By Ctein
Something that really bugs me about the design of personal computers that circulate air internally for cooling is that none of them seem to include any kind of filtering on the airflow entering the machine. It's a basic design flaw that contributes to a significant amount of computer malfunction.
About a year back I wrote "Your Scanner Needs Cleaning!" Here I am, back with a related refrain: clean your computer!
Here's the problem. That circulating air carries dust. The dust settles out on the electrical components inside your computer. Some of those components are insulators, some conductors. Doesn't really matter. They're all designed to operate in open air, not under a layer of dust.
The dust creates problems for two reasons. The first is the dust doesn't have the electrical characteristics of air. It affects electrical signals that pass through it differently. Your computer is just one big radio frequency generator. It doesn't take a lot to distort all those high-frequency signals moving around on those circuit traces. There's a certain safety margin built into the design so that some signal distortion and degradation can be tolerated. There's not a lot. A layer of dust can distort the signals enough that occasionally they will be misread at the other end of the circuit. Programs will abort for no apparent reason, files will get saved with occasional small bits of corruption, and if things get extreme your computer will start crashing unexpectedly...well, more than it does usually.
Basically, the machines aren't designed to run under dust anymore than they are under water. If anyone doubts that and is brave enough to try it, fill up a tank with distilled water, stick your computer in it, and tell us how well it works. Distilled water is an excellent insulator, but you'll find it isn't all that great at supporting computer signals.
The narcoleptic CPU
What's the second
reason? Well, that's why am writing this column. I have my MacBook Pro
opened up pretty frequently, installing upgrades and new components such
as the hybrid hard drive last winter. I always blow out the
accumulation of dust when I do that. Apparently I did not do as good a
job as I thought I did last winter. A month or so ago, my MacBook Pro
started shutting itself off spontaneously. It would simply go into sleep
mode with no warning. I wasn't losing any data, but computer narcolepsy
is really annoying. It shuts down. I wake it up. It shuts down. I wake
it up. Repeat until finally it stays awake...and I'm ready to throw it
across the room. (Remember my motto: never use a computer too big to
throw out a window.)
After a little experimentation I deduced that it had to be a heat buildup problem. The thermal protection built into the system was intentionally putting it to sleep. That's why I wasn't losing any data, being an entirely controlled shutdown so far as the computer was concerned. It just didn't accord with my wishes.
I took the machine apart and used the better part of an entire can of compressed gas to blast every nook and cranny I could get at, paying special attention to getting the little nozzle underneath circuit boards and in corners where I might normally not. (And being very careful not to spray liquid, just gas, into the computer!)
Violating nonexistent warranties
While I
was at it I pulled the DVD burner. It's the third one I've had in this
machine and it had moved far beyond being merely flaky. It not only
usually failed to reliably burn new discs, it often failed to read
previously-burnt ones. This happens regularly on my computers. DVD
burners are not trustworthy beasts. That's why I always double-verify my
disks.
The system was out of warranty, so I had nothing to lose. I took a half dozen screws out of the cover on the drive, sliced through the little stickers that told me how I was violating my nonexistent warranty, and popped off the cover. There was a thin layer of dust inside the drive, not a lot, probably not enough to matter. I was able to get rid of some of it with the remainder of my can of compressed air.
Then I took a PEC Pad, wet a corner of it with optical cleaner, and gently wiped it over the lens in the read/write head in the drive. The lens looked slightly more transparent after I did that. A good sign.
Popped the cover back on the drive, put everything back together. Total time consumed, just over an hour.
It's been two months now, and my laptop hasn't shut down spontaneously once. The internal cooling fans are also cycling off more frequently.
And, wonder of wonders, my DVD burner seems to work perfectly and be completely reliable again. I've burned a good dozen discs with it, they all read perfectly fine on it and other drives I own, and the drive didn't balk once the whole time. Wouldn't have done that two months ago.
Just regained some significant functionality at very little expense. I am happy.
Too bad my iMac isn't user serviceable; its DVD drive has been getting crankier than the one in the MacBook Pro. It'll be back to the Apple Store for a warranty exchange when I can find the spare time.
Friggin' ^%#@!@$*!! unfiltered circulating air.
Ctein
Ctein's regular weekly column on TOP appears on Wednesdays.
Featured Comment by Paul Ewins: "I used to work for a bank doing IT support work and often had to deal with quite filthy PCs. The worst were a set of branch servers that had been kept under desks. There were dust-bunnies inside that would have shamed the worst bachelor housekeeper. I would take them down to the loading dock, pop the sides off and blow everything out with a household vacuum cleaner. We used to joke that PCs were designed to draw dirty air in through one side, filter it carefully over all of the components then expel clean air out the other side."
Featured Comment by Dave Stewart: "I regularly fix desk/portable PCs for folk. Even if they're not having any hardware issues, I make a point of cleaning their machines out, as most are wary of looking inside. It can also be a useful exercise to educate the owners. Shoving the box on the floor wasn't a good idea—dust tends to collect down there, and for one owner, so did lots of pet hairs. Looked like half a cat was inside before it blew up!
"A vacuum cleaner and an anti-static brush are useful, though go easy—make sure the computer power is switched off at the wall, and best to avoid super powerful Dyson and similar vacuums!
"Laptops/notebooks—remove the power sources, i.e. power supply and the battery, prior to working on them. Dust sometimes forms a felt-like barrier across some of the cooling fins at extracts, and getting to such areas can be 'challenging' —not for the faint-hearted. It may be possible to suck or blow some of this out from outside, but again, go easy.
"There can be numerous fans inside that will clog up, and/or seize. Be wary of poking carelessly at them as you can affect their balance. Check they're working okay after any cleaning. Do not put on any lubricant, as it will just gum up with dust.
"Kids laptops are entertaining...gloves may be required. Sticky (as in yuck!) keys, along with all manner of crumbs and other foreign objects are the norm.
"Optical drives—as a first line of defence, there have been cleaning disks available for years. These have with tiny brushes that may resolve some dust problems. Careful how you store them—don't flatten the brushes. Failing that, the old cotton bud with some optical cleaner, as outlined by Ctein above. And note the 'gently' bit; otherwise the alignment may be affected. Optical cleaner doesn't remove jam; water's best for that!"
"The first is the dust doesn't have the electrical characteristics of air. It affects electrical signals that pass through it differently."
As an RF Engineer by trade, I am going to throw a penalty flag here. The traces in your typical motherboard are not controlled-impedance lines - like microstrip or coplanar waveguide. I doubt that they use an EM simulator to make sure that the dispersion is under control. As such I doubt that a little layer of dust is going to make much difference. I have done temperature tests on RF circuits and found that ice on the circuit does not change things all that much, and I am sure that the ice has a higher dielectric constant than dust.
As far as filtered air goes, the mfgs are in a no-win situation. Use a filter and since it probably will not get changed very often, at some point airflow is restricted so much that you get the same thermal issues.
I guess a *clean* hot motherboard is better than a *dusty* hot motherboard, but not by much.
Posted by: KeithB | Wednesday, 20 July 2011 at 02:20 PM
I'm not trying to get into a Mac vs. PC debate, and I know this isn't for everyone (not everyone has the time nor interest), but if you home-build it's not particularly hard these days to find PC cases that include filters over their intakes. Back in the bad old days there was also plenty of lore out there about how to DIY filters for your intakes as well (IIRC, one of the favorite materials for this task was a dryer sheet).
Posted by: Peter | Wednesday, 20 July 2011 at 02:23 PM
Ummm, don't want to open my computer now, but aren't the traces on the motherboard and elsewhere covered with a layer of lacquer? Or whatever it is. So they don't oxidise, yes, but that should prevent the dust settling straight on them, too.
The heat buildup is a problem, though. The worst of it happens when dust enters the cooling fan bearings. You know, when you hear the grinding and squealing, when the fans become really audible. For instance, the fan on my CPU is slightly eccentric. :) It wasn't like that when I bought it. It went off centre after the dust buildup.
Posted by: erlik | Wednesday, 20 July 2011 at 02:26 PM
smcFanControl is a nice little utility that lets you manage fan speed and monitor cpu temps on your portable Macs.
http://www.eidac.de/
Posted by: Matt | Wednesday, 20 July 2011 at 02:39 PM
Dear Keith,
Hopefully somebody who designs computer circuitry today will chime in on this, but I'm going to, very tentatively, disagree with you. I say very tentatively because my information is really 20 years old, but at that time the circuit traces had to be evaluated as coupled pairs/triplets/whatever. The impedance between them mattered. There was a really nice article in one of my computer journals, early 90s, that showed what the actual signals looked like when they got injected into the board and when they left the board, there was substantial degradation along the way, even in a well-designed board.
Maybe I'm misunderstanding the import of that; it's been a hell of a long time since I did any RF stuff, and it was never my favorite thing to do. So if you tell me that that is completely uninfluenced by the impedance of the surrounding medium and I misunderstood the import of what I've just described, I will believe you.
As for filtering, computers have long been designed with thermal protection (viz., the incident I just described). A clogged dust filter will cause a reduction in air flow and a rise in temperature, but both of those are easily monitored and the computer can throw up a warning that says, “Change your filter!”
~~~~~~
Dear Peter,
I would kind of agree with you, except it's really an open-box versus a closed-box thing. You can pretty easily hack any open box desktop computer, Mac Pro or PC, preassembled or DIY, with some filtering over the air intakes and a heavier duty fan to overcome the additional resistance. (Yeah, and how often did I do that. One of those "Do as I say…" things). Not easy with closed boxes or laptops.
~~~~~~
Dear erlik,
Yes, there's a protective overcoat to avoid things like oxidation. And it acts as an insulating layer. But impedance is about capacitance and inductance as well as resistance. Those operate through insulators. Note, though, the possible disagreement between Keith and I. It may be that I am wrong and that the impedance of the surrounding medium doesn't matter at all. Hopefully enlightenment will arrive soon, on the wings of our supremely well-educated readers.
pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
======================================
Posted by: ctein | Wednesday, 20 July 2011 at 03:13 PM
Airflow design in PCs in general (other than some rack-mount machine-room systems) is terrible. They cool better with the case open -- a sure sign of poor airflow design. (It does make debugging hardware easier, though; debugging router hardware at Network Systems required some fairly hefty external cooling expedients to get them to run with the case open and not melt.)
I agree that a blocked intake filter would eventually cause overheating (and remember that early PCs didn't have heat sensors and shutdown software; in the early days it could have meant actual damage and data loss), but so will dust accumulation. These days, with temp sensors and shutdown, it's unforgivable to not filter the intake air. We get thick mats of dust even in the modern clean office space at work, in relatively few months. At home, with the cats, it's far worse. Generally speaking, maintaining positive air pressure in the case, and filtering the input, is what I want. (That's what I wanted in a darkroom, too, but never built one that fancy. Still, the idea of dust filtering OUT of my darkroom at every little crack was very appealing.)
Posted by: David Dyer-Bennet | Wednesday, 20 July 2011 at 03:20 PM
As Peter says many after market cases have filters. My Coolmaster Cosmos case even has a separate filter for the power supply intake. They're easy to remove and washable so I just pull them once a month and run them under the tap. I know that in the past it was possible to "re-case" Macs but I'm not sure about the newer machines.
Posted by: B Small | Wednesday, 20 July 2011 at 03:38 PM
Ctein- you described my problem with my 9 year old iMac it goes to sleep on it's own- and the fan runs noisy. When I brought it in for repair- the logic board was replaced.
It still does the sleep thing- but less often.
Fred
Posted by: Fred Haynes | Wednesday, 20 July 2011 at 04:10 PM
Ctein...
Thanks for the tip.
With best regards,
Stephen
Posted by: Stephen S. Mack | Wednesday, 20 July 2011 at 04:37 PM
My pet hate is the bad design of SATA hard disk drive connecting cables, (desktop machine). I have on more than one occasion starting having HDD troubles with corrupt files and Windows writing umpteen copies of the same files, each one with an added numeric extention. A first I didn't know what was going on but got a clue from some www inquiries. After throwing out some (probably) good HDD drives for new ones, (and I buy the more expensive ones anyway), I finally discovered it was the ropey SATA cable connections giving trouble. They have no grip and if you moved the HDD cage in the sliders they inevitably pull off or pull skew. David.
Posted by: David | Wednesday, 20 July 2011 at 04:49 PM
I think there are a couple of issues here at least:
On a related note: I'd much rather have a machine which actually has some error checking in it: such a machine would not suffer "occasional small bits of corruption" because it would know it had suffered an error (in almost all cases). Doing that is not rocket science.
Posted by: Tim Bradshaw | Wednesday, 20 July 2011 at 05:02 PM
The good news is new macs won't be coming with DVD drives or any moving parts soon (ala the new mac mini), and the case will be welded shut, so you won't have to worry about dust bunnies anymore. I will finally be able to get some work done.
Posted by: yunfat | Wednesday, 20 July 2011 at 05:09 PM
I have an air filtration system that I use in my work room. I used to have it in my darkroom to help keep dust away from my negatives. I find that not only do I breath better with it, but both my printer and computer are free of dust related problems. Mine is pricey but over the years has been worth the investment. You might look into it.
Posted by: Richard | Wednesday, 20 July 2011 at 06:22 PM
I used to use cans of compressed gas to clean film, but, I found that the "air" was leaving an oily residue on the film, so I switched to a camel hair brush. I would check to make sure the gas you are using doesn't leave a residue, which might collect dust, or have even worse electrical properties than the dust.
Posted by: David Duane | Wednesday, 20 July 2011 at 06:23 PM
When my current desktop PC was still almost new, and I had installed more HDs, so there were four, I used HDTune to monitor the drive temps. Three of them were regularly running over 55C.
A few moments of looking at the cooling setup led me to believe that a vent in the side panel next to the bottom and backplane would simply let air go up past the top of the add-in boards and out the case fan.
Simply covering that set of holes lowered interior case temp significantly, presumably by forcing air to travel past the mother board and HDs on it's way through the computer.
A cooling fan mounted in a backplane slot and one each on the current two HDs brought temps down to something reasonable. The primary drive still runs hotter than I think is ideal, but well below the 55C danger point.
You are absolutely right about regular cleaning. Dust gets in the poorly sealed bearing of the backplane fan, so it needs clean and lubes. The nice fan that pulls outside air directly through a tube from a vent blows dust into the CPU heat sink, slowly clogging it up.
I use a vacuum, though, directly through a small brush adapter for the heavy stuff and to catch much of the dust blown from less accessible places. Doesn't make much sense to me to just put it out in the air that I and the computer will be breathing.
Moose
Posted by: Moose | Wednesday, 20 July 2011 at 06:50 PM
And you REALLY think people would clean the filter?
After 25 years as a computer hardware tech I can tell you the answer is NO!
My laptop has a small filter at the air intake. You have to remove 8 very small screws from the underside access to the area where the filter is. You think the average user would know that?
Please. There's no filter for the reason no one would know how or care to clean it.
Posted by: John Krill | Wednesday, 20 July 2011 at 07:30 PM
I can't emphasise enough how important this is. I live an Shanghai which is one of the most polluted cities around.
Recently I noticed that my Macbook would sometimes die under load and that even at idle, the cpu temp was between 80ºC and 90ºC. Initially I thought it was the extra load driving my newly acquired 27" Cinema display causing the extra heat.
However I stripped it down and removed all of the accumulated dust from the fan exhaust. The results were immediate and impressive. Now at idle, the cpu temp sits between 55ºC and 65ºC
Even under load, it rarely tops 85ºC, its previous idle temperature.
Posted by: FrasSmith | Wednesday, 20 July 2011 at 09:04 PM
@Tim Bradshaw: Lenovo makes fans for their laptops that are lower noise because the impeller (or expeller, what do I know from blades?) are modeled after owl feathers. You can read about it here: http://www.lenovoblogs.com/insidethebox/2007/05/what-do-owls-and-thinkpads-have-in-common/
Lenovo Thinkpads are true successors to IBM's Thinkpad heritage, and about as good a portable computer hardware as is available. My W510 workstation has 95% NTSC gamut, full HD resolution, and built in Huey color calibrator.
Within 2 years, most notebooks will probably be passively cooled, so no fan noise possible. And SSDs will be more practical in large capacities.
Patrick
Posted by: Patrick Perez | Wednesday, 20 July 2011 at 10:56 PM
Dear Moose,
Oh yeah, forgot to mention that part. I haul the machine I'm cleaning outside and blow out all the dust there. No reason to make more housecleaning for myself.
I just tell myself I'm "recycling" the dust back into the environment.
pax / Ctein
Posted by: ctein | Wednesday, 20 July 2011 at 11:19 PM
Careful! Vacuuming your computer can generate enough static electricity to damage it. My MacBook Pro luckily seems to accumulate only fairly little dust, so every half year or so, when I'm having it apart anyway, I just blow on the fans and heatsinks, more or less. (I also disagree on the dust = signal degradation theory, but I'll let the experts settle that)
Posted by: Friedrich | Wednesday, 20 July 2011 at 11:52 PM
I once had the RAID memory go poof on a Dell PE 1U rack server. 2 weeks worth of downtime. Fortunately no data was lost. It was working in a dusty environment (textile factory) and naturally it was full of dust inside. Probably full of static electricity as well. This beast has some four or six 2" turbo fans inside that spin so fast that the whole thing sounds like an airplane taking off. Two of them blow the air and dust directly into the CPU radiators. So everything is scattered and deposits inside.
Posted by: ggl | Thursday, 21 July 2011 at 01:29 AM
On the RF issue: motherboards (and ICs) are indeed developed with RF problems in mind, but if I recall correctly this mostly involves keeping extra space and/or groundlines around traces that carry HF signals. So in most systems, the only effect of replacing air with dust (thus changing the dielectic constant) will be (slightly) altered transmission times, but since this should have roughly the same effect on both the data and the clock lines, I doubt this would ever exceed the margin of error build into the design. But like you, it's been a while, and never my favourite subject, so no absolutes here either.
As for the submerged computer, I humbly submit this :-)
However, the heat problem is even nastier than you think. Dust is a very good insulator, so a thin layer of dust on a component will heat it up regardless of temperature or airflow in the rest of the case. And a 10 or 20 degree rise in temperature can have a LOT of effect on the margins in an IC.
And on the subject of optical drives:
I remember a case at a previous employer, where we were designing a new logic chip for a client, a big-name electronics company I won't name, to be used in a CD burner. Their old one had a problem where the heat of the IC and the burning laser combined could deform the lens in the head assembly. No amount of cleaning would have fixed that!
Fact is, home burned optical discs are a bit of a kludge, and optical drives are one of those technologies where making them cheaper has definitely made them worse. I have four different ones in my workstation at home, and there are still discs I can't read.
And of course, of the discs I've burned over the past ten years an alarming number are unreadable. I've stopped trusting (and using) the technology for backups, and have switched completely to server-based backups.
Posted by: Bernard Scharp | Thursday, 21 July 2011 at 01:54 AM
I found out that a computer accumulates less dust if you have the covers of. The fan then blows the dust away with the air and the cooling situation improves also. Only once year I use the VC generously when I also remove all components and leads and reïnstall them after cleaning. Laptops are a different ballpark. But I have not had to carry a PC anywhere (in my free will, my former work required laptops) so I do not do laptops anymore. Carrying data around is much less strain on my back.
Greetings, Ed
Posted by: Ed | Thursday, 21 July 2011 at 02:22 AM
Ctein's suggestion is altogether very sensible. I still make an internal cleaning exercise a yearly-ish event, but like Ctein, if you play around with the insides of your box, perhaps just doing it when you upgrade is enough.
One thing to note is that CPU temperature is closely linked to CPU usage (and frequency, for modern CPUs which can change speed). So in order to diagnose any thermal problems, knowing the load on your box is a good idea if you are trying benchmark before and after.
One final thing: CPU overheating is not generally dangerous to your equipment now. There was a time when only Intel put in thermal self-protection into their CPUs (by throttling), but if you let an old AMD get into thermal runaway, you'd end up with a small burnt crater in your motherboard. So in other words, not cleaning would at worse slow down your machine a couple of notches, and at very worse halt your machine. If you haven't noticed any weird performance problems (i.e. in OS X the spinning beach ball, or in Win7 unusual lockups with the timer, including on mouse movement), then don't stress. Cleaning out the insides of your machine will result in incremental gains at most.
Finally, there are the following apps that help you monitor the temperature of your machine:
OS X:
http://www.bresink.com/osx/TemperatureMonitor.html
Win 7:
http://www.alcpu.com/CoreTemp/
Pak
Posted by: Pak-Ming Wan | Thursday, 21 July 2011 at 02:43 AM
If you will happily open up a MacBook, don't be scared of the inside of an iMac. All you need are some suction cups to lift the front glass off (held on by magnets!!!!) and away you go.
Posted by: Chris Crowe | Thursday, 21 July 2011 at 05:45 AM
First thing I do when building a new PC is to junk all the fans and replace them with quiet, high quality ones - same for the CPU fan and cooler. Good quality air filters as well.
Can't stand noisy computers.
4 CPU cores running at 29 to 31C, hard disks running at 35 and 38C. Doesn't go up very much when it's stressed either.
Much easier now that one can use SSDs instead of SCSI disks for the OS.
If I was a Mac user, it'd have to be a Hackintosh, simply to get some control over cooling and filtration.
Posted by: Hugh | Thursday, 21 July 2011 at 07:24 AM
Actually there's a long tradition in the IT world of honouring dirty kit, for example:
http://www.theregister.co.uk/2009/11/13/ventblockers/
and:
http://xkcd.com/237/
John
Posted by: John Howell | Thursday, 21 July 2011 at 07:55 AM
I have found that my Giotto rocket blower which I bought for my camera is a wonderful tool for cleaning dust out of my computer. Since it's not compressed air it doesn't spew tiny bits of moisture; since it's not electronic I don't have to worry about static (as much); since it's small and manual I can be targeted and purposeful with every gust of air I produce with it.
Posted by: Christian | Thursday, 21 July 2011 at 09:04 AM
Laptops sadly don't have a lot of room for filters. Now on desktops I think it is inexcusable that all modern desktop cases aren't equipped with easily removable, user servicable air filters. That won't stop users from choking their machine (I've inspected more than my fair share of whole house HVAC units that looked like the air filter hadn't been replaced in at least a year, if not several).
Part of the reason why I build all my computers myself. That way I can get the exact case features I want, such as air filtration, number, placement, air flow and sound volume of fans I want (at least as it applies to the case, the exact hardware in the case for cheaper is also pretty nice perk).
Posted by: Matt | Thursday, 21 July 2011 at 10:22 AM
http://ars.userfriendly.org/cartoons/?id=19971203
Posted by: Andreas Weber | Thursday, 21 July 2011 at 11:12 AM
In 1998 I came into possession of a complete DECstation 3100 system (I only wanted the monitor) which had spent the previous 8 years or so in one of my former university's computer suites. Other than occasional power outages they'd been running continually for the entire time. Most of the computer suites were in old, dusty buildings.
I opened the case to have a look and there was a carpet of accumulated dust, hair and God only knows what else covering every square inch of the mainboard. It must have been an inch thick on average. Never did cause a problem. It may explain, though, why our 1985-era Macs in the same room had developed a tendency to catch fire by the time I was using them in the mid-1990s. No fans, but some of that same mess must have been finding its way through vents in the case.
Be thankful that your Mac was smart enough to perform a controlled shutdown. My wife's Toshiba laptop has a thermal cutoff which just shuts the power off immediately without any attempt to put the computer into hibernation, standby or a clean shutdown. Plenty of lost work and some corrupted files before I took it apart and removed the mess clogging the entire heatsink.
Posted by: Paul Glover | Thursday, 21 July 2011 at 12:54 PM
Dear John K,
When the computer pops up a little warning box that says "air filter needs cleaning," most people will figure it out.
~~~~~~~~
Dear Chris,
REALLY???
Can you point me to an instructional guide on this?
I had knowledgeable people, who know what I can do with (to?) the innards of a computer, inform me that a late 2009 27" iMac was not gonna be something I could/should pry into.
If that's wrong info, I'd love to know.
~~~~~~~~
Dear Bernard,
OK, so maybe it WILL run under distilled water. Noticed the impedance issue came up in the comments thread over there, with no clear answer.
My habit on CDs and DVDs from Day One has been to both burn and verify as a matter of course, and then pop the disk into a different computer's drive and see if Photoshop there could open all the files.
So far, after something over a terabyte of disk burns, haven't lost a single file due to a bad DVD.
But, yeah, it's a hinky tech. Unfortunately still the most convenient one for me for backing up and moving offsite work product, the kind that gets generated a few GB at a time.
pax / Ctein
Posted by: ctein | Thursday, 21 July 2011 at 02:41 PM
A comment on the inductance/capacitance issue. If you read the article Bernard linked, in step 8 it talks about how they had to seal the CPU base so oil would not get in there and change the capacitance.
http://www.tomshardware.com/reviews/strip-fans,1203-8.html
Posted by: Greg Roberts | Thursday, 21 July 2011 at 03:39 PM
I used to work in a place with dozens of computers in clean rooms. It was a joy to open up a 10 year old computer and find absolutely no dust inside it at all, unlike the office computers that were filthy after just a few months.
Posted by: David Anderson | Thursday, 21 July 2011 at 05:02 PM
@Ctein, I don't know exactly when (IIRC it has only been in the past year) but Apple has been locking their hard drives to the platform such that only Apple provided drives will run on their PCs. I've only heard about this affecting iMacs (and this isn't anything like encrypting, there wouldn't be any problem moving an Apple provided HD from one Mac to another). This only applies to internal drives.
Aside from the accessible memory slots, there wouldn't be much reason to ever go into an iMac.
Patrick
Posted by: Patrick Perez | Thursday, 21 July 2011 at 07:46 PM
Regarding the the submerged capacitance: It has been a while since I read that article. I didn't remember them mentioning capacitance, my mistake.
Conclusion: Cleaning helps, in many different ways :-)
Regarding optical drives: My problem hasn't been so much with reading directly after writing, but reading them again a few months/years later. The longevity of home burned discs is really bad. Using gold discs and such will help, but that'll draw the cost/GB well back in to flash-drive territory.
Wouldn't a bunch of multi-GB thumbdrives or flashcards be just as efficient for offsite backups? Less error prone, higher data-density and in many cases, higher transfer rates.
(I can't speak from experience, my offsite backups are done by a cron job over the internet, but I probably generate less GBs a day than you do.)
Regards,
Bernard
Posted by: Bernard Scharp | Friday, 22 July 2011 at 12:57 AM
Dear Patrick,
Errr, did you read my column before posting? The reason why I'd have reason to get into my iMac after it's out of warranty is clearly stated.
pax / Ctein
Posted by: ctein | Friday, 22 July 2011 at 01:48 AM
Hah! :D Something brand new:
The heatsink is the fan.
"...immune to the buildup of dust..."
Posted by: erlik | Friday, 22 July 2011 at 11:11 AM
Ctein,
Yes, I did see the specific instant reason, I was just illuminating preemptively (I've recently gone back to doing crossword puzzles). The iMacs, though not impossible to 'work on' (in the maintenance/repair sense) are not for the 21st century equivalent of a shade-tree mechanic.
Since I'm on the soapbox, I'd also like to educate many readers about the warranties on their gear. A prevalent belief is that opening a device (as to perform upgrades/repairs one's self) is a violation of warranty, a belief frequently bolstered by prominent stickers over access screws proclaiming that. In the U.S. at least, that warning isn't sufficient to void an in-effect warranty. The Magnuson/Moss Warranty act puts the burden of proof on the manufacturer that the customer's actions were the immediate cause of the failure. Upgrading ram would not be proof if the CPU failed (well, at least the manufacturer would have to prove it did).
Patrick
Posted by: Patrick Perez | Friday, 22 July 2011 at 08:53 PM
Dear Patrick,
Thanks for bringing up that point. It gets mentioned frequently, but it can't be repeated too often, 'cause I still hear from readers who believe any manufacturer-unauthorized work or entry voids your warranty.
My MacBook Pro keyboard died about a year ago while I was on vacation. I borrowed a friend's screwdriver kit, stripped the machine down in the hotel room, found some loose cables, cleaned and reseated them and, guess what?
It still didn't work.
So I made an appointment to take the machine to the local Apple Store the next morning. Described my problem to the Genius who said, "It does sound dead, but it might be simply a cable that's worked its way loose. If you can spare a half hour wandering the mall, we can check that out and maybe save you a return trip."
So, I explained what I'd already done and he said, "Thanks, you just saved me wasting 20 minutes of my time."
'Course, if he'd found a damaged keyboard cable, when he went to swap in a new cover, I'd have been in trouble. But so long as I hadn't screwed anything up, (I hadn't) he was happy.
Note: KIDS-- DON'T TRY THIS YOURSELVES WITHOUT ADULT SUPERVISION.
pax / Ctein
Posted by: ctein | Saturday, 23 July 2011 at 06:50 PM
This should get you into some of the fans
http://www.ifixit.com/Guide/Repair/Installing-iMac-Intel-27-Inch-Optical-Drive-Fan/1641/1
Haven't been brave enough to try it with my iMac yet though.
Posted by: Chris Crowe | Monday, 25 July 2011 at 06:02 AM
Dear Chris,
Wow. Thanks, I've bookmarked that, but...
If all that goes wrong (post-warranty) with the iMac is the optical drive, I'll just buy another external one, I think. It's not like I port the iMac places, like my laptop. And, man oh man, that disassembly makes stripping down a laptop look like nothing.
But, still, good to have it on file, just in case.
pax / Ctein
Posted by: ctein | Monday, 25 July 2011 at 12:46 PM
Dont forget it is Computer cleaning Month in September
Posted by: Durable UK | Wednesday, 27 July 2011 at 06:55 AM
FWIW in about 15yr of running PCs, only the i920-powered monster server I built a couple of years ago has suffered from dust directly. Specifically, it collects between the CPU heat-sink and its fan, blocking thermal conduction giving all manner of overheating problems that also go away when I wave the Dyson at it.
AFAICT my macbookpro is not so adversely affected; I hate to think what you're doing to yours.
Posted by: Tim | Monday, 01 August 2011 at 02:58 PM