- Fan control depends on the type of hardware, DC/PWM mode and possible conflicts between utilities.
- MSI Afterburner, WattMan, and the manufacturer's software allow for custom curves and stable profiles.
- Adjusting fans from the BIOS/UEFI prevents software issues and facilitates consistent airflow.
- On Surface computers, power plan, ambient temperature, and background tasks influence the effect.
On desktop PCs with a dedicated GPU, thermal management is usually automated (see what to do if the GPU fans not spinning); in portable and compact equipment the control is even stricter. In addition, there are peculiarities depending on the manufacturer (AMD, NVIDIA, motherboards, and even Surface computers) and depending on the type of fan (3-pin DC or 4-pin PWM)Here you'll find a complete, detailed guide in Spanish from Spain to help you understand why the fan speed doesn't change when you adjust it via software and what you can do to make it obey the speed.
Key concepts to understand fan control
Most modern graphics cards come with a factory fan curve which responds to the temperature of the GPU. As the degrees rise, the card increases RPM to prevent thermal throttling or overheating., and when there is a low load, many activate a 0 RPM mode and turn off the fans to reduce noise.
If you use a GPU integrated into the CPU, forget about utilities like Afterburner: On iGPUs, third-party software control does not apply; what you can do is improve processor cooling or disable the iGPU if you have a dedicated one available.. On discrete GPUs, however, you do have some leeway with software and the BIOS/UEFI.
Another vital concept is sensor visibility: the driver exposes applications temperature readings, speed and fan statesIf a program doesn't display RPMs or isn't able to change them, it could be due to a lack of permissions, an active control mode in other software, or the firmware's own policy.
When software doesn't change RPM: Typical causes
There are several reasons why you might move the slider and the rpm doesn't change. Understanding the real cause saves you hours of trial and error.:
- 0 RPM mode or manufacturer curve blocking: although disable 0 RPM, the factory curve may still be in charge if you don't enable user control in the correct software.
- Conflicts between utilities: Having Afterburner, the Radeon Adrenalin panel, or your GPU manufacturer's tool running at the same time can cause one to override the other.
- Fan type and control mode: On chassis and heatsinks, 3-pin fans are regulated by voltage (DC), 4-pin fans by PWM. If the motherboard connector is in PWM mode and you mount a 3-pin fan, the control may not take effect.
- Firmware or BIOS modified: cards used in mining or with flashed BIOS may have altered thermal tables and profiles, including the physical disconnection of fans.
- Sensor reading not available: : There are reports where the software does not show RPM and only by forcing a high state the fan starts; This indicates that the sensor or its exposure to the software is not active at rest. or there is a driver bug.
- Controller or sensor failures: In certain series, controllers have been seen that suddenly shoot 100% or that do not respond; If forcing 100% works, but the middle control doesn't, it could be a hardware issue..
- System power plan: On devices like the Surface, the power mode influences how and when the fans turn on; A high-performance plan allows for more temperature and more noise.
MSI Afterburner: Convenient and Reliable Control for Dedicated GPUs
- Launch Afterburner and unlock the profiles. Save a default profile so you can quickly go back when you need it.
- Go to the settings icon and go to the Fan tab. Activates user-defined automatic control and select a custom curve.
- Drag the points on the graph to set a speed at each temperature. Reducing the top end limits the maximum RPM and lowers noise., always monitoring charging temperatures.
- Apply the changes and test them in your games. Save the profile to a different number to the default to switch between whenever you want.
- If you are convinced, activate the option to start with Windows. Thus the profile will be applied automatically in each Boot without you having to open anything.
In advanced options there are useful technical parameters. The update period and hysteresis prevent sudden RPM changes, which translates into less noise oscillations. And if you'd rather not push it, you can set the firmware's control mode to read the manufacturer's limits and build your curve based on that.
