Slashdot Mirror
Hard Drive Cooling for 10 Cents
David Tiberio writes "I've bought many hard drive cooling solutions over the years, sometimes spending $50 or more on drive cooling systems that were noisy and did little to cool down the drive. After much tinkering, I discovered a simple solution that cost me only 10 cents per drive... the 1/2 inch bracket. Mounts any 80mm fan to the belly of an internal hard drive."
The magnetic field produced by a small DC fan might not be enough to faze the platter. It takes a magnetic field of a certain strength to make a change in the data bits.
I have a few Compaq Xeon workstations that placed the drives transversely in front of the system power supply so cooling air can pass between the drives. I have yet to see a problem. It's designed to cool 15k RPM drives very quietly. The PSU fan itself is a slower 12cm fan, placed on the intake of the PSU, only a few cm away from the drive's edges. It's very quiet for a PC, and very impressively quiet for a system with a 15k RPM drive in it.
If you decide to go with this kind of setup, try to make sure you use a fan with low vibration (well balanced, low speed). The last thing you need with a hard drive is more vibration. The drive head is only flying a few hundred molecules above the drive surface.
It may not amount to much as the vibration needs to be of the right frequency to be really bad. But it is probably better to err on the side of caution with drive lifetimes already being as bad as they are.
I personally use a 120mm fan that is mounted on rubber pegs, perpendicular to the hard drives, but not mounted to the drives themselves. This way, less vibration is transferred to the drives.
Some 7500RPM or higher drives that aren't well engineered can overheat in small spaces.
I personally had a Western Digital 80gb harddrive overheat and cause errors in a normal midtower. (Several of my friends had the same problem with the same model)
Since then my addage is if it's 7500rpm or higher put some fans on it. Since that realization I've had no problems.
A boss of mine who used to work for Air Force Intel told me that the magnetic field used to de-gauss an HDD had to be about as strong as a car-lifting magnet. I seriously doubt that the field generated by an 80mm fan is even enough to penetrate the steel housing of the drive (maybe not even the circuit board, since it's bottom-mounted).
What if this weren't a hypothetical question?
Sure.
Do you have a fan handy? Any sort of window fan or table fan will do, or even one of those 80mm computer fans.
Power it up and aim the air at your face. You should feel a cooling effect, maybe even a lot of cooling if it's a strong fan. Move the fan away a little. Distance decreases the effect but it's probably still noticable, right?
Now turn the fan around. No effect at all at a distance. Move it closer. Still nothing. You're going to have to put your face right up next to the fan to feel anything and even then, it won't be very strong compared to the air coming out the other side.
Bigger fans do have more of a suction effect and it also depends on the design. Vacuum cleaners obviously have quite a bit of intake power, but they also have substantial blower exhaust output.
And what of CPU fans? CPU fans can get away with suction because they are sitting right on top of the heatsink. But, I have to say I have never owned such a CPU cooler. All of mine have blown air down upon the heatsink.
Sig for hire.
I have been lead to understand that drives don't typically fail because their platters got hot. They tend to fail because their circuit boards fail. You'll find advice all over the internet suggesting that you try to fix a broken drive by swapping the circuit board from another of the same model before purchasing expensive data reclamation.
In that context, this fellow's solution doesn't seem so irrelevant. Keeping the circuit board cool is likely to lengthen the life of the drive.
The other replies to your question are good.
... but then he went on...)
But there's more to the answer than they realize.
Let me start with a story...
At my first job about 10 years ago, I wound up helping out at the IT department of a cellular phone company (no, a real cell phone manufacturer, not a service provider). One of the first tasks I had was to replace the CPU cooling fans on a few Sun desktop workstations. At the time, these Sun systems were incredibly expensive - about $40K each if I remember correctly. That, coupled with the fact that this was my first job, made me really nervous. I'd had a little experience with computer cooling fans before, but more with just general electronics. So when I pulled the case off the first system and removed the old, dead CPU cooling fan, I focused 100% on making sure that I matched the polarity of the wires on new cooler with what the old cooler had been using.
And in being so nervous, and so focused on the polarity, I completely missed the obvious question: does the CPU cooling fan pull air upward, or does it blow air downwards? I just stood there next to the system, yelling at myself in my mind for having failed to take notice of the airflow direction.
Now today of course, I've had more experience than I could possibly want with cooling fans. But remember, back then in the early 1990's, A) I was young, and B) CPU cooling fans weren't so common.
Anyway, I immediately realized my problem. But not wanting to let on to the guy who worked in that cube that I wasn't sure how to reinstall the CPU cooling fan inside his $40K system, I thought about it and decided that it would be most efficient if the CPU cooler wasn't fighting the natural tendency of heat to rise. And given that the cooler included a channeled heat-sink, it seemed reasonable that a fan extracting air upwards would work best oriented that way. And hey, power supply fans exhaust heated air from the supply, not force air it into the supply, right? So I installed the fan facing upwards, put the cover back on the system, and moved on.
Imagine my surprise (and fear) when I arrived at the next system with a dead fan, and I realized that the fan was facing downwards. I was now sweating, as I was worried that the CPU in the first system was at risk of frying. That concern was further heightened as all the subsequent CPU cooling fans were also mounted pointing down. Needless to say, except for that first system, I mounted all the other CPU fans facing downwards.
But this still perplexed me. Why would the cooling fan be blowing downwards, fighting the natural tendency of heat to rise? (These were desktop systems -- the CPU wasn't going to wind up sideways or upside-down -- so the potential for changes in orientation didn't occur to me.) At that inexperienced age, and back then when CPU cooling fans were so rare that I'd had no experience with them before, this really baffled me. So much so that, later in the day when I was trying to resolve a computer problem for someone in the RF department, I actually spontaneously launched into the story about what had happened earlier in the day, and asked the guy working in the cube if he, as an EE, could shed any light on the situation. He laughed at the story. As did the guy from a neighboring cube who'd stopped by to chat. But the magic thing was, he actually had an answer.
Not only that, it was a really good answer.
I'm paraphrasing, but here's what he said, roughly:
"Well, you're right. If you mount the fan blowing upwards, you're not fighting the natural tendency of heat to rise, and given the channeling effect of the heat-sink, the fan is going to have a cooling effect."
(Ok, good so far
"But you're forgetting the fan itself in this situation! If you mount the fan facing upwards, then it will actually be pulling the heated air through itself. The fan itself will actually heat up as a result of that, and the bearings aren't going t
Whoever designed level 61 in Frozen Bubble is a sadistic bastard.