I suggest instead looking at the fan speed SETTING as a percent of max fan speed. Different fans will run at different speeds for that SAME setting. The actual output of a fan control system (i.e., from the CPU_FAN header in this case) actually IS a "% of Full Speed" kind of signal, and many displays of "fan speed" will actually report in this form.

 

[CountMike]

Fan's performance in terms of pressure and airflow is not linear with rpm, usually peaks at about 70-80% of full rpm and that is not even taking resistance of cooler itself if it's close enough. When you intersect that curve with curve set in BIOS it results in a straight line anyway, so why complicate settings, leads to nowhere.

 

[35below0]

Too many variables but this is my situation for comparison:

I cannot see my CPU fan(s) RPM as they are connected to a low-noise adapter. I don't remember what the RPM was before i switched to using the adapter.
The cooler is Noctua DH-D15, 2 140mm fans + low noise adapter. CPU (13600K) temp is hovering around 36-37C.

System fans are split into:
2 Fractal Design Aspect 140mm front intake fans, spinning at 700-720 RPM. Temp is 27-33C front side.
1 Noctua NF-P14s redux-1200 PWM, rear exhaust, spinning at 460 RPM. Temp is 29-37C rear side.

When gaming:

CPU temp went up to 70C.

front intake fans, spin at 755 RPM. Temp 28-38C front side.
rear exhaust spins at 510 RPM. Temp is 32-47C rear side.

GPU fans spun up to 2100 rpm - temp should be around 65-70C.
(Gigabyte Eagle OC 4060)

I don't know how meaningful this info is... I do use a custom fan control curve that is a little more silent than Gigabyte's own silent curve. Long story short, fans spin at slightly lower RPMs until temps rise, then RPMs rise more aggressively. I usually don't have temp issues so i prefer the fans to be more quiet. Gigabyte "normal" curve makes fans noticeable without doing much to improve on the already low temps.

https://imgur.com/HDmR346

View: https://i.imgur.com/HDmR346.png

 

[Paperdoc]

Just my comment on the post above. I do NOT recommend using the Low Noise Adapter that comes with Noctua fans if you have that fan connected to a mobo header that IS doing automatic fan control. Low Noise happens ONLY because the adapter is a resistor that reduces the Voltage supplied to the fan, and hence its speed and air flow. This is VERY useful only IF that fan is connected to a fixed 12 VDC power supply (for example, directly from a PSU power output) so the only way to reduce the fan speed is to use that adapter. However, when it is controlled by a mobo header with that adapter in the line, what that header will do is look at the CPU internal temperature sensor (if this is the CPU_FAN header) and discover that the temp is hotter that it should be (because the fan is running too slow and delivering less air flow), and increase the fan speed. It will keep doing that until the temp comes close to the right temp. BUT if you remove that adapter, the system will see that the temp IS maintained at the correct value by using a LOWER fan speed setting, and be happy. Result is that you get almost the same temperature at the same fan speed (same air flow for the same CPU workload and heat generation), and thus the fan NOISE is still the same! The main REAL difference you get by using the adapter is that at very HIGH workloads when high air flow is needed, the fan can NOT deliver the air! Its power supply has been restricted! You MAY experience overheating in that case.

If you WANT your fan to run slower (and quieter, with less cooling) that the default Standard "fan curve", you can set up your own custom fan curve for that. Within that you can choose to have the max fan speed used at the highest temps.

 

[35below0]

I think this was mentioned in the Noctua documentation, in more or less the same words. Power supply will be limited and with it max RPM lowered. But it doesn't say anything about headers. Most headers today are PWM headers so either Noctua is providing a pointless adapter or it doesn't interfere with PWM operation in the way you describe.

The system can read temperatures and can adjust RPM in response. They only thing that's changed is, it cannot exceed a lower power treshold.

The reason i chose to use the adapter is ... well, it was there in the box. I was getting very good results with a single fan installed, so when i installed the secondary fan i figured i'd increase cooling capacity enoutgh to not run into any regrets over using lower RPM limits. Silence is really, really precious to me. I don't mind high workload increasing fan speed, but i do love quiet idle RPMs.