Radeon Adrenalin (WattMan) and manufacturer software
On AMD cards, Radeon Adrenalin's own software includes WattMan to manage voltages, frequencies, and fans. The same principles of a custom curve apply here., with the advantage that everything is done from the official drivers.
Many assemblers include their application with predefined modes such as Silent or Eco. These profiles usually modify the fan curve and power limit from behind., and they can be a quick fix if you don't want to overcomplicate things. Just avoid having multiple utilities open at once to avoid overwriting settings.
Control from BIOS/UEFI and motherboard: DC vs PWM
If fine control of chassis and CPU cooler airflow is your thing, the BIOS/UEFI is your best ally. There are two ways to regulate speed: varying the voltage (DC) or using a PWM signal.. PWM is a pulse train with a duty cycle between 0% and 100%, and offers greater precision and better minimum RPM.
4-pin fans are designed for PWM and are identified as such by the manufacturer; 3-pin fans rely on DC. Many boards allow you to choose by connector whether they control by PWM or by voltage., and they can also regulate a PWM fan by reducing voltage, although this is not ideal.
Important: At 0% PWM most fans stop, but some continue to spin slowly by design. This behavior is usually specified by the fan manufacturer., and it is worth knowing if you are looking for absolute silence in repose.
To enter the BIOS/UEFI, you usually just need to press Delete when you turn on your PC (sometimes F2). The fans section may be called Hardware Monitor, Q-Fan, Smart Fan or similar. depending on the board. Some include a quick calibration that measures the minimum speed and minimum usable cycle for your fan.
Once inside, you will see a curve where the X axis is temperature and the Y axis is speed. Dragging points allows you to decide how many RPM you want at 40°C, 60°C, etc., in a stable manner and free of software problems in the operating system.
Two very common curve approaches: 1) Gradual: for example 20% at 20°C, 40% at 40°C, 60% at 55°C, 70% at 70°C and 100% at 90°C, ideal for maintaining a smooth progression of noise and temperature. 2) In steps: maintain a flat range and suddenly increase to 60% at 60-65 °C and to 100% at around 80 °C, useful if the temperature remains stable under load and you want to avoid constant changes in noise.
For CPU coolers, it is advisable to be a little more aggressive. A rule of thumb is to add approximately 10% PWM to the case fans. at the same temperature, especially if you do heavy tasks or play with sustained high loads.
BIOS or programs in Windows? If you are only setting the chassis ventilation, the BIOS/UEFI is more robust and avoids software conflicts.For the GPU, the manufacturer's tool or Afterburner offers more specific control over the card.
Laptops and Surface: Why the fan doesn't listen to you
On devices like Surface, the system adjusts fan speed based on load and environment, prioritizing sustained performance. If you work at temperatures above 25°C or while charging the battery, the fan may activate earlier and more intensely..
The Windows power plan directly influences this. Using Recommended mode reduces fan aggressiveness at the expense of limiting performance spikes, while in Better Performance the device will accept more heat and more RPM.
After the initial setup of the device, indexing, update downloads, and account synchronization processes take place for about 24 hours. This traffic increases CPU usage and wakes up the fan more often. until everything settles.
To detect gluttonous processes, open the Task Manager and sort by CPU. CLOSE apps that you don't need reduces heat and noise instantly. Keep drivers and firmware up to date from the Surface app to ensure thermal management is up to date.
Beware of abnormal noises: squeaking, rubbing, or knocking are not normal. The fan may sound loud when spinning fast, but it should never sound like a loose or damaged part.; in that case, contact support.
Real situations and warning signs
Cases like that of a Sapphire RX 7900 where RPM is not displayed on the panel, but when pushing the first state to 100% the fans start, are illustrative. That pattern suggests that the fan table exists and is working, but the RPM report or idle reading is not exposed to the software., or that there is a bug in the specific version of the driver/panel.
If your fans are not spinning at idle, this may be due to the 0 RPM mode thermal threshold. If they do not spin under sustained load, then there is a problem that you need to address., whether with a custom curve, a cleaning and inspection of connections or, if necessary, an RMA.
Is 100% speed dangerous? No. The card is prepared to run at full fan speed when needed.What can be annoying is the noise, which is why adjusting the maximum limit on the curve is so useful for balancing temperature and noise.
Will increasing RPMs give you more FPS? In theory, keeping the GPU cooler prevents it from throttling due to temperature. You won't exceed specs by going cooler, but you'll be able to maintain boosts longer., which stabilizes performance.
Will lowering RPMs cause performance loss? Yes, if the card reaches its thermal limit and reduces frequencies. An effective strategy is to combine a sensible fan curve with a slight undervolting of the GPU., minimizing heat and noise without penalizing performance.
There are also controller and reading failures in specific models, with sudden jumps to 100% or inaccurate readings. In those cases, a custom curve may mitigate the symptom, but the root cause is hardware or firmware. and may require official support.
How to create and test curves that actually work
When building your curve, consider three scenarios: low load (office work and browsing), medium load (moderate gaming), and high load (demanding gaming, rendering, editing). Look for a low load point above 40 °C of around 35-45% for acceptable silence., and rises decisively under medium/high load to ensure safe temperatures.
A gradual curve keeps noise changes smooth. Another option is a tiered curve that avoids continuous changes if your temperature stabilizes in a specific window during the game.If your GPU is always around 65-70°C under load, a flat area there will reduce noise fluctuations.
Saves multiple profiles: silent, balanced and performance. So you can alternate depending on what you are going to do. without having to redesign the curve each time.
Case and CPU fans: fine-tune the whole thing
It's not all about the GPU: the airflow in the chassis and the CPU cooler determine the graphics card's temperature. Well-tuned BIOS curves for the case and CPU offload the workload to the GPU fans., allowing to maintain lower RPM on the card without increasing temperatures.
If your motherboard allows it, run calibration to find out the actual minimums for each fan. That minimum may be lower than you thought and give you extra silence at rest., or conversely, prevent falling below a certain threshold without intermittent stops.
Special scenarios: used cards and firmwares
Second-hand cards used in mining often come with custom firmware, different power limits, and disconnected fans or extreme profiles. Check cables, restore the original manufacturer firmware and validate that the curve responds before accepting the card as valid.
If you flash the BIOS of another model, the fans may become uncontrolled. Always work with the specific BIOS of your card and, from there, we fine-tune with software. to avoid erratic behavior.
Rapid diagnosis when RPM "doesn't change"
- Check that there are not two programs controlling fans at the same time; leave only one.
- Disables 0 RPM and enables “user-defined control” in the chosen tool; apply and save.
- Force a high point in the curve to verify that the fan is operating and that the RPM reading appears.
- On motherboard, set correct DC/PWM mode per connector depending on the fan you use.
- Update drivers and, where applicable, firmware from the manufacturer's official software.
- On Surface computers, check power mode and background tasks that shoot CPU.
- If you hear abnormal noises or no response under load, consider support or RMA.
For more detailed questions, technical communities are a good resource, although they sometimes require registration to participate. Sharing screenshots of your curve, temperatures, fan model, and whether you use DC or PWM speeds up diagnostics..
With all of the above, you should be able to get the fan speeds to truly respond to your settings, either on the GPU with MSI Afterburner or WattMan, or in the motherboard BIOS for both the case and CPU. The optimal balance usually combines a logical curve, sensible maximum limits, an appropriate energy plan and, if appropriate, a slight undervoltage.With this base, you can forget about thermal shocks, maintain performance, and enjoy a much quieter PC.
Passionate writer about the world of bytes and technology in general. I love sharing my knowledge through writing, and that's what I'll do on this blog, show you all the most interesting things about gadgets, software, hardware, tech trends, and more. My goal is to help you navigate the digital world in a simple and entertaining way.