| My mother always told me to dust, but I never | | | | idle and when working. The tests are all run |
| did, mostly because I was lazy, but also | | | | are on the following system:CPU - Dual AMD |
| because I couldn't find any tangible benefit | | | | Athlon MP 2000+ (1.67 GHz) |
| to dusting. I just didn't see how I'd benefit | | | | |
| from my room or my things being less dusty. | | | | Motherboard - AMD K7-D |
| Well now I've gotten a bit older, and I | | | | |
| finally found a reason to dust - a cooler | | | | RAM - 1024 MB RAM (2 x 512 MB registered DDR |
| running computer. I'm going to give some | | | | 2100) |
| background on my own system and circumstances | | | | |
| first, then run tests before and after | | | | Video Card - Matrox Parhelia AGP 128 MB |
| dusting, as well as explain how and what I | | | | |
| used to dust. Be sure to take a look at this | | | | HD - Segate 5400 rpm- st320410a |
| article - with pictures and graphs - plus a | | | | |
| whole lot more, at I've had the computer in | | | | Windows XP SP2I picked that unexciting hard |
| question for a little over two years, a dual | | | | drive because it was the only one I had that |
| AMD Athlon MP workstation that while no | | | | reports temperature.For the idle readings the |
| longer the top of the line, is still plenty | | | | computer was freshly booted into Windows. To |
| fast enough for what I do. The computer is | | | | get the computer running at full load I ran |
| running at standard speeds and | | | | two instances of Prime95, a math application |
| specifications, and has always been very | | | | that will max out a CPU (2 copies running, |
| stable - but not 100%. When the system was | | | | one each for 2 CPUs), and copied 2 GB of |
| about a year old I had been getting by with | | | | Music on the hard drive to another folder on |
| the occasional, roughly once weekly lock up. | | | | the same drive. The entire process took about |
| At that point, I finally spent the time to | | | | half an hour.Without anything to compare them |
| try and diagnose the problem.To be perfectly | | | | to, those numbers for the most part aren't |
| clear, I'm talking about a lock up, where | | | | terribly interesting. My only reaction was |
| everything stops responding, the screen | | | | that 63 degrees Celsius seems pretty hot, and |
| freezes and I have to reboot, not simply an | | | | the idle CPU temperatures aren't too |
| application crash, which I can usually just | | | | wonderful either. Remember, each CPU type has |
| blame on Microsoft. My initial thought was | | | | a different recommended temperature range. 63 |
| that the computer was overheating, | | | | degrees may be too hot for mine, but could be |
| specifically the CPUs. I was a little | | | | either acceptable, or perhaps way too hot for |
| hesitant though because I was running AMD | | | | your own. Regardless, you hopefully will |
| retail processors at standard specifications | | | | notice a relative reduction in temperature |
| with AMD retail heatsinks and fans, and I | | | | after dusting.Dusting the computerAs I said, |
| figured that should have been a fine setup. | | | | I've never dusted a computer before, but I |
| But I've had CPUs overheat before when I was | | | | came up with what ended up being a reasonable |
| sure that was the problem and this just felt | | | | plan. I bought compressed air and a small |
| like it now. I did some research online and | | | | brush from staples for $7 total and used a |
| it looked like the AMD cooling solutions were | | | | rag I have here. I made sure to unplug my |
| somewhat underwhelming performers, so I broke | | | | computer, grounded myself to discharge static |
| down and bought new heatsinks and fans. These | | | | electricity by touching something metal other |
| still weren't top of the line, but they | | | | than my computer, opened up the case and was |
| reduced my CPU temperatures immediately by | | | | ready to go.As long as you make sure your |
| about 20%.In unscientific testing I'd say my | | | | computer is unplugged, don't get anything |
| computer was absolutely more stable after the | | | | wet, and are gentle while you have it opened, |
| reduction in temperature. I estimate the | | | | you really shouldn't damage anything. For the |
| weekly lockup became a monthly or even every | | | | most part, the inside of your computer is |
| other monthly lockup. This clearly isn't | | | | delicate, but it isn't brittle. The most |
| perfect for a system that really should be | | | | notable exception is your CPU. If you were to |
| 99.99% stable, but it was a big improvement, | | | | disconnect it and pull it out of the |
| and I let the problem go for a while. I will | | | | motherboard you would expose pins on the |
| note that as many of you many assume, this | | | | bottom that are in fact extremely |
| computer is always on, twenty-four hours a | | | | delicate.You could do a really thorough job |
| day, seven days a week.Anyway now it's | | | | by taking everything apart and getting all |
| another year later and my computer is | | | | the dust out, but I didn't want to spend that |
| increasingly unstable again. I'm not going to | | | | much time doing it and I figured if I left |
| go out and get better heatsinks and fans | | | | everything connected their was less chance |
| again, as I'm sure the improvement would be | | | | I'd mess something up. So I used the air, |
| less than before. My next though was about | | | | brush and rag and got the dust off the |
| how dusty the whole system is. I know I | | | | computer, taking the most time to clean the |
| should have dusted it once in the last two | | | | fans and the holes they blow out of.I ran the |
| years, but I never got around to it. I'd say | | | | same tests again, and the results were |
| I live in an average environment in terms of | | | | dramatic.Idle, CPU 1 saw an 18% reduction in |
| dustiness, not especially better or worse, | | | | temperature, while CPU 2 is 10% cooler. |
| and I just never thought it would make a very | | | | That's very important, as all those hours my |
| significant difference in my CPU temperature. | | | | computer is sitting doing very little it's |
| As you'll see, I was completely wrong - which | | | | going to be considerably cooler.At full load, |
| incidentally might make my mother | | | | CPU 1 is 21% cooler and CPU 2 is 12% cooler. |
| right.Almost all users should really consider | | | | These are again very impressive results, and |
| the results of both tests, possibly giving | | | | very important as well. I have had lock ups |
| more weight to the one which most closely | | | | when video encoding and doing other CPU |
| matches your typical computing. Even if you | | | | intensive tasks, and now the chips should be |
| run predominantly business applications, | | | | running cooler while doing those.It's |
| you'll almost certainly occasionally do | | | | interesting that originally CPU 1 was hotter |
| something that falls under this content | | | | than CPU 2, and after the cleaning they |
| creation test, editing pictures or an | | | | switched. An important fact is that the |
| occasional home movie, for example. So | | | | temperature reporting on this type of dual |
| consider all the tests, don't just focus on | | | | AMD Athlon system is generally uneven. That |
| one graph.Should you dust your computer? Yes, | | | | said, CPU 1 was easier to dust around because |
| why not, it can't hurt. But really, there are | | | | the DVD drives were in the way of CPU 2, so I |
| tangible benefits of cleaning your computer, | | | | probably did a better job cleaning the former |
| even if it seems stable right now.Computers | | | | and there is still likely room for |
| and electronics in general don't like heat. | | | | improvement in cooling the latter.One other |
| Dust blocks fans in your case, which | | | | note, I kept track of the fan speeds before |
| generally cool you CPU, video card and | | | | and after cleaning, and I saw about a 2% |
| motherboard components. Dust also blocks fans | | | | increase in the CPU fan speeds after I |
| and their airflow into and out of your case. | | | | dusted. I wasn't expecting this at all, and |
| Cool air needs to be brought into a case, and | | | | certainly can't be certain that this will |
| then the host air dispelled. If the airways | | | | always result from dusting, but it could be |
| are blocked, system temperature can rise | | | | very interesting for overclockers trying to |
| quickly. If your computer is stable but the | | | | eak out every last bit of fan performance.The |
| CPU is running too hot, you cut down on its | | | | hard drive temperature saw a 9% improvement |
| lifespan, potentially quickly.More important | | | | while idle after dusting, and again a 9% |
| to many people though, may be the result of | | | | improvement at full load. It's a nice |
| that first heat related computer lock up. | | | | temperature reduction for a mechanical |
| Even if it's never been a problem before, if | | | | component that will be grinding away for |
| your computer crashes at the wrong time it | | | | thousands of hours over the life of your |