BTW, even though i've lost the ability to see exact RPM, i can see % of max RPM. So it's possible to get a close enough number. Wonder why GCC isn't doing that? Maybe it doesn't know what the max RPM is?

 

[Paperdoc]

As I understand the LNA supplied by Noctua, it simply contains a small resistor in the line for Pin #2, which is the DC power supply line from any header. For older 3-pin fans the Voltage Control Mode of control by the header varies that supply Voltage from 12 VDC for max speed down to about 5 VDC for min speed without stalling. But in the new PWM Mode the voltage on Pin #2 is always 12 VDC, and the PWM control signal from Pin #4 is used by a special chip in the motor to modify the flow of current from that power supply through the windings to achieve reduced fan speed. So in BOTH cases the reduction of Voltage supplied on Pin #2 of the header imposed by the resistor in the LNA means a reduced fan speed compared to NOT using the LNA.

The matter of WHICH speed the header decides to request based on the temperature sensor reading is the fan PROFILE setting (commonly Normal, Quiet, Full Speed or Manual). Once that decision is made, the MODE setting controls how the signal is sent out to the fan to achieve the speed chosen. In either signal form, though, it really is a "% of Full Speed" signal - for example, see the manual Fan Curve graph in your post above. The system does not try for a specific speed, and actually does NOT use the speed reading to do its job. It only manipulates the fan speed signal to meet it TEMPERATURE target.

I am surprised, however, to see that using the LNA prevents you from "seeing" the fan speed. The speed signal (a series of pulses generated by the fan) is sent back to the header in Pin #3. I would have expected that the LNA unit would have a connection through it from female to male side for that pin so the signal can reach the header. From your posts it appears that is not connected through.

 

[35below0]

I will investigate further. But maybe it's because two cooler fans are connected via a Y-splitter to the fan header?
From the motherboard: CPU_FAN -> Y-splitter -> 2 low-noise adapter cables -> 2 140mm fans.

To be honest, noise has not changed at all sine i installed the adapters, so i do not need to use the them.
Or i could use the CPU_FAN and CPU_OPT headers? It is advised not to overload headers, so maybe i should. 😒

Thank you for your advice.

 

[Paperdoc]

You should not need to use two separate headers for 2 fans. One header can handle both via a Splitter.

Any header can deal with the speed signal sent back to it from only ONE fan. So when you use a Splitter, it will send back only one fan's speed, and all others on that Splitter are ignored. So with 2 fans on a Splitter with 2 outputs, you should see ONE fan speed on the host header. But your post suggests you see NO speed there. So either the LNA units are interfering with the signal, OR you are using a Splitter with THREE or more outputs, and you have failed to connect one fan to the ONLY Splitter output that can send back a speed.

HOWEVER, there is one small advantage to using two headers since you have the extra header to use. IF you connect only ONE fan to EACH header, then both headers WILL be able to display their respective fan's speed. IF you do this, then you must ensure that the configuration options for BOTH those headers are set the SAME so the fan behave identically. I do recommend that you do NOT use the LNA units. Plug each fan into a header directly, OR into two Splitter outputs if you choose that route. IF using a Splitter, look closely at the outputs and make sure to plug one fan into the only output that is marked differently.

 

[35below0]

The fans are connected like so: CPU_FAN header -> Y-splitter -> two low-noise adapter cables -> two 140mm fans.

https://imgur.com/32kwGqe

View: https://i.imgur.com/32kwGqe.jpg


My guess is the LNAs are interfering with the signal.

As for chaining fans together vs. using separate headers. I guess the advantage to using a single header is being able to control both fans using the same settings. Since they are tower cooler fans, they ought to be synced.

Giving each fan it's own header ensures no damage is done by excessive power draw? I'm not sure how common this is on a CPU_FAN header.

 

[Paperdoc]

FYI, the reason that Noctua recommends using a separate LNA unit for EACH fan us that the resistor value in the LNA is selected for the normal CURRENT draw of one Ncotua fan. IF you use it the WRONG way in the Noctua diagram, the current through the resistor is the total current for ALL fans connected after that, and that means the VOLTAGE DROP created by the resistor is too big!