Why does the central processor on a computer get hot? Why does my computer overheat?

Hello to everyone who came to the light. Summer has come, a wonderful time, wonderful weather, warm and hot. This is the time when you need to periodically clean your computer from dust, check the temperature of the processor (CPU) and other PC components. It’s just that in the summer, a huge amount of dust accumulates in the computer case, which can be a source of overheating, and overheating will cause failures and other troubles in the operation of the PC. In today's post we will talk about how to measure the temperature, what it should be and what to do if overheating is detected.

Reasons for overheating of the processor and other components

Before we start looking at temperatures, we should list the causes of overheating. And symptoms that indicate your PC components are overheating.

Poor cooling system, or rather not designed properly.
Dust that gets clogged into the cooler's radiator, obstructing air flow. In some particularly neglected cases, the thickness of the dust can reach several tens of millimeters.
Poor contact between the surface of the heatsink and the CPU. Drying the thermal conductive paste.

Symptoms of overheating of the processor and other components

The computer turns off spontaneously(). During a critical temperature, the protection mechanism is triggered and the PC reboots.
The computer freezes and slows down during games, graphics distortion ( video card processor overheating).
Periodic braking. This can happen if the processor starts throttling— skip cycle mode to relieve the load.
If, after turning on the PC power, a message appears like: ***Warning: Your computer CPU Fan Fail or speed too low*** You can also disable this warning message in SETUP, then this is a problem with the cooler, which means the processor is overheating. I once had a wire get tangled in a fan and wouldn’t spin at all. Good percent. didn't burn

It is worth mentioning here that brakes can occur not only from overheating, but the culprit for this can be: viruses, file garbage, a huge number of programs in the tray ( registered in startup), file fragmentation.

How to measure the temperature of a video card and processor

For temperature measurements There are a huge number of utilities. I will offer you HWMonitor. You can download it here. It's in English, but that won't be a hindrance.

This program shows the current (Value), minimum (Min) and maximum (Max) temperatures of PC components since startup. It is worth paying attention to the maximum indicators .

In order to find out the maximum temperature, you need to launch some powerful application or game and look at the readings in HWMonitor.

Important: The HWMonitor utility should be minimized when working with a powerful program or game. If you start it later, then there may no longer be accurate readings, because The CPU drops temperature very quickly.

I launched HWMonitor on different computers and on each of them the components were displayed differently. In the screenshot above, everything is basically clear - this is my work PC. If you have any difficulties determining the temperature, write in the comments and I will help you in any way I can.

Standards for indicators

For the processor, the norm is 35-45 degrees when idle. Above 60, throttling begins—the mode of skipping beats. The CPU specifically skips clock cycles in order to reduce the load and then the temperature. At this moment the slowdowns begin. Above 80 degrees, a spontaneous reboot or shutdown may occur.
The norm is a temperature of 70-90 degrees under load, although if the video card is old, then this temperature can cause problems. In general, you need to observe and if signs of overheating begin to appear, then you should think about how to reduce the temperature.
For a hard drive, I consider the temperature to be up to 50-55 as normal. Over 60, it is worth thinking about where to copy important data, otherwise it can be lost.

So what to do if overheating is detected? Here you can give some practical advice. Now I will simply list them, and we will consider some of them in detail in the following notes.

Today our topic is processor overheating. We have already touched on it a little in our article on processor burning due to overheating. Now let's look at this point in more detail.

Firstly, I want to say: if it heats up, this is normal, since electrical voltage is applied to it. But if it overheats, this is no longer good and must be dealt with.

Secondly: modern chips (from " AMD" or " Intel" - not so important) as well as, have a built-in system for preventing overheating and forcing shutdown if it is detected. So processor overheating and failure is not as important now as it used to be.

How can you monitor the processor temperature? First of all, these are “bios” indicators. Depending on its model and implementation, the parameter we need may be located in different sections of the BIOS. Most often it is located in the “Hardware” or “Power” sections and is called “ Hardware monitor"(monitoring).

In the figure above we see the processor temperature (CPU Temperature) and motherboard temperature (MB Temperature). Temperatures are presented in degrees Celsius (C) and Fahrenheit (F).

