CPU cooling solution maintenance ... rant
- Thread starter Psi*
- Start date
gibbersome
Coastermaker
But do I get to use the Stargate? Now THAT would be beyond cool....
Sure....here you go:
http://ny1.com/9-staten-island-news.../new-dyson-fan-blows-away-the-need-for-blades
Psi*
Tech Monkey
Just what I figured, just big enough to fit my head or arm or leg thru, but not my whole body.
2Tired2Tango
Tech Monkey
WOW... way cool (pun intended)... I would be really interested in how much air that moves and even more interested in how it works... Google time....
Found it... Bernoulli's principle...
http://blogs.howstuffworks.com/2009/10/14/invention-the-bladeless-dyson-fan/
Last edited:
I'm not a fan (pun also intended). Sure, it's a cool concept, but $300 for a fan? Especially a fan with such an ugly base? I'll take a bladed fan, thanks.
If you're going to use those Scotch brand pads, I'd recommend ya find the cheap offbrands. Dollar stores usually have them at half the cost and they are just a tiny bit thinner, so airflow works a bit better through them.
I get a bunch of mine from a local flea market... hard to believe, a cheap China-import knockoff that was actually more useful than the name brand version. 15 pads in a pack for a buck, no tax, and still a little thicker than needed.
They are dense, but airflow is only slightly decreased. The problem is putting them directly in front of a fan will raise the noise level slightly... add at least half an inch of space though and it the fans will stay pretty quiet. For really low speed fans (I'd say sub 10,000RPM 120mm's) it doesn't even matter. To be honest I've not used them on my main rig though, and they don't really fit 200mm fans either...
I get a bunch of mine from a local flea market... hard to believe, a cheap China-import knockoff that was actually more useful than the name brand version.
15 pads in a pack for a buck, no tax, and still a little thicker than needed.They are dense, but airflow is only slightly decreased. The problem is putting them directly in front of a fan will raise the noise level slightly... add at least half an inch of space though and it the fans will stay pretty quiet. For really low speed fans (I'd say sub 10,000RPM 120mm's) it doesn't even matter. To be honest I've not used them on my main rig though, and they don't really fit 200mm fans either...
2Tired2Tango
Tech Monkey
Ummmm... a little "aifrlow" lesson, if I may....
More fans is not always the answer. Often less is more... especially when cases are getting smaller and people are increasingly outspoken about noise.
PC cases, by design, are intended to draw cool air into the least occupied part of the system (lower front) move it through the entire case and then push it out from the hottest part (power supply). The thermal solution in "tower" cases relies upon a slight depression (lower air pressure) inside the case to create eddy currents that disperse themselves throughout the case, mixing air to provide more even cooling.
You probably won't believe me but the last thing you want is fans pushing air into your case. This tends to create jets of organized air that move directly from one fan to another... without cooling much of anything along the way. I've often noted that taking intake fans out of computers results in cooler operation ... and by more than a mere couple of degrees.
In fact most computer cases can be adequately ventilated by the fans in the power supply... and FWIW that is why the case design positions it at the upper back of the case. For computers under heavy load an auxiliary fan just below the power supply is usually more than adequate. Covering unused fan openings on the back and sides of cases can also result in an improvement in overall cooling.
The best solution is to use active exhaust ventilation at the upper back of the case and passive intake ventilation at the lower front.
Also you shouldn't put filters anywhere near fans, ever. The intake of a fan has to represent the lowest air pressure in a system. Anything that inhibits that, even slightly, is going to severely reduce airflow and affect cooling throughout the entire case. Air filters should be used only on passive air intake ducts where they catch dust before it gets inside the case.
To be perfectly honest with you, I figure the use of intake fans does little more than blow dust into a case. I doubt they contribute much to the overall cooling solution.
Attachments
Psi*
Tech Monkey
I just saw these today with a professional shef friend of mine Scotchbrite griddle cleaning screens. They seem to come in brown only, but they are perfect screens, 4" x 5 3/4".
I never would have noticed them had this not come up in the thread.
I never would have noticed them had this not come up in the thread.
b1lk1
Tech Monkey
You cannot use a blanket statement like this in describing the efficiency of case fans and cooling potential. In fact, each case has different requirements as well for the types and power of fans. I have tested quite a few cases that have positive pressure and are running large quiet fans which cool the system far better than anything with negative pressure. In fact, I would bet money we will see more and more positive pressure designs. Most negative pressure cases available today are the lower end and the companies cheap out with giving us less fans. Placing more fans in them, and almost always as intake fans, helps them as well.
