By Fred Langa
Overheating in PCs can cause unexpected hangs and shutdowns — and even shorten the life of your computer.
Fortunately, it’s easy to monitor your system’s temperature and to correct the most common causes of overheating.
Heat is the inevitable byproduct of all electronic operations. Have you ever wondered at the size of that heat sink sitting on top of a desktop system’s much smaller CPU?
Heat is also the scourge of all electronics hardware. Mild overheating will shorten the life of a system’s components; excessive overheating can cause a PC to cook itself to death. At the very least, a PC that’s running too hot can have erratic behavior, data errors, spontaneous reboots, and other intermittent — and often baffling — problems.
Sometimes, the first and only sign of serious heat-related trouble is a sudden failure such as the one described by reader Richard Thornton:
- “I have a laptop with Windows 7 Home Premium [installed]. Recently, while I was watching a streamed Netflix movie, everything froze. It was a panic-initiating experience. I had to use a forced shutdown.
“I noticed that the power AC adapter was very hot. Upon restart, CHKDSK ran. It found and fixed some damaged files. Afterwards, the computer seemed to function properly, but I haven’t tried to watch a Netflix movie yet.
“I’ve tried many support venues, but so far nobody can answer why the computer froze.”
This sounds like a classic case of overheating. If the ability of a PC (or laptop, or netbook, or whatever) to cool itself is partially compromised, the system may work fine under normal load but suffer heat stroke when asked to work hard. Streaming a high–bit rate, movie-quality video feed is indeed a hardware-intensive task.
What can cause a cooling system to fail? Sometimes it’s a dead fan, but most often it’s simple grunge — dust and dirt — that slowly builds up in a system’s air passages and prevents proper cooling.
Obviously, it’s better to find out about thermal problems before you experience freezes, hangs, data loss — or premature system death!
That’s what the rest of this article is about: How to tell whether your PC is being properly cooled — and what to do if it isn’t!
Is your personal computer headed for a meltdown?
Most motherboards, CPUs, and hard drives have temperature sensors built in. Oddly, most operating systems largely ignore these sensors. But with the right software, you can tap into your PC’s built-in sensors to tell exactly how hot it is inside the case.
My favorite temperature-related tool is SpeedFan, a free, multipurpose program that can monitor your system’s temperatures, fan speeds, and internal voltages. If your hard drives are S.M.A.R.T.-compliant (definition) — and most are — SpeedFan also can show you your hard-drive temperatures.
With additional configuration, SpeedFan can also let you control your system’s fans, adjust the CPU clock speed, and more. These are advanced, expert-level features that should be approached with caution, if at all. But the basic temperature readings require no special configuration and are safe for anyone — novice to expert.
SpeedFan’s download page makes the actual download link somewhat hard to find. On that page, the download-initiating link is the phrase SpeedFan 4.44 in the paragraph labeled Download.
Running on one of my laptops (see Figure 1), SpeedFan shows that the temperature of the hard drive (HD0) is a relatively cool 32°C and dropping (the blue down arrow). The app also indicates that the motherboard (Temp1) and the two CPU cores (Core 0, Core 1) temperatures are all within safe temperatures and steady (green check marks).
Figure 1. SpeedFan can show that the current temperature of hard drives, CPUs, the motherboard, and other components are within safe ranges.
SpeedFan also lets you set alarms to provide early warning of imminent overheating, so you can take corrective action before damage occurs. The software runs on all current Windows versions, and it supports a wide range of common hardware sensors and motherboard types.
If SpeedFan doesn’t work on your system or if you’d prefer something different, you can choose from plenty of other temperature-monitoring software. I’ve used the following to good effect. Except as noted, they’re free downloads from the listed sites.
- MobileMeter: Monitors CPU temperature, CPU clock speed, battery charge/discharge rate, and HDD temperature.
- Core Temp: Monitors CPU temperature. The software is excellent, but it’s distributed via a very aggressive co-marketing installer package. Read the download dialog boxes carefully to decline software you don’t want.
- Intel Active Monitor: Monitors CPU temperature, motherboard temperature, voltage and fan speed. Compatible with most newer Intel motherboards. See also the related Intel Desktop Utilities (free; site).
- Hmonitor: Monitors voltage, CPU temperature, motherboard temperature, and fan RPM. (14 days free, then U.S. $25 and up).
What temperatures are OK — what’s too hot?
Different systems and components are designed for different temperatures. For example, many laptops and portable devices are built to run at much higher temperatures than desktop systems.
Most monitoring software will try to identify your system type so it can accurately interpret the temperatures a PC generates. But this tends to be a broad-brush, approximate approach.
For greater precision, you can look up the optimal operating temperatures for your system on the maker’s website. For example, Intel says my laptop’s CPU chip has a maximum safe operating temperature of 100°C. With that information in hand, I can more accurately confirm that the machine is not overheating.
All the major system vendors and component makers publish similar data, often in technical sections of online product spec sheets.
As a shortcut, some third-party sites aggregate temperature information for you. For example, a Panther Products page lists “CPU maximum temperatures” for a number of common CPU types. If yours is a listed type, you can save yourself some digging on the vendor’s pages. (If you don’t know your CPU type, a freeware tool such as CPU-Z [site] can help.)
To be certain, do a thermal stress test
After you have temperature-monitoring software installed and you know what your system’s safe thermal values are, you can perform a controlled stress test to make sure your cooling system is working correctly.
The concept is simple: start your thermal monitoring software and then run your PC’s CPU at 100-percent load for a set time (say, 10 minutes). Watch what happens to the temperature readings.
In a healthy system, the temperatures will climb rapidly for a minute or two, then stabilize safely below the allowable maximum temperature.
In a poorly cooled system, the temperatures will keep climbing until they approach — or even threaten to surpass — the maximum allowable temperatures. If you see this happening, abort the test, shut down your PC, and attend to the PC’s cooling system. We’ll cover this in a moment.
The easiest, most reliable way to run a CPU at full power for a set time is with specialty testing software such as the free OverClock Checking Tool (download site) or Prime95 (free; site). Both come with complete instructions.
You can try commercial options, too — such as PassMark’s BurnIn Test ($39 and up; site).
If your system runs hot or really overheats
As I stated at the top, the most common cause of heat trouble is dust and dirt clogging a PC’s fan, heat exchanger, or other critical internal air passages. Fortunately, it’s easy to clean most PC desktop and laptop cooling systems.
Desktops are typically easy to open up and clean with swabs, cleaning cloths, and judiciously used compressed air. See my Feb. 28, 2005, Langa Letter story for a step-by-step guide.
You can clean the cooling system of a typical laptop in less than 10 minutes, as I’ll show here in five easy steps:
- 1. With the laptop’s files backed up and the system completely powered off, place the laptop upside down on a soft surface (such as an old towel) in a well-lighted location. Remove the laptop’s battery, as shown in Figure 2, and locate the fan. It’s usually near where warm air normally exits the laptop. In this example, my laptop’s fan is located in the lower-right corner of the machine.
Figure 2. Start by removing the battery.
- 2. Carefully examine the area around the fan to determine the least amount of disassembly necessary to gain access to the fan. In this case, just four screws hold a plastic cover in place over the fan and nearby components. (See Figure 3.)
Figure 3. After removing the access-cover screws, gently pry it off.
- 3. With the cover off, the fan and heat exchanger (in the lower-right corner of the laptop) are clearly visible, as you see in Figure 4. The details vary from system to system, but the basic cooling components — some kind of fan, heat exchanger, and vents — are fairly universal.
Figure 4. Examine the notebook’s cooling system, noting how it’s laid out.
Unscrew, unclip, or otherwise release the fan from its mount. Unplug its electrical connector. Lift the fan out of the laptop. Gently clean the grunge from the fan blades with a cotton swab (as shown in Figure 5), a barely damp cloth, or short bursts of compressed air.
Figure 5. You can use swabs to carefully clean the fan.
- 4. Clean the vents and heat exchanger the same way. As you see in Figure 6, a barely damp cloth works well.
Figure 6. Clean the vents and heat exchanger.
- 5. Carefully reassemble the parts, and you’re done!
Want more detail on cleaning? See the article I wrote a long time ago, “Curing laptop overheating.” It uses a now-obsolete system for an example, but the cleaning principles remain the same.
With your PC’s fan, vents, and heat exchanger now clean and unobstructed, your system should now be able to stay cool, even when running at full speed!
| Feedback welcome: Have a question or comment about this story? Post your thoughts, praise, or constructive criticisms in the WS Columns forum.|