Here, for example, is what an external thermal sensor looked like for taking temperature readings from old AMD processors (now similar sensors are built into the core itself):

To measure temperature, there are also various system utilities, of which there are a great many, but you need to understand that all programs use the readings of hardware sensor chips located on the computer motherboard.

Previously, monitoring functions were performed by specialized thermal and electrical sensors. Now special microcircuits do this “ Super Multi IO" They are also called multicontrollers (or “cartoons”), since they not only remove and process various indicators from tracking sensors, but also control fan speed, implement the functions of parallel and serial ports, contain mouse and keyboard controllers, FDD, game port, etc. We looked at how to repair the board by replacing the multicontroller in our article.

We can see an example of one such “cartoon” in the photo below. As you can see, the multicontroller is built on the basis of the Winbond W83627THF chip.

Note: The normal operating temperature of the processor should be between 30 before 60 degrees Celsius (depending on its model), chipset - from 25 before 50 , and the graphics core (video card) - from 40 before 70 degrees. Naturally, depending on the load on a particular component!

And now I want to give one example from practice when the BIOS temperature readings helped me detect processor overheating.

At one of my previous jobs, I was replacing a burned-out motherboard on a computer. Naturally, I removed the processor with the cooling system and installed a new one. Assembled, screwed together, launched - everything works. About a week later, they ask me to come and take a look: the computer is making a lot of noise, which hinders the self-realization of the accountant working on it :)

I arrive and from the threshold I hear the howl of a fan. At first, I admit, I thought (this happens to them), but it turned out that the cooling cooler was noisy (the processor was installed for the LGA 775 socket).

I decided to look at the temperature indicators of the motherboard and core in the BIOS, and was quite surprised that the sensor showed “85” degrees Celsius. Now, at least, it was clear why the fan was working at maximum speed (the motherboard or the “stone” itself, having detected overheating, increased its rotation speed).

Another thing is noteworthy here - under the conditions of a week's operation in this temperature regime, no consequences of such overheating were found, leading, as a rule, to reboots and freezing.

By the way, the reason that overheating occurred was that I did not fully press one of the clamps of the cooling system. As a result, the radiator did not fit tightly to the protective cover of the processor and could not effectively “take” heat from it. After fixing the problem, the temperature readings were fixed at “59” degrees.

It must be said that signs of processor overheating on new computers may not appear even if you forcibly turn off the active cooling system (pull the fan power out of the motherboard). If you do not run fairly “heavy” applications (computer games), such systems can function stably for a long time.

I remember we had another case in our IT department: they brought in an old Athlon with symptoms of periodic “freezing” and unstable operation of the software. They asked me to reinstall Windows.

Intuitively understanding that the reason was not Windows, we opened the case and found that the fins of the radiator of the processor cooling system were so densely clogged with caked dust that it formed a kind of “shield”, which, in fact, isolated the radiator from the fan, which was unsuccessfully trying to dissipate the emitted the core is warm.

I had to remove the entire structure, clean the radiator from dust with a screwdriver, apply new thermal paste and put everything back together. After this, the computer worked stably and the question of reinstalling Windows automatically disappeared :)

What do I want you to understand? That attention to overheating of the processor (like any other PC component) should be given due attention.

How to avoid CPU overheating? When installing (or replacing) a radiator, be sure to use thermal paste. Carefully remove the old one and apply a new one to its top protective cover in an even thin layer, as shown in the photo below:

You can evenly distribute the paste over the surface with your finger (as shown in the photo). Personally, I use an old credit card that doesn't work. Its hard plastic allows for uniform distribution of thermal paste over the surface.

I use paste from Zalman. It is sold in special bottles and equipped with a brush for more convenient application.

Why do you need thermal paste at all? She plays a very important role in preventing processor overheating! To effectively remove heat from the hot core, the lower surface of the radiator must fit very tightly to its protective cover (in older models, directly to the processor chip).

The fasteners provide the proper degree of fit, but the problem is that no matter how good the fasteners are and no matter how well the bottom surface of the radiator itself is polished, there are still micro-scratches and micro-gaps between the processor and its cooling system.