As for dust, as long as one regularly cleans intake filters there is never an issue. My house is somewhat dusty and as long as I maintain mine monthly my PC stays consistent.
As for best fan placement, so far the Silverstone Raven RV02 with it's positive pressure design and 3 large 180mm intake fans VS a single 120mm exhaust fan was the best case I have tested in a very long time in regards to quietness and effectiveness. Only the Coolermaster HAF932 and 922 came close to it in overall performance.
Finally, since positive pressure has been proven to be far more effective, covering any unused openings would not help. In my experience, low pressure negative airflow cooling is just inadequate at best and as noted is used in primarily low end market offerings.
2Tired2Tango
Tech Monkey
Does it not strike you as just a little odd that in order to improve over the "low end" solution of 80mm fans in the power supply and case back, you have to go to multiple 120mm fans placed all over the case?
My current case is an older version In-Win mini tower. The only air intake on it is on the front of the case. The only exhaust air is the power supply fan. The system is an AMD X64 dual core 4200+ running a mild overclock at 2.4ghz, with the supplied cooler, on an Asus M2N micro-atx board with an Asus passively cooled 8400 series Nvidia card... And yes, I have a scotchbrite pad over the intake... The chipsent runs warmer than the CPU at 38c. The CPU even under full load while rendering video or compiling code never gets over 50c.
Many times the big, impressive, expensive and noisey cases we see in use today are nothing more than a way of getting deeper into your pockets. Think about it...a $300 case no matter how flashy and dressed up, is still a rectangular box, dimensioned to fit the same equipment as that $79 cheapy is... How much better can it be??? If we are arguing over 2 or 3 degrees, we are wasting each other's time, my friend.
While I agree that in machines under heavy load a somewhat different cooling solution is often required. Water cooling, ducted cold air, etc. I have a couple of servers where I've gone some lengths to keep them cool... But these are on 24/7 with average CPU loads of 75 to 80%. The average "sitting at my desk typing along" user doesn't need anything even remotely so exotic... In fact, even gamer systems often don't need half the fans people put in them.
I can give you a good argument against over-cooling.... Just walk into an office with 20 computers running 4 or 5 fans each. Just the noise is enough to change your mind on the issue. In fact, it might even explain why some people "go postal"...
Last edited:
b1lk1
Tech Monkey
You missed my point completely. The cases I am talking about are not even close to noisy while the low end cases always use far noisier and very much audible fans. The low end cases also do not cool that well and can barely handle stock running systems.
You are forgetting something majorly important. High end cases appeal to high end users. People running quad core CPU's and multiple graphics cards. That may only be 10% of all users, but that 10% are the ones spending 90% of the money on expensive high end parts. We are not Dell/HP/etc.. users here. We are tech enthusiasts and for most people posting on this site and pretty much all tech sites, money is not an object.
The example you give about the temperatures of your PC sounds spot on. Now try putting an X4 in there and a couple HD4870's and a sound card. Your cooling would be horribly overwhelmed to the point of causing the PC to throttle back or even have parts failures.
As for your argument that the bigger more expensive cases are merely a bigger space to fit parts, you need to do alot more research. The design and construction of high end cases is not even remotely similar to low end. Low end cases never take proper airflow into consideration while this is the primary concern in high end cases.
You also must remember that nearly all high end cases are costing from $150-200CAD right now while the best low end cases are in the $75-100CAD range. I have not seen a $300 case in some time. Even after spending $50 for a few new fans for that low end case you would get far better results by just spending the cash the right way the first time.
I suggest you start reading up on case technology and as many reviews as you can. You are really far off base here. For an office PC, there is no need for additional cooling since the heatload is so small it only needs one exhaust fan to do the job. A high end PC, especially a full on gaming PC, absolutely must have high end cooling including a properly designed high end case. I am not arguing over 2-3 degrees. I am explaining the difference between melting parts and running them at a reasonable temperature. Try more like a 20C+ difference between the best possible low end case and the cheapest possible high flow case. I guarantee that the high end case will always be quieter as well by a large margin.
You are forgetting something majorly important. High end cases appeal to high end users. People running quad core CPU's and multiple graphics cards. That may only be 10% of all users, but that 10% are the ones spending 90% of the money on expensive high end parts. We are not Dell/HP/etc.. users here. We are tech enthusiasts and for most people posting on this site and pretty much all tech sites, money is not an object.
The example you give about the temperatures of your PC sounds spot on. Now try putting an X4 in there and a couple HD4870's and a sound card. Your cooling would be horribly overwhelmed to the point of causing the PC to throttle back or even have parts failures.
As for your argument that the bigger more expensive cases are merely a bigger space to fit parts, you need to do alot more research. The design and construction of high end cases is not even remotely similar to low end. Low end cases never take proper airflow into consideration while this is the primary concern in high end cases.
You also must remember that nearly all high end cases are costing from $150-200CAD right now while the best low end cases are in the $75-100CAD range. I have not seen a $300 case in some time. Even after spending $50 for a few new fans for that low end case you would get far better results by just spending the cash the right way the first time.
I suggest you start reading up on case technology and as many reviews as you can. You are really far off base here. For an office PC, there is no need for additional cooling since the heatload is so small it only needs one exhaust fan to do the job. A high end PC, especially a full on gaming PC, absolutely must have high end cooling including a properly designed high end case. I am not arguing over 2-3 degrees. I am explaining the difference between melting parts and running them at a reasonable temperature. Try more like a 20C+ difference between the best possible low end case and the cheapest possible high flow case. I guarantee that the high end case will always be quieter as well by a large margin.
Last edited:
2Tired2Tango
Tech Monkey
No, i didn't miss your point at all.
Based on experience that started in the 70s and progressed into a career as a design level technician (i.e. "Technologist") I quite simply disagree with you.
Last edited:
In some instances I would definitely agree. In my opinion it does depend on the hardware inside the case, including any overclocking on top of that.
After realizing the mistakes with my first case purchase I bought a mid-size Antec case, it came with a single 120mm on the back below the PSU. While very effective, the case still cooled better when I added a second 120mm intake on the front, and it also helped improve the airflow by disrupting the bubble of heated air underneath the GPU.
A good example of what you're talking about could be one of those cases with a giant 250-300mm fan on the side... the airflow from those just doesn't go anywhere. They create a cushion of air that hugs the motherboard and insulates it from the cooler air that bounces off and gets drawn or forced out the sides of the case. They most often tend to cool components worse than two or three 120mm fans in a push/pull config.
This is true, but some manufacturers have tower designs that are no longer designed around solely this principle. Quite a few cases in recent years have been designed around having a positive internal air pressure, while others focus intake fans to breakup deadzones that are otherwise created... One reason is to use your single rear 120mm fan scenario, a single high-end GPU that doesn't exhaust any heat out the back of the case will quickly create a hot spot underneath itself that can grow far hotter than even the PSU exhaust at the top of the case. It just comes down to the hardware in the build more than anything else. In your example you use a passively 8400GT, and that's almost as cool a GPU as one can get.
A single Pentium 4 and an ATI 9600XT will get way to warm with just a 120mm exhaust fan, forget about removing that and going with just a PSU fan to vent the entire case! Such ideas were practical when CPU's didn't need fans, but today only highly-efficient SFF builds could get away with the PSU cooling the system.
More efficient and cooler running hardware can be had today, but the average computer is going to require more than a single PSU fan by long and far. Especially as the average computer owner almost never blows the dust out of the PSU fan until after the machine has overheated and shut down a few times, or a component in the PSU cooks itself to death.
I would agree, it's not optimal and does wear out the fan quicker to place a filter over it. It would be best to have at least a full inch of open air in front of any intake fans to minimize disrupting the low air pressure they create, and the best possible solution would be to use any filters as a passive intake duct where possible. But it is not always an option.
Such a system would be sufficiently cooled by a PSU fan, but that's about as cool running hardware as you could expect to find short of buying power efficient CPUs, or someone that underclocks and undervolts their hardware and runs an IGP. For quite a few years Dell multipurposes the CPU's 120mm fan to double as a case exhaust fan in their desktops. Another issue is many enthusiast PSU's keep their fans running at very low speed, far to low to effectively cool the entire computer let alone properly cool the PSU components.
I don't know about others, but I don't buy cases based on 5 degrees. Once they meet my airflow requirements I look to the features and what I can do with such a case. I'd have to say most people buy cases based on features and ease of use.
If everyone used GPU's cool enough to run passively, then sure. I remember my first 8800GTS... huge chunk of metal, really felt like a brick. Unfortunately regardless of the cooling it always got hot enough to burn skin if you touched any part of the metal not hidden by the plastic. Even at idle it was to warm to comfortably touch. Recent GPU's like the GT200 series, HD 4800 and HD 5800 series are no different.
Number of fans doesn't have to equate into decibels of noise... an office doesn't look at fan speeds or cooling efficiency when buying their desktops. If I replaced the Delta fans in my case with 9 Noctua's it would be truly silent even with the AC and everything else turned off. Presently I can only hear that it's running if the room is dead quiet as it is, so I haven't done so.
Some poeple go postal on their fans, but that's their choice to do so... if I wanted to go nuts I'd raise the fanspeeds in order to run at 4.4GHz all the time. Personally if given the decision I think 200MHz is an easy sacrifice for a quiet, well-ventilated system.
Last edited:
2Tired2Tango
Tech Monkey
Of course it does. I haven't totally divorced reality on this one.
But that doesn't stop me from seeing better solutions in many cases (pun intended).
I've even had people come to me about my insistance on quiet PCs and pay me to get the racket out of their offices. In one such job I walked into the office and everyone was literally shouting over the din... I later found out their in-house computer genius was a hard core gamer who insisted on exorbitant hardware with way too many fans for what they were doing. First pass through this I went around and underclocked all their systems --most were overclocked-- and pulled out 3 and 4 fans per computer... they loved it. Second pass I hooked them up with a bunch of ASRock S330s and set up a proper server (in the Boss's office of course) to handle the heavy load.... The server is quad core AMD, water cooled with dual external radiators. Quietest office I've ever been in; but they still talk in raised voices. I'm thinking it will take a while.
Your commentary raises several very interesting points...
In particular the bit about multiple fans run at low speeds. I'm guessing that most people don't realize airflow through a fan will actually stall if the low rotational speed fails to create pressure and reverse airflow can actually happen when other fans do create pressure. The whole 10 fans at 25% speed thing probably has less cooling effect than a couple of fans running near full capacity... and it's probably no quieter. This is one of those "What were they thinking" deals where consumers are actually sold on methods that are both self-defeating and expensive. (In another thread on this topic, I posted a simple schematic for a 50 cent fan quieter that works wonders in this situation.)
For your point about the heat bubbles that can form near GPUs and some PCI cards... simple trick, remove an unused expansion cover and let cool air vent in underneath them. I would recommend this as a first course of action over adding more fans. Enlarging passive air intakes can also be a big help in this department... Many manufacturers do not provide adequate intake openings, burying them in case covers etc. A few minutes with a nibbling tool can bring a definate gain in cooling.
The recent shift to positive pressure cooling is interesting but it has a couple of it's own drawbacks... the first being an increased need for maintenance. Fans blowing into a case will also force dust into the case. With high velocity airflow this dust is going to disburse itself uniformly throughout the case. Worsen this with nicotine or high humidity passing through the case and you have a sticky little dust magnet on your hands. Again lack of consumer awareness comes into play when you realize that a layer of dust so thin you can barely see it can reduce a passive heatsink's efficiency to dangerously low levels. While this solution may work better in the very short term, keeping it going is going to require greatly increased maintenance on the system... and as we both seem to agree, mostly that just doesn't happen.
A very interesting rebuttal Kougar... and thanks for not telling me how to spend my money.
Good to hear! And I'm sure there are better solutions to many things, that's the great advantage to information technology... getting access to or figuring out how to find that information to arrive at a better solution.
Seriously, even I'm a gamer and would know better. Hmm, maybe I should start a computer consulting firm, that's the sort of job I'd excel at...
You are quite true, but that nasty word "depends" comes into play here again. A proper low-RPM fan has the blades designed to suit the low RPM's, Noctua offers several radically different blade designs on their fans to fit everyone from the ultra-low noise audiophile to the silent PC enthusiast, to the gamer that just wants some airflow at reasonable volume levels he won't hear over the explosions.
As I've said in past discussion about low speed fans with you, I use Panaflo/Panasonics & Deltas. Admittedly they came with the case, but I initially kept them due to their great acoustics while maintaining good airflow. The specific model is designed to operate as low as 7V and shows the CFM/Air Pressure dropoff in relation to the voltage range on their PDF specifications.
Frankly I do think the airflow design of this cube case itself sucks, but there isn't much I can do about that. What matters is the motherboard +GPU get airflow and the radiator can exhaust smoothly out the back of the chassis. The cube case at least works well enough for that. Soon as I find an affordable tower that fits a 3x140mm radiator on the top or back without some major metalwork, I'll be upgrading... if my PA12.3 was the newer model that allowed for standard fan spacing I'd already be using a HAF 932 right now.
Removing brackets certainly helps if the case is solid, but most high performance cases already use ventilated or open PCI slot covers for this reason. Additionally if you use a positive air pressure case, the hot air is forced out of the PCI slots lowering GPU temps further.
Yes, with greater amounts of airflow come greater amounts of dust.
For heatsinks that are poorly rated for the amount of heat they must disperse I might agree, but a capable heatsink isn't going to lose much until it has more than a tiny layer of dust on it.
With my Thermochill radiator, I've noticed the thing seems to be more effective at collecting dust than even my air cleaner! However, the surface area is so great that there can be moderate amount of dust in the unit and I still won't notice a temperature change on my CPU cores. This thankfully means I can stick to a 3-4 month cleaning schedule without anything getting thickly covered or any noticeable loss of cooling performance.
Regarding smoking... I don't smoke, but it's a dead giveaway by opening any computer if the owner does. It's not just fatal on their lungs but it can get back enough to gunk up the inside of a fan housing and kill the fan after a year if they smoke heavily or use cigars.
Last edited:
2Tired2Tango
Tech Monkey
Yep, I've seen fans that will move more air at 500rpm than some of their 3600rpm cousins. But seriously, how many people do you know who A) have even the first clue of this or B) would know where to get them even if they did?
I wouldn't exactly call Panaflo a "low speed" fan. They spin at 3600rpm like most fans... but they do it quietly. Yep... clover leaf bladed fans where the blades overlap slightly will work better a lower RPM than hook bladed fans with no overlap... but once again, we're back to the question of how many people actually know or care?
Lets be real... a lot of computer builds are done for pure machismo... I find it particularly interesting that the computer store I deal with tells me their sales of "replacement fans" has nearly tripled since they started stocking the ones that glow while spinning. Really a lot of this is directly from the "what you got under the hood" hot rodder mentality of the 50s.
(Hint... chromed valve covers do not make your car go faster.)
Ever had one on a proper hot-plate test rig? Spray water across the top of it with a misting bottle... the temperature will go up. Tap a little baby powder onto it and the temperature really goes up. Unless a heatsink is grossly overrated (and the ones in PCs seldom are) they are going to be badly largely affected by dust accumulations.
Yep... seen more than one PSU with it's intakes all gummed up and closed over. The stuff you take out of there has roughly the consistency of putty...
Well, if the person knows enough to undervolt fans the I would surmise chances are they would figure some fans are better than others at being undervolted and why. Almost every fan maker offers multiple default speed fans for their lineups.... I think most people buying fans don't even realize they come in several speed classes or know anything about dBa levels, and would agree with you on those points.
About Panaflo, if you want to be specific then yes. They push a surprisingly good amount of air when silent and that's the big reason they're still in use! They do offer several models at different speed grades.
But Chrome bumpers are the best for stopping vehicles cold during an impact...
About Panaflo, if you want to be specific then yes. They push a surprisingly good amount of air when silent and that's the big reason they're still in use! They do offer several models at different speed grades.
But Chrome bumpers are the best for stopping vehicles cold during an impact...
2Tired2Tango
Tech Monkey
Please don't take this the wrong way and do note that I am including myself when I observe how easily the techy crowd loses track of just how little the general public actually knows about anything.
Example: I have an Uncle who recently retired from a GM assembly plant. His entire working life was spent installing parts on engines and rear axels. One day he was telling me about some problem he was having with his spark plug wrench, how long it took to get a mechanic to look at it... and in the process of this I realized that after 30 years in that plant, he didn't know how anything he worked with or on actually worked.
Sadly I think this is the new normal in technical understanding. Even so called technicians today are dumber than fence posts compared to electronic hobbiests of only a couple of decades ago... Most of this new crowd is totally lost around test equipment, about half don't even know how to solder... and we trust them with the most complex systems mankind has ever created.
All that being said... we return to my original point that case ventilation is not simply a matter of stuffing in more fans. There is a definate science to cooling systems and very often those who practice the science would benefit from a better understanding of it.
For example... Have you ever played with stirring fans? A fan inside the case, moving air inside the case, used to create turbulance and break up heat bubbles...
Yeppers... as I said in our previous discussion, best darned fans I've ever used.
But Chrome bumpers are the best for stopping vehicles cold during an impact...
LOL!
Last edited: