Guide: How to set up a fan curve in the BIOS

[citay]

Setting up a fan curve involves a balance between airflow and noise. But it is not that difficult when you know a good way to do it, and you only have to do it once.

First, a word about the fans. I'll try not to make it too difficult, there is just some precise terminology i have to use, but in the end it's not that hard to understand, i promise.

If you want to skip this part and go right to the information about setting up fan curves, scroll down to FAN CURVES.


There are two different types of fans, which you can tell apart by their plug. Let's look at this picture:

A three-pin plug means the fan is DC (= direct current) voltage-controlled. A 4-pin plug means it's PWM (pulse width modulation) controlled.
With a DC-controlled fan, the fan speed is regulated by the board powering it with lower or higher voltages (say, between 3V and 12V) instead of steady 12V.
With a PWM-controlled fan, the board powers it with a steady 12V, and the fan speed is controlled through the fourth pin via a PWM signal.

Note: The "RPM Speed Signal" pin (rotations per minute) in the picture is telling the motherboard at which speed the fan is spinning, it is not controlling the fan.

So for each fan, you have to select the right fan control method: DC for a 3-pin fan or PWM for a 4-pin fan.


Beware: Some cheaper motherboard models may only allow DC control for most of the fan headers, or only allow PWM control for all of the fan headers, regardless of the fan headers all being 4-pin headers or not. This may save them a little money on components around the fan headers, but it can be very inconvenient if you have the wrong type of fans and you can't control them properly as a result. So pay extra attention to that.

It's easy to notice in the BIOS: When you can't control some 4-pin fan headers with a PWM signal and instead only have DC (voltage) control available. Or when you don't have the ability to set a DC voltage for the fans, and you can only set a PWM percentage (then, when using 3-pin fans, they would always run at full speed here, because they would get constant 12V).

But you can also find out about this in the manual, before purchasing such a board.
Two examples of such boards. First, a board where they saved money on two of the headers:

Both CPUFAN headers can control the fan speed via PWM signal or DC voltage (depending on the fan and which control method you select in the BIOS).
But the SYSFAN headers at the bottom can only control the fan speed via DC voltage, despite having a 4-pin fan header.
"NC" means Not Connected, so they didn't implement PWM control on those headers for cost-saving. On more expensive board models, all the fan headers should be able to control fans with both methods. But in this example, you'd want your 4-pin PWM fans on the two CPUFAN headers, if possible.

Second example, from the MSI PRO Z790-S WIFI, the worst Intel Z-series board on the market:

This board simply doesn't offer DC Voltage control for any fan headers, only PWM control. So if you use 3-pin case fans (or CPU fans) on this board, since they don't have the fourth wire for the PWM speed control signal, they're stuck at a full 12V. There is no way to control 3-pin fans on this board at all, the fan curves will be useless for them. You need some kind of solution like a seperate fan controller, which some cases may offer.

These shortcomings are mostly on lower-end board models. Once you go for a slightly nicer board model (from lower mid-range onwards), they should be able to have all fan headers controlled in both ways, DC voltage or PWM.


By the way: Every PWM fan can also be DC-controlled, it's just a slightly worse method of controlling it. One advantage of PWM control is that the fan will always turn on, even at a very low setting (unless it's purposely configured to stay off below 5% or 20% PWM signal for example). But when you go too low with the voltage on DC control, the fan might not turn on reliably, as the voltage is not enough to overcome the fan motor's resistance, so you'd have to add a bit of extra margin on the voltage.

The general target for the fan curves is:
- Nice low RPM (fan speed) at low temperatures
- Let the RPM ramp up gently with rising temperatures
- Only ramping up the RPM faster when the temperature approaches a quite high level.


Now, before setting the fan curves: Since we're doing this in the BIOS, this is a good time to first update the BIOS to the newest version. Because if you decide to update the BIOS later, it will reset all settings (on a lot of boards, that includes the fan curves), and you have to enter everything again. But since we'll start from scratch now anyway, updating the BIOS beforehand is a good idea.

A quick how-to on BIOS updates:
1) Get the latest BIOS. It's always the topmost one when you click on "BIOS" on the MSI support page for the mainboard.
2) Extract the file and you will get a text file and the BIOS file. Put the BIOS file into the root folder of a USB stick/drive.
3) Enter the BIOS by pressing DEL during boot, go to "M-FLASH" in the BIOS.
4) Once M-Flash (the updater) is loaded, it will show a list of your drives. Select the USB stick and select the previously extracted BIOS file on there.
5) It will ask for confirmation and then update the BIOS. It's fully automatic from there, takes about two minutes.