These same “gaps” are what high-quality thermal paste is designed to fill. Since it is thermally conductive, it significantly increases the overall cooling efficiency by collecting heat from the entire surface and transferring it to the radiator, from which the heat is blown away by a fan.

There are also funny cases: one PC came to us for repair/maintenance. The diagnosis is overheating, the consequence is spontaneous shutdown (overheating protection is triggered). We blew it out and decided to change the thermal paste on the processor. Imagine my surprise when, after removing the old paste underneath, I discovered this picture:

The computer came to us from China (along with a set of measuring equipment), but why weren’t the stickers removed when it was carried out?! In addition to everything else, a transparent polyethylene blotch was carefully glued to the oval! What do you think is the thermal conductivity coefficient of such “nice” additions? :) The processor heated up to 76 degrees Celsius!

With the help of a stationery knife, alcohol and such and such a mother, they removed this disgrace, applied new thermal paste and the heating temperature, in the end, was 61 degrees.

Personally, to eliminate the problem of overheating, I used a tower-type radiator in my home computer. Copper (unlike aluminum) conducts heat better, and the overall design of “tower” cooling systems makes it much more effective to combat the problem of processor overheating.

This is what my Scythe model looks like Katana3».

I would also like to draw your attention to the fact that good (high-quality) fans have three or even four wires for connection. Additional wires allow you to programmatically control the speed of their rotations and adjust the pitch of this rotation.

Here, for example, is what the packaging of an 8-centimeter LED fan I recently purchased looks like:

Let's go through the basic notations:

RPM- "Rounds Per Minute" number of revolutions per minute
Voltage- supply voltage of the device
Current- current consumption (in Amperes)
Air Flow- air flow created by the cooler
Noise- noise produced

All the main parameters related to temperature, rotation speed and voltage can be measured using the wonderful program “ Speed Fan", which I offer you. This program receives readings from a multicontroller (we talked about it above), which, in turn, takes data from various counter and sensor chips located on the motherboard.

I can’t resist, and I’ll dwell separately on the products of the Zalman company, I really like them! :) In general, Zalman in computer cooling systems is the same as Intel in the production of processors (a respected and very respectable brand).

Here, for example, is what their packaged cooling looks like, designed to prevent processor overheating:

The delivery set includes a mounting kit for different boards. Moreover, the fastener is so universal that it allows you to install the structure on a whole range of processor sockets (connectors). It is noteworthy that both Intel and AMD CPUs are supported!

Here are additional accessories for the cooling system itself:

A little more detail:

(1 and 2) - fastening elements
3 - “backplate” (The lower part of the cooling mount. Located on the back side of the motherboard)
4 - fan speed controller on the processor (displayed on the front or rear panel of the case) Scientific name - “ rheobass»
5 - power supply cables

Look at the radiator itself. This is the thing I’m holding! :)

Massive and at the same time elegant design:

So, don’t skimp on the cooling system: overheating the processor has never brought anyone any harm. And yours will look simply gorgeous after installing such a “toy”! :)

Let's say a few words about the general (classical) organization of the cooling system inside. It can be described as follows: a fan is placed on the front wall of the case (under the front decorative panel), running for air injection inside the system unit. Another fan is mounted on the rear wall of the case, but this one works at blowing air out.

In a closed case, such a combination provides quite significant traction. Inside, it’s as if a large turbine begins to work: cool air from the room gets inside and, passing through the entire “stuffing” of the computer, is thrown out.

There is another (non-traditional) cooling scheme: many fans are installed in a closed system unit, all of which work to blow air inside. During operation, they create excess pressure inside the housing, under the influence of which hot air is literally “squeezed out” from numerous holes in the casing.

And for a better understanding of the importance of cooling and preventing overheating, I once again present a photo of mine and hope that the recommendations we outlined in this article will help you avoid such an unpleasant phenomenon as processor overheating.

And finally, watch a video about how computer processors are assembled:

Hello, I built a PC two years ago, the budget was small - 400 dollars, and for a gaming PC this is not enough, so I decided to take a risk and buy a used motherboard and processor. g41MLX3 and xeon e5450. Everything else is new: gtx 950, goodram 4 GB 1333mhz ×2 pcs, deep cool 300 cooler, chieftec smart 750wats power supply and a 300 GB hard drive from an old PC Sata 2, assembled everything, started up, installed the OS from scratch. All games ran with a 1400×900 ultra monitor and high fps of at least 45, I played 15 hours a day all winter and some micro-freeze glitches started, I started checking everything with different tests, I sinned on hhd but everything is fine, I installed a new OS and it didn’t help, and I took it to the service center as it was a pity, they said the memory bar was dead, I bought a new one, came home but nothing changed, about every 20 minutes it freezes for 30 seconds and online games every 10 and disconnected from the server, so I suffered and burned out I didn’t turn it on for 6 months, I bought a new 25-inch fullhd monitor. I turned it on and oh gods, it worked for a week without glitches, and again the hhd light froze and froze as always, I bought a new terabyte P300 new OS for half a day and again glitches and more often, I found a swollen capacitor on the motherboard, changed all the capacitors at once, nothing helped . I noticed that with the Internet cable disconnected, it hangs much less often, I took it to a very good service, they tested it and said that everything was fine (they tested it separately), but the problem did not go away. Then I started researching myself, bought a new sata cable, cleaned everything, changed the paste (PC on an open bench), formatted everything, installed a proven old Windows 7-2009, did a stress test of the processor for 1 hour without trolling, max temperature 61 degrees with Aftoburner and Aida, the video card showed 66 degrees under stress also for a whole hour everything was fine, four mark, Aida, aftoburner, ran the memory 5 times, everything was fine, the disk was checked by Victoria, the condition is excellent, the read speed is excellent, the power supply was checked under load, the stress test showed 11.9, 5.1, 3.3, the slightest deviations according to the table the condition is excellent, all this for an hour, after these tests there was a blue screen of death and often writes that Windows cannot be found, then it says no hhd, I installed an old sata cable and nothing helped. In one hdd program it was shown that the connection to the host was lost 46754 times. As a result, according to these tests, everything is in excellent condition, we launch Word of tanks with crashes, we enter the hangar in battle every 15 seconds, a glitch for 30 seconds and a server break, and so on on all 10 servers, not playable, on different tanks, from the observation of another PC the tank drives straight on its own until they kill you, and the depiction of the tanks and locations as some kind of square boxes with a 6-sided wheel barrel, the buildings are soapy, the sound and crash disappear, every other time it doesn’t matter whether the settings are ultra or minimal. The disk light freezes or goes out for a long time and you can hear it gaining high acceleration and starting to respond. The new Doom glitches every 20 seconds, and many others, otherwise the PC works fine except for games, it also slows down when watching movies, on YouTube 2k runs fine without glitches, how is that? It’s a pity to throw it away on ultra WOT, Far cry 4 Gta 5, and Battle 4, and all Call of duty, before I played 50-60 frames on this hardware, but now I can’t even complete the mission. Dear experts, help me solve the problem.

Each computer component performs its specific role. But perhaps the processor can be called the “brain” of a computer. It is with its help that information that is in the memory of a computer and other devices is processed. And if any problems arise with the processor, this will affect the entire computer as a whole. Very often there is a problem with the processor overheating.

What to do if your computer’s processor starts to get very hot or even overheat? First, let's try to understand the reasons for processor overheating.

If you go back a little, you can remember that due to overheating, the processor simply burned out. Today, many computers have a built-in overheating warning system on the processor. That is, if the processor starts to overheat, this . Therefore, the most common accompanying processor overheating is the spontaneous shutdown of the computer.

It’s worth noting right away that the processor always heats up. This is logical, since voltage is applied to it. But the operating temperature of the processor should not be confused with its overheating. can range from 35 to 55 degrees. If the temperature is higher than the extreme limit of this interval, then the processor is overheating.

Moreover, if the processor overheats, the computer will not always turn off on its own. Very often at such moments he begins to slow down. This will immediately become noticeable, as the data processing speed will decrease significantly.

So, if your processor gets very hot, and the computer spontaneously reboots, then the overheating problem is most likely related to the computer’s cooling system. In this case, you need to perform a number of actions.

