Saved Bookmarks
| 1. |
Solve : Computer switching itself off - overheating?? |
|
Answer» I don't think iNtel's ship with thermal paste applied.... I don't think iNtel's ship with thermal paste applied....The lowest thermal resistance is CPU direct to heatsink. Properly machined heatsink transfer most heat through that ‘two medium’ interface. A tiny amount of thermal compound will fill fewer holes. The ‘CPU to air pocket to heatsink’ (a three medium interface) has higher thermal resistance - is more like an insulator. If that air pocket has thermal compound, it is more conductive - but no where near as conductive as CPU to heatsink. Still is a ‘three medium’ interface. Thermal compound causes a single digit temperature reduction if using properly machined heatsinks. Most of the heat transfers directly ‘CPU to heatsink’. And most heat transfers only in the CPU center section. A poorly machined heatsink will have many more air pockets. Therefore thermal compound causes a greater temperature reduction. Thermal compound is not required (but usually used anyway) on better machined heatsinks. Essential to better heat transfer is very little thermal compound so that more of the heatsink makes direct contact to the CPU.Can you post new sets of temperatures?These are the new temperatures, taken whilst PLAYING the video game.. By the way, the thermal paste was already applied to the fan when purchased, ready for sticking onto the CPU ... I could see it when I pulled it apart to put it back on again. [attachment deleted by admin]Temps look fine to me Quote from: westom on July 08, 2009, 08:28:38 AM The lowest thermal resistance is CPU direct to heatsink. Properly machined heatsink transfer most heat through that ‘two medium’ interface. A tiny amount of thermal compound will fill fewer holes. The ‘CPU to air pocket to heatsink’ (a three medium interface) has higher thermal resistance - is more like an insulator. If that air pocket has thermal compound, it is more conductive - but no where near as conductive as CPU to heatsink. Still is a ‘three medium’ interface. So now we should recommend NOT using Thermal Compound ? ? Sorry but that's unacceptable here...Quote from: patio on July 11, 2009, 08:29:24 AM So now we should recommend NOT using Thermal Compound ? ?Can you guarantee that a heatsink assembly marketed to people with minimal knowledge is poperly machined? Install the heatsink without thermal compound. Measure CPU temperature. Then repeat the installation with thermal compound. If that temperature decrease is double digits, then post the name of an inferior heatsink here. First thing viewed by anyone selecting a heatsink is the 'degree C per watt' number. How many do what the informed would do? Meanwhile test for a better heatsink is simple and not destructive. However in a world where Arctic Silver is hyped as superior, even that simple test would not be performed. (Yes, a swipe at Arctic Silver whose massively profitable thermal compound is only equivalent to less expensive thermal compounds.) heatsinks aren't machined as precisely as an optical telescope lens... and if they were they would be prohibitively expensive. Thus the trade-off of having slightly-less precisely machined heatsinks with something to fill in the gaps that would otherwise be filled with insulating air. We must also take account that the CPU would need to be just as precisely machined for even the most expensive lenscrafter heatsink to work properly, and Processors are expensive enough. Its far more economical to use thermal grease or paste for a fraction of the cost then to pay a premium for some utopian scenario where heatsink and processor have a perfect, airless fit. your theories are wild and based purely on conjecture. the ideal scenario's you specify would require insane amounts of extra precision passed onto the consumer in the form of a wildly inflated price margin, with little actual gain other then the knowledge that their heatsink and processor, for the additional hundreds of dollars, now do the same task as a 5 or 10$ thermal paste. Meanwhile, you shun the use of other products such as "monster cable" (which I agree is hugely overpriced) and "artic silver". Both "monster Cable" and "artic Silver" you claim are hugely profitable. While I don't doubt that less expensive products can perform a job at least on par with theirs, I have to refute that despite their similar function, they do use more expensive components... in a way, it's similar to a golden toilet seat. Doesn't actually improve the experience, and it's really cold. But it sure as heck raises the budget of your bathroom remodelling. Monster Cable, for example, USES gold plated contacts, perhaps because gold is the most conductive material- for this same reason higher end motherboards use gold plated connectors. Artic Silver's composition includes, surprisingly, Silver. Both of these, as I'm sure we all know, are precious metals which generally are more expensive then the more common materials. (honestly I wouldn't mind a "arctic Copper" or "arctic Aluminum" compound, Since the silver really doesn't have a huge benefit over them). But to call these products hugely profitable is to believe that it is merely the brand that the consumer is paying for, as opposed to the passed on price of the more expensive composition.Quote from: BC_Programmer on July 11, 2009, 09:34:03 AM heatsinks aren't machined as precisely as an optical telescope lens... and if they were they would be prohibitively expensive.So heatsink manufacturers and Intel application notes were only theories? Nobody said anything about machining for telescope tolerances. Why did you invent that myth? Many heatsinks sold to computer assemblers have the cheapest machining possible. Are they are SELLING on engineering numbers? Or price? Where is the ‘degree C per watt’ number? Why are you arguing about using no thermal compound? On better heatsinks, compound is unnecessary. Then use cheaper stuff (that is equivalent to overhyped Arctic Silver), so that CPU is a few degrees cooler (and so that crappy heatsinks stop causing problems). Will thermal compound prevent processor damage? Of course not. A bare heatsink mounted on the CPU does not create hundreds of degrees necessary to cause hardware failure. As stated by those who do this stuff, demonstrated by the equations, and documented in applications notes. If you have a problem, also provided was an experiment. Post your numbers. The point: thermal compound is overhyped by myth. Often will mask inferior machining on some heatsinks. Any heatsink, properly machined, should provide sufficient cooling to any CPU. And then add thermal compound for a minor decrease in thermal resistance - to reduce CPU temperatures by single digit degrees. If you have a problem with this, then post numbers from your experiment without and then with thermal compound. the only way it can be machined enough to make thermal compound unnecessary IS to machine it to telescpe tolerances. Why do I need to do the experiment? How about instead of throwing out theories, you post YOUR numbers using a "proper" heatsink WITHOUT thermal compound. You refer to documents. Well guess what, everything looks good in theory. Thermal paste was created as an economically feasible alternative to the more expensive usage of a "properly machined" heatsink. Besides, even if I posted the 80(c)+ temps of my Socket 7 K6-2 BEFORE applying thermal paste, and now the "magic" reduction to 29(c) With it, you'd only pass off the entire issue as a "improperly machined heatsink" each time, with no regard for the FACT that you cannot buy this mythical "Properly machined heatsink". Instead you troll on about spec sheets and equations, none of which actually exist. Then you insist that others post the numbers to prove wrong these non-existent spec sheets and "equations" with an experiment that, if anybody actually did it, you'd invalidate with the "fact" that the heatsink was improperly machined, so why bother. The real question is why WE have to perform these experiments. Your the one with the unconventional ideas that obviously need a reality check. try frying a few processors, then maybe you'll decide on using thermal paste. Unless your using a 386 (which doesn't need a heatsink at all) or a 486 (which could usually get away with just the heatsink) your the one vending the fantasy. Post YOUR numbers proving the conventional knowledge to be a "myth", instead of proving yourself a insipid pseudosoph and telling others to prove your hare-brained scheme wrong. A common myth says that a processor rated for 70 degrees C will be damaged if heated to 100 degrees C. Why is that 70 degrees C so low. Because timing and voltage thresholds change so much as to adversely affect its operation - not life expectancy. Some have confused timing changes with hardware damage for the same reason he attacks this poster rather than provide facts or knowledge based in a few generations of experience. So what happens if you put no heatsink on a Intel processor? Nothing destructive. Also well known. And just another reason why that heatsink experiment is peformed by those who would learn rather than just attack. |
|