There are times when the video card resets the driver, or the image may disappear for a short time during intense gaming. One of the reasons for this behavior may be VRM Overheating of the video card. Let’s talk about diagnostics and ways to solve this problem.
What is VRM?
Before determining the causes and solutions for overheating of the VRM (voltage regulator and converter) of the video card, you should consider the device of the video card itself. Nearing the end of 2021, modern video cards consist of several tens of millions of transistors, consume up to half a kilowatt of energy and have very complex cooling systems. Readout Motherboard VRM Heatsink.
But it was not always so. To find out what key elements a video card consists of and how these components have evolved, for example, let’s take the Nvidia GeForce 7800 GT video card, which appeared back in 2005.
So, the main components of the video card:
- Graphics Processing Unit (GPU)
- Memory chips (VRAM)
- GPU and VRAM (VRM) system and power circuits
- Interface connector for connecting to the motherboard (PCI-Express)
- Power connector(s)
- Interface connectors for connecting monitor/s
- Image input/output connector(s) to/from other sources (optional)
- Connector for combining video cards in SLI or CrossFire mode (optional)
- Video card cooling system
In the mid-2000s, powerful graphics cards typically had a single power connector, and their TDPs were less than 100W. The power consumption of the Nvidia GeForce 7800 GT considered as an example was then only 65 watts.
Now let’s figure out what components the VRM of the video card consists of. To regulate and convert the voltage supplied to the video processor and video memory, a voltage supply regulator or VRM is used, which consists of the following main components: MOS transistors (they are also called mosfets, MOSFETs), inductors (chokes) and capacitors. These elements are controlled by a PWM or PWM controller.
The VRM of video cards from the mid-2000s usually consisted of 1-3 phases, where the GPU accounted for one or two phases, and one phase accounted for the power of the video memory. Since the amount of energy consumed by such video cards was small by modern standards, the VRM zone was cooled by the airflow blown out by the video card cooling system or was left without active airflow and any radiators at all. Below is a photo of an Nvidia GeForce 7800 GT with a cooling system, where the uncovered VRM zone is clearly visible.
Now let’s fast forward to today. Let’s take the Asus ROG Strix LC GeForce RTX 3080 Ti OC Edition as an example. This modern and powerful graphics card has 12 GB of video memory, 22 power phases (18 for GPU and 4 for VRAM), and its peak power consumption is as much as 400 watts. For 15 years, the appetites of modern video cards have increased markedly.
Accordingly, all VRM elements occupy almost half of the entire PCB area. And the cooling system has been transformed from air to hybrid.
For modern video cards, VRM can be located both to the right of the GPU and to the left. If the number of power phases of the video card cannot be placed on one of the sides, then the VRM elements can be located on both sides of the GPU, as is the case with the above-mentioned video card from ASUS.
The VRM is responsible for regulating and converting the video card’s power supply, and currents with a force of 50-60 amperes pass through it. Accordingly, the longer the video card is busy rendering 3D polygons, the higher the heating of the power circuits and all VRM components, as well as the video processor and video memory. MOSFETs generate a lot of heat during operation, so they need passive cooling in the form of radiators at least, and even better, the video card cooling system additionally blows over the VRM zone.
How to Check VRM Overheating?
So, heating and overheating of the VRM system of the video card can become a serious problem and the result of the unstable operation of the video card and even the cause of its failure.
To determine what VRM temperature your video card has, you should use diagnostic utilities:
Aida64, GPU-Z, MSI Afterburner, HWMonitor, SiSoftware Sandra and others. Typically, the GPU VRM1, GPU VRM2 and so on sensors are responsible for displaying the temperature of the VRM of the video card, depending on how many sensors are provided by the manufacturer of your video card.
VRM temperatures in GPU-Z:
If, when fully loaded in 3D, the VRM temperature is in the range of 50-70 ° C, then you can not worry – this temperature is considered normal. Although the lower the better naturally. But, if the temperature readings step over the mark of 85 ° C and above, then it is worth considering and starting to take some measures.
Although here it is necessary to make a small digression. The maximum allowable operating temperature of VRM power elements can be 100 and even in some cases 110 ° C, but sound logic suggests that a video card will be able to work at such temperatures, but its service life will be significantly reduced. Therefore, you should not focus on such figures, even if they are considered conditionally acceptable.
How to reduce VRM overheating Temperature?
And now how to solve problems if, according to monitoring data, the VRM temperature of your video card shows numbers over 85 ° C.
The first thing to do is to clean the video card cooling system from dust. It is possible that dust has clogged the space between the heatsink fins and is interfering with airflow or blowing through the VRM heatsink, depending on the design of the VRM cooling system.
Then you need to disassemble the cooling system and check the fit of the CO heatsink of the video card and the VRM components of the video card the thermal paste may have dried up or the thermal pads used have expired. In this case, both thermal interfaces will need to be replaced. When choosing thermal pads, pay attention to the thermal conductivity parameter, measured in W / (m K). The more watts the thermal pad you purchase, the better.
If the two simple methods described above did not help, you need to resort to programmatic methods to solve the problem. If the video card is overclocked, you need to remove overclocking. If the VRM temperature has not changed or you have not overclocked the video card, you can lower the supply voltage to the GPU and reset the frequencies a little, or reduce the Power Limit (PL) value. As a result of these actions, the power consumption of the video card will decrease and temperatures will decrease. There are also several options for how to do this. The first is using the MSI Afterburner utility, the second is to change this setting in the BIOS of the video card, which will require a BIOS flashing procedure.
If these three methods did not help, then two cardinal ones remain. Let’s start with the first – this is the replacement of the video card cooling system. There are different types and types of cooling, from purchasing a full-cover water cooling for a video card to using special solutions or improvised means. Nothing prevents you from fixing a fan under the video card and adjusting its airflow to the VRM zone of the video card. Everything will depend on the design features of your video card and your budget.
The last option might even be to replace the case. If you are using an old case, or it is not ventilated enough, and due to the high temperature inside the case, the video card and all system components heat up, then you should consider purchasing a new, well-ventilated case.
Conclusion
You will definitely like this solution to the issue of VRM overheating since a new case from the day of purchase will delight you every day.
The user should closely monitor the temperature indicators of all components of modern computers. Even if you got a new computer or bought a new graphics card, install one of the diagnostic utilities and run through it. All operating temperatures of all system components must be within reasonable limits.
Often, users pay attention only to the temperature of the CPU and graphics core of the video card, forgetting about the VRM of the video card. Now it should be clear how to monitor the temperature readings and what to do if the VRM temperature of the video card is close to critical. It is better to diagnose the problem at an early stage than to contact the service later for repairs.