First of all, in this situation you should. Dust accumulated over a long period of time can be one of the reasons for overheating of the processor. After this, you need to inspect the cooling system. In most cases, the problem will be a lack of thermal paste. It's also possible that the cooler cable may have become disconnected and it simply isn't working. It is likely that there is a problem with the cooler itself.

If the above problems are detected, they need to be eliminated. If you don’t have thermal paste, you need to apply it to the necessary areas of the processor. Here we should not forget that thermal paste is applied in a thin, even layer. If the problem is in the cable, then simply connect it to the cooler. Well, if the cooler in the cooling system does not work, then it is best to purchase a new one.

By the way, there may be such a situation: the processor overheats, but everything is fine with the cooling system. That is, everything is cleaned of dust, thermal paste is applied, the cooler is working. In this case, the fact is that your computer’s cooler is not doing its job. In other words, it is simply not capable of cooling your computer's processor. If this is your situation, then buy a more powerful cooler. Luckily they are not that expensive. You can also install an additional cooler on your desktop computer.

You need to monitor the temperature of your processor regularly. You don't have to do this every day, but once a month is worth it. To do this, you can use the Speccy utility, which can be downloaded from this site. It displays all the data on your computer, among which you can find information about your processor, including its temperature.

It can be seen that there are not so many reasons for processor overheating. And all these reasons can be eliminated yourself.

The situation is as follows. I was ready for an upgrade, took advantage of a promotion at a computer store, and bought a lot of things in installments. Including:
motherboard ASUS Socket-AM3+ SABERTOOTH 990FX
processor AMD FX-8150 3.6GHz 16Mb DDR3-1866 Socket-AM3+ OEM
video card GigaByte GeForce GTX 670 2048MB 256bit GDDR5
Left with the old configuration:
Cooler IceHammer IH-4500 (Al+Cu, copper base, 1000-2000rpm 15-25dBA)
memory 2 * Patriot PC3-12800 / DDR-III / 4Gb / 1600MHz
The new configuration is assembled and launched. This is not the first time I’ve done this, so everything happened quite quickly and relatively professionally. Laying wiring in a housing is an art, so with each attempt I become more and more skilled. It would seem that it’s time to bang a bottle of champagne on the system unit, fortunately the case is made of good material and will withstand a blow. But it was not there. I have a habit (and rightly so) of running a temperature test every time I assemble a PC. It doesn’t matter where - the BIOS or an already installed system. Measurement time, noise from fans, launch of “heavy” applications and temperature sensor readings can tell a lot about our assembly. And once again I am convinced how bad cheap ready-made assemblies from stores are in this regard.
But let's not get distracted. In general, the temperature of my processor is somehow suspiciously strange:
http://litl-admin.ru/wp-content/uploads/2012/07/03da71e5e4cc3fb7c80e2f19da81e6fc.jpg
For older generation AMD processors, temperatures of 60-70 degrees are normal. But we are in the age of nanotechnology, so such temperatures are already considered enormous. In the screenshot of the HWMonitor64bit program, you see that the processor temperature has reached 47. It would seem like nothing, but this is practically a simple computer. Without thinking twice, I ran the AIDA64 processor stress test. Unfortunately, a photograph of that process was not preserved, but take my word for it - in 20 minutes the temperature rose and continued to rise after 57 degrees. At the same time, in the manual fan speed control on the IceHammer IH-4500, I set the speed a little higher: from 950 to 1250 revolutions. I didn’t want to make more revolutions, because... noise appears. My processor is not overclocked, so I won’t increase the rotation speed as a matter of principle. Based on the description of my processor on the manufacturer’s website, it is clear that the critical temperature is +61 degrees. Then I stopped the test and started thinking and writing on forums.