| can be catastrophic. Usually mine just locks | | | | computer.The case temperature stayed the same |
| up when I'm away from it, or overnight, and I | | | | throughout all the tests.ConclusionAll in all |
| just turn it back on and restart Firefox and | | | | I was surprised the results were so dramatic. |
| haven't lost anything important. But last | | | | I expected few degrees for the CPUs and maybe |
| week it locked up with unsaved graphs for my | | | | one for the hard drive, but to see a 20% |
| last article and Excel chose no to auto save. | | | | reduction in temperature is just great. There |
| I spent the hour it took to redo them | | | | is defiantly room for improvement as well. |
| considering ways to eliminate these lock | | | | The CPU fans have covers that I could have |
| ups.Of course, reducing heat is also always a | | | | taken off and gotten more dust out, and as I |
| priority for people who overclock their CPU. | | | | mentioned reaching one of the CPUs was harder |
| For those that don't know, overclocking is | | | | than the other. Further, I'd say I got most |
| running a CPU at a higher frequency than it | | | | of the dust inside my computer, well over |
| was sold to run at. For example, you could | | | | 90%, but I still saw some that proved more |
| take your Intel Pentium 4 that is running at | | | | challenging to clean that I was willing to |
| a "clock" rate of 2 GHz, and try to run it at | | | | bother with. Taking all the cards out and |
| 2.1 GHz, 2.5 GHz, faster speeds, or anywhere | | | | thoroughly cleaning the whole system would |
| in between. I have an old dual CPU system | | | | have no doubt led to even better results.My |
| that was supposed to run at 366 MHz. Instead | | | | system has maintained the cooler temperatures |
| I ran the chips at 500 MHz each, which was a | | | | so far, a week after the cleaning, and I |
| huge performance gain. Overclocking is | | | | haven't had a lock up since. I can't be |
| actually a great way to get more "free" | | | | certain I've solved the lock up problem, but |
| performance out of a system, as long as you | | | | it does appear I have for now. At the very |
| can maintain stability. Usually the single | | | | least, I can be sure CPU heat is not the |
| biggest factor for success is reducing heat | | | | problem, as now they are running at very |
| as much as possible.Another thing to note is | | | | acceptable temperatures. Your own results |
| that while it is very important to keep CPU | | | | will vary, and may not be as dramatic if your |
| heat to a minimum, hard drives, video cards, | | | | computer wasn't as dusty as mine. Regardless, |
| and other components all need to be kept cool | | | | you should see some reduction in component |
| as well. In fact, I don't really know for | | | | temperature after cleaning your system.All in |
| sure that my CPUs are the current problem. I | | | | all I recommend you take a look around and |
| think they are, but my next guess (if I'm | | | | inside your computer to see how it looks. If |
| correct that it's a heat problem) would be my | | | | it's dusty, $7 and an hour of your time seems |
| video card, since I've checked, and it runs | | | | like a good investment to increase the life |
| really hot.Consider this as well, if my CPU | | | | of your computer and prevent system lock ups. |
| were to actually stop working because it was | | | | I suspect many people suffer the occasional |
| too hot, it would probably be a gradual | | | | heat caused lock up, it just isn't obvious |
| process, and I could fix the situation by | | | | what the problem is, and they don't occur |
| purchasing a replacement. If my hard drive | | | | often enough to be a serious concern. But |
| crashes and ultimately loses data, that could | | | | remember, even if you only have one heat |
| be a much more problematic situation. I could | | | | caused lock up in the entire life of your |
| replace the drive, but recovering the data | | | | computer, it could come at the worst time, or |
| could be far more difficult than just | | | | cause permanent damage.Steve Perlow is the |
| replacing a CPU.I opened up the system and | | | | founder of aworldofhelp.com, where you can |
| saw more dust than computer. All the fans | | | | find the aworldofhelp Top Picks in desktop |
| were covered in dust, and their airflow was | | | | and notebook systems. Visit to get answers |
| totally blocked. I put the case back on and | | | | from real people to your questions about |
| took temperature readings of my computer both | | | | technology, travel and more. |