Now, before we come to the topic of how to set good fan curves, let's first look at a proper way to create an airflow through the case. Normally there should be at least one intake fan (usually at the front) and at least one exhaust fan (usually at the rear, in line with the CPU cooler). This will ensure a defined airflow through the case which can extract the heat from the components.

All the fans should work in unison within that airflow, to get one stream of cold air coming in from the front, picking up some heat (mostly from the CPU and GPU), and the warmed-up air being exhausted out of the rear. So in modern cases, the airflow usually looks something like this:

Perhaps minus the fans on top of the case. But there often will be at least one fan (if not two or three) in the front, and one exhaust fan in the rear behind the CPU cooler. Each fan will have one or two arrow markings on one of the sides, showing the rotational direction as well as the airflow direction, and all the "airflow arrows" should be pointing towards the rear of the case. When the fans come pre-installed in the case, they will already be installed like that.

If the front intake fans outnumber the rear exhaust fan by 2:1 or 3:1, then the rear fan ideally runs on slightly higher RPM than the front fans, to create a more even airflow.

About over- vs. underpressure, or positive vs. negative pressure:

Usually there should be a slight overpressure / positive pressure, meaning, slightly more air intake than air exhaust. Because the problem with underpressure or negative pressure is, it can't be controlled where the air is being pulled from. So when there are more exhaust fans than intake fans (or there's more powerful exhaust fans, or they spin faster), they will usually pull air from the shortest path. Meaning, all the additional air does not come from the front, it will be pulled from every vent and case opening that is nearby.

But since we want a defined airflow, usually front-to-back/top, the best thing is to have a slight bias towards the intake fans. And most cases will naturally do that, as they tend to have more intake than exhaust fans. Then the exhaust fan should be on a slightly higher fan curve to spin a bit faster, but it's ok to still have some air blowing through the case vents below the rear exhaust fan, having the the front fans move slightly more air into the system than the exhaust fan alone can move out. Then the warm air that collects below the graphics card, or generally in the lower half of the system, can also slowly be replaced by fresh air as it exits through the rear case vents. You don't have to feel real wind coming through the vents, but it's good if the air is not totally stagnant, and you definitely don't want the air to be pulled in from there.



FAN CURVES

Now it's time to set the fan curves. Enter the BIOS (press DEL after power-on/reboot) and open the "Hardware Monitor" which offers the fan control.
For each fan, you can set four points of a curve, MSI calls this the "Smart Fan Mode".

This is how it might look:

Note that each of the four points of a fan curve is restricted by the points next to it, it can't go lower than the previous point or higher than the following point.
So you may have to move a neighboring point if you hit a restriction on the point you want to adjust.

The goal for the first point of the curve is to find a setting with a bit of airflow, but where the fan is very quiet. This will be the setting when the CPU is doing nothing (idle).
You don't need a lot of airflow when the CPU temperature is low anyway. My fans spin only at around 400 RPM there, as you can see, just enough to keep some air moving through the case. Note that i have a high-end air cooler with two fans on there, as well as three 140mm case fans in a large case. So in a small case with fewer fans, you will need a bit higher RPM to keep it this cool inside. But there is no need to have the fans spin for example at 1000 RPM in idle.

For testing, it's good to open the case and put your hand behind the fan or behind the cooler (where the air gets blown through) to feel how much airflow the fan generates with different values. As long as you don't touch the motherboard or other components, there's no danger. And you can immediately feel the results of your adjustments.

For testing the airflow, turn off "Smart Fan Mode" for a while, so you can influence the fan speed directly. You can also turn off the other fans, so you can find the sweet spot for low temperatures with the particular fan you're checking. Remember, you only need a slight airflow for this starting point of the fan curve, the goal is not to have unnecessary noise when the temperature is low.

For this goal (a bit of airflow, but being very quiet), the resulting PWM % value (or DC voltage on 3-pin fans) is your starting point at 30°C or 40°C for this fan's curve. The temperature you select depends on where you want the fan to first start ramping up. It makes no sense to define a temperature below the ambient temperature, or below the minimum CPU temperature, you'd just be wasting the whole adjustment point. So anything below 30°C only makes sense with powerful cooling methods that can actually hold the CPU below 30°C.

Next, don't go to the second point, but the third one. Find a good level where the airflow is strong but the noise is still bearable, and use this as your "full CPU load" setting for higher temperature values like 65°C or 75°C. To fine-tune this point of the curve, you might have to go back and forth from the BIOS to Windows, where you monitor the fan speed with certain CPU load and fine-tune the setting afterwards. But it doesn't take that long to do, and you only have to do it once.

As for the second point of the curve, the inbetween point: Set it slightly below a straight line between the first and third point, to not make the fans ramp up too fast at medium temperatures.

For the final point, set it for 85°C or 90°C CPU temperature and 100% PWM value (or full 12V DC with a 3-pin fan). This is the "worst case" point for safety.
Now you should have all four points of the curve set to a sensible value, and most of the time, the fan should stay between the first and the third point. The highest last point is just a safety measure.

I would always create such an "ascending dominant" curve:

A perfectly straight line makes no sense, you'd be wasting the two middle adjustment points.
A descending dominant curve makes no sense, as it will make more noise than necessary.

Here are examples. You might of course have to use different values, but just to see how it should rougly look like. The curve needs to be done for your particular fans.

This is for a PWM fan (4-pin), which is controlled by a PWM % signal:

I would always prefer "Temperature source" CPU for the CPU fan(s), and probably most other fans too.
Usually, the CPU temperature is the most important one, and will influence the other temperatures.
By the way, a "Step down time" of 1.0s makes the fans spin down less audibly.

However, it's also feasible use "Temperature source" System for case fans. Here's an example for a DC (voltage-controlled) 3-pin fan.
Note: The following picture is not how it should be set when using the CPU temperature as the source (which is the usual temperature source),
it's an example of a fan curve when using the System temperature as the source, which could be done for case fans.
When using CPU temperature as the source, the fan curve should not be this steep, it should be more like the pictures above.

Of the six total fans in my PC (two on the CPU cooler, four in the case), this lower front intake fan is the only one for which i use the "System" Temperature Source. I want it to react to the system temperature with a steeper curve, since the system temperature will obviously increase much more slowly than the CPU temperature. The graphics card can be a major contributor to heating up the whole system, and since i can't use a graphics card sensor as the temperature source, this is sort of a roundabout way to handle that. Of course, you can also have the CPU as the temperature source for all the fans, then you should use a shallower curve like in the pictures before this one.

Here's a 4-pin system fan (in my case, a Noctua NF-A14 PWM) with the CPU as the temperature source:

In my system, the CPU temperature actually stays very low under load, since i have a mid-range CPU and a high-end CPU cooler. Therefore, even for the system/case fans, i can set this steep of a fan curve and not worry about getting jet engine noises under load, it will still stay pretty quiet. Depending on the individual configuration, it might be better to have the system fans on less steep of a curve compared to the CPU fan(s), just so the noise is less annoying under load. It all depends on your setup, what kind of fans you have, and noise/temperature preference.

When you have a graphics card (GPU) with a pretty high power draw, and/or there will be a lot of gaming done on the PC, then it's also a good idea to observe the various temperatures during a stint of gaming, running HWinfo Sensors in the background, set up as mentioned in step 1) of this guide with the sensors expanded. Because apart from the CPU temperatures, the GPU and other temperatures (SSD, RAM, System...) are of course also important. A GPU can take quite high temperatures, and for some cards it's not unusual or concerning to see high 80°C there under load, that can be pretty normal. But it shouldn't cause excessive temperatures for other parts. So while thinking about achieving a good balance of temperatures vs. noise, it's good to keep the entire system in mind, for example in a gaming scenario, and not only test full CPU load.


Each fan model has a different RPM range and therefore needs different values or voltages to reach a certain airflow. Also, each different PWM-controlled fan model can interpret the PWM signal differently. This is because some mainboards don't allow a PWM signal lower than 20% for example, to never have the fan turn off. So to circumvent that, a fan maker might decide to let the fan interpret a 20% PWM signal from the mainboard as "still turned off", and turn on at 21% PWM. Another fan might interpret 0% PWM as the turn off signal and 1% as the lowest possible RPM, or may never turn off and just map the entire PWM signal range to the lowest and highest RPM. And all variations in between.

Here's an example of an Arctic fan which has implemented a semi-passive mode, it stays off with any PWM signal below 5%:

These Arctic P12/P14 PWM PST are very good fans by the way, almost unbeatable in price/performance.

Another example of a fan with a semi-passive mode is this Noiseblocker one, which has the following PWM-signal-to-fan-speed mapping (depending on the variant):

So for each fan model, you will need different PWM values (and for 3-pin ones, different voltages), but just go by airflow and noise.
And the concept of the fan curve is always the same.

Once you're done, it's a good idea to write down your settings or make a screenshot/picture of them (in the BIOS, F12 saves a screenshot to a FAT32-formatted USB drive).
Because whenever the BIOS settings are reset (due to BIOS update, CMOS Clear or empty battery), you'll need your notes or pictures to know what fan curves you had before.
Although in BIOS updates for newer boards, the fan curves are actually kept between updates nowadays, which is good.

My other guides:
RAM explained: Why two modules are better than four / single- vs. dual-rank / stability testing
Guide: How to set good power limits in the BIOS and reduce the CPU power draw
Guide: How to find a good PSU

Someone asked me if they can thank me for my work by sending me something via Paypal: Yes, that's possible, just write me a message and i'll tell you my Paypal

 

[Changeemailaccount2025]

I think you mean me, not the post above you that you quoted.

Yes, for 4-pin PWM fans, you set PWM control which is a value in percent. For 3-pin voltage-controlled fans, you set DC control. The PWM fans always get a constant 12V, their speed is fully controlled via the PWM percent value. The DC fans get a varying voltage, the minimum voltage to start spinning might be roughly around 3-5V for some fans, and with 12V they have their maximum speed. Your board will take the 12V straight from your PSU, and for the 3-pin fans, it can adjust it freely between 0-12V.


MY question - Bios sets up the DC speeds to run as pwm entering not volts what is the problem with this they are not all running on dc just three and two pwm fans?

 

[Changeemailaccount2025]

Hey there again and (soon to be) merry Christmas o/


Apologies for taking so long with the fans, they did arrived around 2 weeks ago but one of my shoulders was so sore I couldn't lift heavy things so I had to wait xD

I've finally changed them today,
I had to use one of them in the Pump connector since fan 5 pins were at the bottom of the mobo and the cable didn't reached that far, so I hope it's alright.


Here the 3 fans settings so far,
System 1 is the 1200 rpm rear fan,

System 6 and Pump are the 1800 rpm fans you suggested me to get.

View attachment 196998
View attachment 196999
View attachment 197000

Btw when I've reconnect the PC it took it few mins to load up, just like how it happens when I change bios settings but I didn't changed anything and old settings were kept as is, and the mobo battery is far from being dead so I hope it wasn't some bad sign...

Just wondering why dont u have Smart Fan Mode connected to this DC curve and source up to CPU temperatures sensor on the pictures u listed for bios control>>????

 

[citay]

Here is the curve so far, by middle you mean mostly messing with the 30c to 50c dots?

Mainly the 50°C and 75°C ones, but you can also lower the 30°C one a bit, i think this fan can easily go to lower RPM in idle. Remember, when the CPU temperature is very low, you barely need any airflow through the system, because there's barely any heat produced. So the fan speeds can drop to 500 RPM or lower, if the fans are capable of it. No need to cool something much that is already cool. And if you get rid of an annoying sound at the same time, that's ideal.

Consider I don't hear them much at 600-650 rpm, and that the CPU fan also at 600 rpm most of the times, won't it be better?

Whatever gives you better results is better. Since i don't sit in front of your PC, i cannot give absolute advice. You are the best judge of it.

Though the upper fans got their own job taking out the hot air so maybe they do need to be faster than the CPU's, but as I'm uncertain in it considered I'm no expert in these kind stuff I rather wait for your answer (sorry if I'm making something obvious sound more harder than it should be xD).

Just keep an eye on the temperatures and the noise. If they don't get out of hand, then the configuration can't be that bad. You can compare the results quite easily with the help of HWinfo and Cinebench, or just gaming, letting the sensors open in the background as well as judging the noise. Make a few notes, and then do another comparison run with modified fan curves.

It's good to theorize about things, but sometimes you simply have to do a back-to-back comparison to see which settings are the best compromise between cooling performance and noise.

 

[citay]

3-pin set that came with my case at 1200 rpms bios curved dc running in the smart fan mode curve selector it does not give VOLTS just a system fan temp and rpm % with no volt in bios hardware monitor, cause they have intake setup i would want to know why my bios does not connect to the motherboard by volts and *10mv is the standard on it which means nothing to me right? I just get the basics of bios fan control on intake fans running lower than aio top mounted fans exhausted to push neutrally but the dc system does function on its curve of two bottom pwm fans and three dc front intake fans.

Not sure i understood all of this. You're saying, you have 3-pin fans, but you can't set DC Voltage control for them, you only have PWM control available?
Yes, depending on the board model, this can also happen. I just edited the first post with an example of this, here is the relevant section i added:

Example from the MSI PRO Z790-S WIFI, the worst Intel Z-series board on the market:

This board simple doesn't offer DC Voltage control for any fan headers, only PWM control. So if you use 3-pin case fans (or CPU fans) on this board, since they don't have the fourth wire for the PWM speed control signal, they're stuck at a full 12V. There is no way to control 3-pin fans on this board at all, the fan curves will be useless for them. You need some kind of solution like a seperate fan controller, which some cases may offer.

These shortcomings are mostly on lower-end board models. Once you go for a slightly nicer board model (from lower mid-range onwards), they should be able to have all fan headers controlled in both ways, DC voltage or PWM.

how is it that i dont see a pump option for my computer bios ? i understand the point of connection via i dont have a connection for my aio production year, it only has a cpufan plug so i cant not do that part right?

What's your board model? Then i can look up in the manual what is going on.

Hi read this if its accurate and critique =- point one is at 500rpm for system fans intake and exhuast for idle, point two is mid level temps at 1000rpm for intake and exhaust, what are these temperatures for bios editing? also might as well throw out a 100% temp cap too, my next question is what is a good dot to add for the fourth in between temp and fan% ?

Personally, i use these four points for the CPU temperature: 30°C, 50°C, 75°C, and 90°C. Look at the screenshots in the first post to get an idea. Note, my fan curves might be close to ideal for my system, but everyone has a different system, so it's better to explain the principles than to give exact settings for someone. Just do some testing to find a good compromise of cooling performance and noise, working with the general principles i explained.

 

[Changeemailaccount2025]

Not sure i understood all of this. You're saying, you have 3-pin fans, but you can't set DC Voltage control for them, you only have PWM control available?
Yes, depending on the board model, this can also happen. I just edited the first post with an example of this, here is the relevant section i added:

Example from the MSI PRO Z790-S WIFI, the worst Intel Z-series board on the market:

This board simple doesn't offer DC Voltage control for any fan headers, only PWM control. So if you use 3-pin case fans (or CPU fans) on this board, since they don't have the fourth wire for the PWM speed control signal, they're stuck at a full 12V. There is no way to control 3-pin fans on this board at all, the fan curves will be useless for them. You need some kind of solution like a seperate fan controller, which some cases may offer.

These shortcomings are mostly on lower-end board models. Once you go for a slightly nicer board model (from lower mid-range onwards), they should be able to have all fan headers controlled in both ways, DC voltage or PWM.




What's your board model? Then i can look up in the manual what is going on.



Personally, i use these four points for the CPU temperature: 30°C, 50°C, 75°C, and 90°C. Look at the screenshots in the first post to get an idea. Note, my fan curves might be close to ideal for my system, but everyone has a different system, so it's better to explain the principles than to give exact settings for someone. Just do some testing to find a good compromise of cooling performance and noise, working with the general principles

PWM FAN + DC fans seem to sync if i enter enough different rpms on the bios control however i keep changing the numbers on it over the last year but still confused on making a perfect 1 and 3 number for idle and mid loading gameplay for silent rpm of 400 and mid level 1000rpm, cause it wants to ramp up the fans rpm and i keep changing my exhaust AIO in ICUE CORSAIR Progam to keep up the intake flow of air above 725 rpm now idle 46 c-55c, and another 10c change in games but the two fan exhaust and intake arent made on bios to match. the fans are 5 intake and 2 aio exhausting.
Motehrboard MSI B450 Tomahawk the type u wanted to know, beta bios controlled on new bios.
ive copied the CPU temperatures for the four points of cpu curve above , and tested on games, it performs but its not silent on noise ramp up for the discussion on setting one idle and thirdly go game curve and one safety measure, but adjusting the rpm slightly is my goal just wanting to get the best performance on my system, and could i copy ur settings for cpu temperture source dc and pwm fans like the discussion said or does it not get support like for me changing dc to match rpm and noise changes like the pwm non 12v gets.?

 

[citay]

but the two fan exhaust and intake arent made on bios to match.

You don't necessarily need to match everything 100%. If you use different fan models, and even different ways of controlling them, then it's enough to have them behave somewhat similarly and not be too far apart. Because you also have to take into account that the fans don't all create the same airflow at the same RPM, due to different rotor design and so on. So if you get two different fan models to exactly the same RPM for any given CPU temperature, then they will still have a discrepancy between them, one will move more air than the other.

Instead, go more by the results you get, look at the temperatures and listen to the noise. This is where you have to find a good compromise via trial and error.

Motehrboard MSI B450 Tomahawk the type u wanted to know

Looks good, seems that all fan headers support both modes, DC and PWM control:

BTW, on these 300- and 400-series non-MAX boards, the BIOS chip size was only 128 MBit (16 MB), so in order to save space and still be able add support for those newer CPU models, they had to remove support for the oldest CPU models from a certain BIOS version onwards, and switch to a GSE-Lite mode, also see here. The MAX versions of the boards later came with a 32 MB BIOS chip instead of 16 MB, to support more CPU models at once and still keep the full BIOS layout available. So on the non-MAX boards with the smaller BIOS chips, they are trading some BIOS functionality for having the newest CPUs supported on such an older board. The result is that you should see a text-based BIOS without the nicer fan curve adjustments.

ive copied the CPU temperatures for the four points of cpu curve above , and tested on games, it performs but its not silent on noise ramp

As i mentioned before, my fan curves are ideal for my system, and i have very low noise with them, because my cooler is overkill for my CPU. If your cooler has more trouble with your CPU, and the CPU gets hotter, and you have different fans than me etc., then the result will be louder. Either you have to accept higher temperatures and have less noise as a result, or if you want the same temperatures at lower noise, you ultimately might need a more capable CPU cooler. But usually a decent compromise can be found, unless the cooler is completely underpowered.

but adjusting the rpm slightly is my goal just wanting to get the best performance on my system, and could i copy ur settings for cpu temperture source dc and pwm fans like the discussion said or does it not get support like for me changing dc to match rpm and noise changes like the pwm non 12v gets.?

Think of my fan curves as examples, as starting points. Of course you can copy them, but you won't get the same results if you don't have my system. I pair one of the best and largest air coolers from Noctua with a mid-range CPU that doesn't need that much power. For your circumstances, you might need something a bit different, otherwise it gets too loud.

 

[Changeemailaccount2025]

You don't necessarily need to match everything 100%. If you use different fan models, and even different ways of controlling them, then it's enough to have them behave somewhat similarly and not be too far apart. Because you also have to take into account that the fans don't all create the same airflow at the same RPM, due to different rotor design and so on. So if you get two different fan models to exactly the same RPM for any given CPU temperature, then they will still have a discrepancy between them, one will move more air than the other.

Instead, go more by the results you get, look at the temperatures and listen to the noise. This is where you have to find a good compromise via trial and error.



Looks good, seems that all fan headers support both modes, DC and PWM control:

View attachment 197231

BTW, on these 300- and 400-series non-MAX boards, the BIOS chip size was only 128 MBit (16 MB), so in order to save space and still be able add support for those newer CPU models, they had to remove support for the oldest CPU models from a certain BIOS version onwards, and switch to a GSE-Lite mode, also see here. The MAX versions of the boards later came with a 32 MB BIOS chip instead of 16 MB, to support more CPU models at once and still keep the full BIOS layout available. So on the non-MAX boards with the smaller BIOS chips, they are trading some BIOS functionality for having the newest CPUs supported on such an older board. The result is that you should see a text-based BIOS without the nicer fan curve adjustments.




As i mentioned before, my fan curves are ideal for my system, and i have very low noise with them, because my cooler is overkill for my CPU. If your cooler has more trouble with your CPU, and the CPU gets hotter, and you have different fans than me etc., then the result will be louder. Either you have to accept higher temperatures and have less noise as a result, or if you want the same temperatures at lower noise, you ultimately might need a more capable CPU cooler. But usually a decent compromise can be found, unless the cooler is completely underpowered.



Think of my fan curves as examples, as starting points. Of course you can copy them, but you won't get the same results if you don't have my system. I pair one of the best and largest air coolers from Noctua with a mid-range CPU that doesn't need that much power. For your circumstances, you might need something a bit different, otherwise it gets too loud.

yea thank you, the information of smaller support mb helps me out occurred with the change back when i found out the support changed i never gamed on the system for long cause i did not know the issue in bios settings changed smaller form than a non max b450 tomahawk.

 

[Changeemailaccount2025]

yea thank you, the information of smaller support mb helps me out occurred with the change back when i found out the support changed i never gamed on the system for long cause i did not know the issue in bios settings changed smaller form than a non max b450 tomahawk.

Any chance you read a solid temperature for b450 motherboard intake and exhaust and CPU low load temp + gaming temp for fans running on intake 1400 rpm pwm and 1200 rpm DC and exhaust AIO control for bios IE 30c low load temp, 75c high load gaming temp, but those could be different for fans in intake rpms. thanks

 

[citay]

I'm not sure what information you want, that sentence was difficult to understand.

 

[Changeemailaccount2025]

I'm not sure what information you want, that sentence was difficult to understand.

thanks for reply, update to the part im confused on, when i manually input the PWM PERCENTAGE/DC RPM FOR BIOS ON AM4 ryzen 5000k motherboard b450 tomahawk it has to be entered lower than the AIO by 10 percent rpm difference, according to the readout on this thread, thats enough to make it sound less sounds during any part of the temperature changes, here is the update to that photo attached. cause its only in idle i have not put one in for a separatee graph on loads, what is the problem here? such as i set the system fans to work with a little more temperatures than the cpu temperature target would that cause the system intake fans to get more rpm than the regular Exhausted fans?

 

[citay]

So you get different RPM for the intake vs. the exhaust fans, on the same exact fan curve, is what you're saying? And these are identical fans? Well, fans naturally have a bit of a variance/tolerance in RPM, they are never 100% the same. Then, the RPM is also influenced by something being in the way, either in front or behind the fan.

Of course, if the fans are different models, then this is completely normal. They all behave differently to the same PWM percentage or DC voltage. To start with, different fan models have different RPM ranges. For one PWM fan, it might go up to 1600 RPM, another fan model might go up to 2000 RPM. So right there, for a 100% PWM signal, you have 400 RPM delta between the two fan models. And according to their amount of rotor blades and their geometry, they will additionally differ in the air they move at a certain RPM. So with different fan models, you don't aim for them to have the exact same RPM, you only set them so that the end result is good. Good temperatures at as low noise as possible. Just a bit of trial and error.

 

[Changeemailaccount2025]

My problem of aim results are the good temperatures, and noise level ive done this for months but recently followed this guide to do it once (Cpu Temperature source for PWM and DC signal on the curve) sources are hard to match the intakes temperatures to the idle noise increased rpm of likely around 900rpm, what sort of curve do i make to get the easiest non ramp up curve signal? IE. the CPU Fan 1 is set around 1045 rpm after putting them on the source it gets 150 rpm less for intake, and cpus aio fans aren't controlled for bios only software, so they make rpm degrade on the case cooling as well.

 

[Changeemailaccount2025]

So you get different RPM for the intake vs. the exhaust fans, on the same exact fan curve, is what you're saying? And these are identical fans? Well, fans naturally have a bit of a variance/tolerance in RPM, they are never 100% the same. Then, the RPM is also influenced by something being in the way, either in front or behind the fan.

Of course, if the fans are different models, then this is completely normal. They all behave differently to the same PWM percentage or DC voltage. To start with, different fan models have different RPM ranges. For one PWM fan, it might go up to 1600 RPM, another fan model might go up to 2000 RPM. So right there, for a 100% PWM signal, you have 400 RPM delta between the two fan models. And according to their amount of rotor blades and their geometry, they will additionally differ in the air they move at a certain RPM. So with different fan models, you don't aim for them to have the exact same RPM, you only set them so that the end result is good. Good temperatures at as low noise as possible. Just a bit of trial and error.

Same question above last post- My problem of aim results are the good temperatures, and noise level ive done this for months but recently followed this guide to do it once (Cpu Temperature source for PWM and DC signal on the curve) sources are hard to match the intakes temperatures to the idle noise increased rpm of likely around 900rpm, what sort of curve do i make to get the easiest non ramp up curve signal? IE. the CPU Fan 1 is set around 1045 rpm after putting them on the source it gets 150 rpm less for intake, and cpus aio fans aren't controlled for bios only software, so they make rpm degrade on the case cooling as well. I dont want to see temps currently as on and using barely windows get to 60 C with the cooler dropping with loud noise fans to 49c, and not changing correctly

 

[citay]

"Non-ramp up curve signal"? Not sure what you want here. Like i said many times in this thread, for idle, you want to aim for very low RPM, i have my fans at around 500 RPM there. So you set the first point to 30°C or 35°C for example, and then check what kind of PWM percentage (or DC voltage for a 3-pin fan) result in such a low RPM. Because in the BIOS you should have a low CPU temperature, so it's easy to check this. If your CPU temperature in the BIOS is higher, because it's an AMD, then you can even set the first point at 40°C.

But yeah, i don't quite understand what the exact problem is, i explain all the principles in the first post, i can only explain them all over again in different words. I'm happy to do so, but first i have to understand better what the issue is. Maybe it's a language barrier, but somehow i don't understand it properly, sorry.

 

[Changeemailaccount2025]

"Non-ramp up curve signal"? Not sure what you want here. Like i said many times in this thread, for idle, you want to aim for very low RPM, i have my fans at around 500 RPM there. So you set the first point to 30°C or 35°C for example, and then check what kind of PWM percentage (or DC voltage for a 3-pin fan) result in such a low RPM. Because in the BIOS you should have a low CPU temperature, so it's easy to check this. If your CPU temperature in the BIOS is higher, because it's an AMD, then you can even set the first point at 40°C.

But yeah, i don't quite understand what the exact problem is, i explain all the principles in the first post, i can only explain them all over again in different words. I'm happy to do so, but first i have to understand better what the issue is. Maybe it's a language barrier, but somehow i don't understand it properly, sorry.

yea would i need to keep system fan headers on system temperatures rather than the cpu curve for fans on intake and exhaust? Im happy to try both, however its not reading rpms correctly, i set the cpufan pump header to its 40c and still went even further for the problem of noise of bios curves with AMD /b450 chipset, ive got better shot of the way i know according to HWINFO i get loads temps of 37-40 with the fans correctly at 600 rpms, with 37c temperatures on cpu, i can sustain that with it atm idle, but im thinking i just enter the 35c for the first for fans at 26% cooling, 40c for the cpu source first at 33% cooling volume, and on the system i adjust the SECOND on a middle between 1 and 3 setting* system 53C at 42% along with the CPU #2 53C at 54%, i adjust the third point on the graph bios system 70C / 67% and CPU#3 at 75C/75%, all the next curve point is 85C/100%. Do u get what im saying that its still configured to run off of header that seems to keep the operation above the current question somehow to massively keep the rpm above its needed source temp.? DC and PWM were pretty close to each other though, but id like to stay around the reading here. ill type over the settings or screenshot as well. -

 

[citay]

yea would i need to keep system fan headers on system temperatures rather than the cpu curve for fans on intake and exhaust?

If you want to use System temperature as the source for the fan curve, then the curve would have to be quite different from one that's using CPU temperature as the source. Because the System temperature has a much smaller range and reacts much more slowly. So i would usually say to use CPU temperature as the fan curve source for most fans.

Most of the other things you write, it's difficult to understand, i just cannot make sense of it. Let's look at one sentence for example:

"Do u get what im saying that its still configured to run off of header that seems to keep the operation above the current question somehow to massively keep the rpm above its needed source temp.?"
->
"Do you get what i'm saying? That it's still configured configured to run off of the header that seems to keep the operation above the"... and then it just doesn't make that much sense anymore.

Something about the RPM being higher than necessary for the current CPU temperature, i guess? Well, if you see higher RPM than you like, just set a lower PWM percentage in the BIOS. Experiment a bit. You don't need me for that, once you understood how things work. Just set the curves how they work best for you.

 

[Changeemailaccount2025]

If you want to use System temperature as the source for the fan curve, then the curve would have to be quite different from one that's using CPU temperature as the source. Because the System temperature has a much smaller range and reacts much more slowly. So i would usually say to use CPU temperature as the fan curve source for most fans.

Most of the other things you write, it's difficult to understand, i just cannot make sense of it. Let's look at one sentence for example:

"Do u get what im saying that its still configured to run off of header that seems to keep the operation above the current question somehow to massively keep the rpm above its needed source temp.?"
->
"Do you get what i'm saying? That it's still configured configured to run off of the header that seems to keep the operation above the"... and then it just doesn't make that much sense anymore.

Something about the RPM being higher than necessary for the current CPU temperature, i guess? Well, if you see higher RPM than you like, just set a lower PWM percentage in the BIOS. Experiment a bit. You don't need me for that, once you understood how things work. Just set the curves how they work best for you.

Might be a heat issue on the chips old thermal paste on startup i wont get under 48.8 celcius on the motherboard header reading, therefore isee a minimum of 890 rpm with the cooling completely on quiet setting for the pump, the rpm fans in bios have a auto detect feature i notice if i put on the auto feature it turns my AIO PUMP fans to the DC not pwm selection but when i select the PWM it gets the read out as DC did, would this matter to auto cpufan header option or select the four pin like they are plugged into the AIO adapter to Cpu-fan 1 header on motherboard, thats all im really asking thanks for the help.

 

[citay]

Well, for AMD, you're not gonna have as low BIOS temperatures as on Intel, but at least in Windows it should probably stay below or closer to 40°C in idle (unless it's quite hot in your room). But yeah, if you never get below 40°C, then set the first point of the curve at least to that, so you can properly control the RPM for that state.

What i find weird on your recent screenshot from a couple replies ago, the CPU has way too high power consumption there, if that was supposed to be idle. Over 47W CPU Package Power as the minimum, too much even for AMD. Are you maybe using a different power plan than "Balanced", the only proper one? Or is there something taking up CPU resources in the background?

I would accept maybe up to 20, 30W for an AMD in idle, but almost 50W? This is way too much in idle. So no wonder the temperatures are not that good.

 

[Changeemailaccount2025]

package power is 45W min to 64.6W average, of max 145W. My plan has separate upgrade with the ryzen cpus called Precision boost overdrive gets me better mhz, i forgot about the curve optimizer let me know if its just wasting energy with that it should consume less too, but idle i could use ryzen master to put it into a eco mode, or better put new thermal paste on the cpu aio.

 

[Changeemailaccount2025]

Well, for AMD, you're not gonna have as low BIOS temperatures as on Intel, but at least in Windows it should probably stay below or closer to 40°C in idle (unless it's quite hot in your room). But yeah, if you never get below 40°C, then set the first point of the curve at least to that, so you can properly control the RPM for that state.

What i find weird on your recent screenshot from a couple replies ago, the CPU has way too high power consumption there, if that was supposed to be idle. Over 47W CPU Package Power as the minimum, too much even for AMD. Are you maybe using a different power plan than "Balanced", the only proper one? Or is there something taking up CPU resources in the background?

I would accept maybe up to 20, 30W for an AMD in idle, but almost 50W? This is way too much in idle. So no wonder the temperatures are not that good.

But yeah, if you never get below 40°C, then set the first point of the curve at least to that, so you can properly control the RPM for that state.-????
Does that mean first point should be 46C?