Thinking about a solution

So, there is a processor that, by definition, should not get very hot. There is a cooler that is not a high-end product, but still has 5 heat pipes, a huge radiator with a quiet fan. Several options arise:
1. Bad thermal paste. I don’t remember the model, but it seems that I installed it on a simple CoolerMaster, which was included with some old cooler. I have always neglected thermal paste, believing that it does not play a big role. The one that comes with the kit will be enough. In most cases this is true. But in my case there was none, so I smeared it with what I had on hand. We will change the thermal paste.
2. The cooler itself was poorly secured to the motherboard. It may very well be that later you will see that this was the case. Let's check the fastening.
3. Weak cooler. This is unlikely, but I'm ready to buy a new one, especially since IceHammer will be useful to me for the old system. We'll buy and install a new one.

Given

Remember chemistry lessons? So we are now writing the conditions of the problem. There is a system unit:
http://litl-admin.ru/wp-content/uploads/2012/07/7faf655e9cc7ca7e1fe108158836e344.JPG
The cooling in the case is more or less adjusted. Turntables are located everywhere: 1 piece at the back for 120 mm, 2 pieces at 92 mm at the back on the side, 1 piece at 92 mm at the front on the side wall:
http://litl-admin.ru/wp-content/uploads/2012/07/06736fef88efec5c21f5e42a71981e93.JPG
... which connects to a special adapter to reduce the speed:
http://litl-admin.ru/wp-content/uploads/2012/07/24039c09ce8b9aae4b1881952fb13e03.JPG
Our task is to replace the cooler. Now everything looks like this:
http://litl-admin.ru/wp-content/uploads/2012/07/4a80030cc02dbd9e7ce95a65795ad291.JPG
All good and powerful coolers are always installed by completely unwinding the motherboard. The original fasteners that are on the motherboard are unscrewed and the ones included with the cooler are installed. To do all this, you need to remove the motherboard from the case. To do this, we free it from the “load”: we take out the pieces of iron and various cables.
http://litl-admin.ru/wp-content/uploads/2012/07/80edfcda461f53be1f49ded53dd1fb64.JPG
After I removed the entire structure, I begin to look. We check the fastenings, the second point of my guesses. From below everything seems to be normal. Red gaskets create a sealing effect. Everything seems to be fine for now:
http://litl-admin.ru/wp-content/uploads/2012/07/f6a0f4e274c68f7b6e8e21511eda372f.JPG
I look at the front view. I don't see any gaps. Everything seems to be fine too:
http://litl-admin.ru/wp-content/uploads/2012/07/d59e69bfc83b3b0fd8139fcdfa7cf80c.JPG
But on the other hand, this view appears:
http://litl-admin.ru/wp-content/uploads/2012/07/4618801a979b56b32f8f1259ce3e14ec.JPG
http://litl-admin.ru/wp-content/uploads/2012/07/1699630239be12845a5cabfe1dbf1683.JPG
This is where the dog is buried. Now the big picture emerges: I made a mistake when installing a new platform. Namely, he did not check the tightness of the structure. Those. The platform on which the cooler stands was slightly higher on one side than on the other. As a result, a microgap appeared, which affected the temperature of the processor. I'm removing the entire structure.

Cooler replacement

Since I decided to change the cooler, I ask. There are many offers. Almost everyone is eliminated due to the fact that many coolers are simply not available. Finding the optimal cooler is hard work. The budget has been cut a bit. As a result of a private conversation with, I set my sights on Deepcool Frostwin.
http://litl-admin.ru/wp-content/uploads/2012/07/57c1957d1982f297ef4b56a84dfb033f.JPG
An ordinary box, nothing special. Inside there is thermal paste (I couldn’t read the markings) and a set of various fasteners, including AM2/AM3. There are also instructions.
http://litl-admin.ru/wp-content/uploads/2012/07/c0b411536ef17dfd125e490517ddee58.JPG
As we can see, there are only 4 heat pipes on the cooler versus 5 on the former IceHammer. Well, let's rely on a new radiator system, including 2 quiet coolers.
http://litl-admin.ru/wp-content/uploads/2012/07/a03730711e935ee6f7cae887722317dc.JPG
Next is a matter of technology. We assemble everything according to the instructions. By the way, it’s easier to assemble than the old one. Much is provided for user convenience. I applied the thermal paste that came with the kit. Although it was necessary to play it safe and buy a better one. But I’m not going to overclock it in the near future, so it should be more than enough. This procedure did not take much time. We get at the output: