Post by johnnybee on Oct 4, 2012 19:12:48 GMT
Sorry to resurrect this thread so late on, but it's important that members remain aware of the benefits of keeping the working parts of a PC as cool as possible; not only does the system work better, day-to-day, but it's also more likely to stay that way for longer.
Kev has outlined the basics, and he's absolutely bang-on in what he's said, but there's more to it than that - all you need is a bit of logic and imagination to improve the inner environment of your PC case, thereby making your surfing life both quicker and longer.
Firstly, look at your system and decide where the heat actually is - most folks would say that the CPU is the most critical point, and to an extent they'd be right; however there are other, much bigger components in the system that generate much more heat than a CPU, and this is where we need to take a good, hard look at what can be done to get rid of that heat.
As Kev says, air enters your case via the front vents - in most cases passing over the optical drives as well as flowing into the void above the motherboard - and in some cases it flows over (and maybe around) the hard drives mounted, in general, at the front of the case in the 3.5" bays.
If you look at the scene I've just outlined, you'll see that by the time the air has made its way into the central part of the case, it's already gained several degrees in temperature; however a few degrees won't make that much difference as long as there is sufficient flow throughout the casing.
This is where the question of overall flow comes into the equation; given that in a "standard" installation you will only get one extractor fan at the back of the case - normally an 80mm fixed-speed fan that does make discernible noise - plus the extraction effect of the PSU fan, normally positioned at the upper rear of the case above the motherboard level.
Some cases have provision for two extractors at the rear, in which case you can fit a second fan; however against that you have to account for the extra noise that a second unit will make - this can be significant, especially if a heterodyne builds up in service through differing fan speeds.
So, we have an overall situation where there is more extraction than supply - effectively encouraging a vacuum; however if you look at most cases, there are vents in the side casing that correspond to the AGP/PCI slots, thus drawing in cool air above any components in those slots to dissipate the heat that they generate.
This, then, is the "standard" cooling method for most ATX cases, and in fairness, as long as there is adequate room around the major components, it works.
The problems arise when ambients begin to rise, and effectiveness of fans atrophies through accretion of dust and crud; it doesn't take much to reduce the efficiency of a fan by up to 30%, and a 30% reduction in flow can easily result in an internal increase of 50% in terms of peak temps on memory modules and CPU's.
Most of you will have noticed that on most 'boards, the CPU and memory slots are mounted right up there next to the PSU in the top of the case - bloody silly place to put 'em, but we're stuck with it unless someone in mobo design finally wakes up to the basics of heat technology.
Now with most midrange 'boards this doesn't matter too much, as all that's needed is the addition of a front-mounted intake fan; however that then alters the balance of airflow through the case, bringing about the need for increasing the extractor effect to retain the cooling on the AGP/PCI bridge - in effect, a second, or bigger extractor fan.
So we now have one inlet and two exhaust fans in the case, hopefully moving enough air to keep everything within working temperature limits no matter what the ambient might be.
This however brings a price with it, that of the noise that these fans will make in operation.
Now you know and I know that every fan you see for sale is labelled as “whisper-quiet” or “silent in operation” in its description; however we all know that such isn’t the case – ALL fans make noise, big or small, so we have to look at minimising that effect.
It’s a well-know aerodynamic fact that big fans generate less perceived noise than small ones, and that is entirely due to the fact that they rotate more slowly, and have much longer blades than the smaller units; air disturbance is therefore much lower in both frequency and amplitude, hence much quieter.
In an ideal world, all fan feeds off the motherboard should be equipped with a PWM output voltage to control the fan speeds as required; however this is rarely, if ever the case, leaving us with the options of (a) running the case fans at full speed at all times or (b) fitting a dedicated fan controller into either a spare 3.5” or 5.25” bay on the front of the case.
Obviously choice (b) would be preferable, as the prospect of a 120mm fan bellowing away at full chat isn’t a good one, and two 80/94mm extractors are going to make a fair old racket too.
There is one old dodge that can help us here; working on the principle that the efficiency of a fan falls away rapidly past its optimum flow – around 60% of full speed on an 80mm unit – so wiring the two rear extractors in series instead of individually, effectively reduces the fan speeds without needing any further control or drawing any extra power – a win-win situation.
Okay then, so where does the heat come from?
The first source is the motherboard itself; there is around a mile and a half of microscopic conductors in the average board, and they all get warm during normal operation; however there are three other, more definitive heat sources mounted on the board, namely the CPU VRM, the board chipset and the graphics/sound chip, all of which generate significant heat output. Obviously the chipset has its own heat sink and sometimes a fan too; however the majority of the heat from the board is dissipated by the cooling air flowing through the case and around the back of the board itself.
Second and most obvious if the CPU – the heart of the system, and potentially the hottest component of them all.
Much nonsense is talked about this issue, but it needs to be stressed that although a CPU will destroy itself within seconds without any cooling pack, the heat is very much localised within the die itself; the spreaders fitted to the later ‘lidded’ CPU’s are the first stage of heat release from the die.
Back in the days of S370 Intels and AMD Athlon XP’s the die was miniscule, and merely booting the PC would be enough to fry the core without a heatsink attached.
Suffice to say that with a dedicated heat removal pack within the case, the CPU is the best protected component within the machine.
The question of where the hot air goes from the CPU is a moot one, and depends very much on the board architecture and the relative location of other components; broadly speaking the air within the case is moving from front to back, so it’s to be expected that *at some point* the warm air will be picked up in the flow towards either the PSU inlet at the top of the case, or the extractor fans lower down.
It’s as well to remember when looking at that particular flow, that on most boards the efflux from the CPU heatsink passes over the chipset cooler, and to a lesser extent, onto the grapics card adjacent to it.
That brings us nicely on to the graphics card, if fitted, which is the next largest producer of heat within the case.
Most GFX cards these days have their own heatsink/die arrangement that picks up cool air coming through the casing side vent and exhausts it radially in the same plane as the card itself. The big danger here is that most cards have their HS/fan comboes facing downwards, which means that the underside of the combo could easily be masked off by another card being mounted underneath it.
At best, it’s going to be a potential overheating problem, but with the masking effect of an adjacent card it could easily lead to local air recirculation and consequential GPU overheating.
This is where a side-mounted fan would come into its own; however it’s als a scenario that could easily occur where a passive GPU heatsink is used, so if there’s any doubt at all, a side fan is a must.
Coming now to memory modules, the unsung backroom boys of the system; in general, the heat they generate isn’t regarded as significant – certainly nowhere near CPU levels, and as such they are largely overlooked.
There’s nothing special that can be done about the cooling aspect of the modules themselves, except to make sure that the airflow around them isn’t impeded by cabling, etc. You could fit them with proprietary heat spreaders, but they are largely aesthetic, and won’t make much difference to their overall temperature.
The power suppy unit, or PSU is basically a self-contained, self- cooled transformer/rectifier that supplies various voltages to components as appropriate; the fan(s) that it comes with are internal, and in general are extremely reliable. It does however have a significant extractor effect on the PC case, drawing air from the interior and ejecting it out of the case at the rear; it is therefore part and parcel of the overall cooling system as a whole.
Look now at your optical drive(s), mounted in the front of your case; you will see that in most cases, they sit in enclosures within the case, and as such have no direct cooling airstream to ventilate them. Now as mentioned earlier, the case has more extraction than supply so there exists a very slight low pressure inside it – leading to air being drawn in wherever it can. The gaps between the CD/DVD drive and the front case faceplate, plus the small air gap between the unit and its disc drawer is normally enough to provide adequate flow over the gubbins to keep it workable.
Last but certainly not least, we come to those huge slabs of machinery, AKA hard drives, and this is where it does get interesting; these components vary in weight from 1.6lb to 3.7lb according to specification, and also contain their own control system mounted on a PC board under the frame.
Under idle conditions, they consume little power – typically 20watts or less – but under load the bigger units start drawing 60 to 70 watts, which in turn generates a considerable amount of heat.
Now again, the majority of PC cases put the 3.5” drive bay right at the front of the case, well above any incoming airflow and in a dead air area below the CD/DVD enclosures.
In addition, the bays tend to stack drives vertically with very little clearance between them – made much worse when using bigger, full-height drives over 100gb.
The result of this can be catastrophic; if the drive is used hard for any length of time, heat buildup reaches a point where the drive will cease to function, or worse, seize up solid.
There is no hard and fast solution to the problem, as the variations in case size and configuration are so many, however common sense and ingenuity should be the yardstick when addressing it; hard drives don’t have to be at the front of the case, and nobody’s stopping you from installing an extra fan to provide airflow where it’s needed.
I hope some of this dissertation has been helpful to you, and maybe given you an insight into an area of computers that rarely, if ever, is looked at in any depth.
I wish you well.
JB.
Kev has outlined the basics, and he's absolutely bang-on in what he's said, but there's more to it than that - all you need is a bit of logic and imagination to improve the inner environment of your PC case, thereby making your surfing life both quicker and longer.
Firstly, look at your system and decide where the heat actually is - most folks would say that the CPU is the most critical point, and to an extent they'd be right; however there are other, much bigger components in the system that generate much more heat than a CPU, and this is where we need to take a good, hard look at what can be done to get rid of that heat.
As Kev says, air enters your case via the front vents - in most cases passing over the optical drives as well as flowing into the void above the motherboard - and in some cases it flows over (and maybe around) the hard drives mounted, in general, at the front of the case in the 3.5" bays.
If you look at the scene I've just outlined, you'll see that by the time the air has made its way into the central part of the case, it's already gained several degrees in temperature; however a few degrees won't make that much difference as long as there is sufficient flow throughout the casing.
This is where the question of overall flow comes into the equation; given that in a "standard" installation you will only get one extractor fan at the back of the case - normally an 80mm fixed-speed fan that does make discernible noise - plus the extraction effect of the PSU fan, normally positioned at the upper rear of the case above the motherboard level.
Some cases have provision for two extractors at the rear, in which case you can fit a second fan; however against that you have to account for the extra noise that a second unit will make - this can be significant, especially if a heterodyne builds up in service through differing fan speeds.
So, we have an overall situation where there is more extraction than supply - effectively encouraging a vacuum; however if you look at most cases, there are vents in the side casing that correspond to the AGP/PCI slots, thus drawing in cool air above any components in those slots to dissipate the heat that they generate.
This, then, is the "standard" cooling method for most ATX cases, and in fairness, as long as there is adequate room around the major components, it works.
The problems arise when ambients begin to rise, and effectiveness of fans atrophies through accretion of dust and crud; it doesn't take much to reduce the efficiency of a fan by up to 30%, and a 30% reduction in flow can easily result in an internal increase of 50% in terms of peak temps on memory modules and CPU's.
Most of you will have noticed that on most 'boards, the CPU and memory slots are mounted right up there next to the PSU in the top of the case - bloody silly place to put 'em, but we're stuck with it unless someone in mobo design finally wakes up to the basics of heat technology.
Now with most midrange 'boards this doesn't matter too much, as all that's needed is the addition of a front-mounted intake fan; however that then alters the balance of airflow through the case, bringing about the need for increasing the extractor effect to retain the cooling on the AGP/PCI bridge - in effect, a second, or bigger extractor fan.
So we now have one inlet and two exhaust fans in the case, hopefully moving enough air to keep everything within working temperature limits no matter what the ambient might be.
This however brings a price with it, that of the noise that these fans will make in operation.
Now you know and I know that every fan you see for sale is labelled as “whisper-quiet” or “silent in operation” in its description; however we all know that such isn’t the case – ALL fans make noise, big or small, so we have to look at minimising that effect.
It’s a well-know aerodynamic fact that big fans generate less perceived noise than small ones, and that is entirely due to the fact that they rotate more slowly, and have much longer blades than the smaller units; air disturbance is therefore much lower in both frequency and amplitude, hence much quieter.
In an ideal world, all fan feeds off the motherboard should be equipped with a PWM output voltage to control the fan speeds as required; however this is rarely, if ever the case, leaving us with the options of (a) running the case fans at full speed at all times or (b) fitting a dedicated fan controller into either a spare 3.5” or 5.25” bay on the front of the case.
Obviously choice (b) would be preferable, as the prospect of a 120mm fan bellowing away at full chat isn’t a good one, and two 80/94mm extractors are going to make a fair old racket too.
There is one old dodge that can help us here; working on the principle that the efficiency of a fan falls away rapidly past its optimum flow – around 60% of full speed on an 80mm unit – so wiring the two rear extractors in series instead of individually, effectively reduces the fan speeds without needing any further control or drawing any extra power – a win-win situation.
Okay then, so where does the heat come from?
The first source is the motherboard itself; there is around a mile and a half of microscopic conductors in the average board, and they all get warm during normal operation; however there are three other, more definitive heat sources mounted on the board, namely the CPU VRM, the board chipset and the graphics/sound chip, all of which generate significant heat output. Obviously the chipset has its own heat sink and sometimes a fan too; however the majority of the heat from the board is dissipated by the cooling air flowing through the case and around the back of the board itself.
Second and most obvious if the CPU – the heart of the system, and potentially the hottest component of them all.
Much nonsense is talked about this issue, but it needs to be stressed that although a CPU will destroy itself within seconds without any cooling pack, the heat is very much localised within the die itself; the spreaders fitted to the later ‘lidded’ CPU’s are the first stage of heat release from the die.
Back in the days of S370 Intels and AMD Athlon XP’s the die was miniscule, and merely booting the PC would be enough to fry the core without a heatsink attached.
Suffice to say that with a dedicated heat removal pack within the case, the CPU is the best protected component within the machine.
The question of where the hot air goes from the CPU is a moot one, and depends very much on the board architecture and the relative location of other components; broadly speaking the air within the case is moving from front to back, so it’s to be expected that *at some point* the warm air will be picked up in the flow towards either the PSU inlet at the top of the case, or the extractor fans lower down.
It’s as well to remember when looking at that particular flow, that on most boards the efflux from the CPU heatsink passes over the chipset cooler, and to a lesser extent, onto the grapics card adjacent to it.
That brings us nicely on to the graphics card, if fitted, which is the next largest producer of heat within the case.
Most GFX cards these days have their own heatsink/die arrangement that picks up cool air coming through the casing side vent and exhausts it radially in the same plane as the card itself. The big danger here is that most cards have their HS/fan comboes facing downwards, which means that the underside of the combo could easily be masked off by another card being mounted underneath it.
At best, it’s going to be a potential overheating problem, but with the masking effect of an adjacent card it could easily lead to local air recirculation and consequential GPU overheating.
This is where a side-mounted fan would come into its own; however it’s als a scenario that could easily occur where a passive GPU heatsink is used, so if there’s any doubt at all, a side fan is a must.
Coming now to memory modules, the unsung backroom boys of the system; in general, the heat they generate isn’t regarded as significant – certainly nowhere near CPU levels, and as such they are largely overlooked.
There’s nothing special that can be done about the cooling aspect of the modules themselves, except to make sure that the airflow around them isn’t impeded by cabling, etc. You could fit them with proprietary heat spreaders, but they are largely aesthetic, and won’t make much difference to their overall temperature.
The power suppy unit, or PSU is basically a self-contained, self- cooled transformer/rectifier that supplies various voltages to components as appropriate; the fan(s) that it comes with are internal, and in general are extremely reliable. It does however have a significant extractor effect on the PC case, drawing air from the interior and ejecting it out of the case at the rear; it is therefore part and parcel of the overall cooling system as a whole.
Look now at your optical drive(s), mounted in the front of your case; you will see that in most cases, they sit in enclosures within the case, and as such have no direct cooling airstream to ventilate them. Now as mentioned earlier, the case has more extraction than supply so there exists a very slight low pressure inside it – leading to air being drawn in wherever it can. The gaps between the CD/DVD drive and the front case faceplate, plus the small air gap between the unit and its disc drawer is normally enough to provide adequate flow over the gubbins to keep it workable.
Last but certainly not least, we come to those huge slabs of machinery, AKA hard drives, and this is where it does get interesting; these components vary in weight from 1.6lb to 3.7lb according to specification, and also contain their own control system mounted on a PC board under the frame.
Under idle conditions, they consume little power – typically 20watts or less – but under load the bigger units start drawing 60 to 70 watts, which in turn generates a considerable amount of heat.
Now again, the majority of PC cases put the 3.5” drive bay right at the front of the case, well above any incoming airflow and in a dead air area below the CD/DVD enclosures.
In addition, the bays tend to stack drives vertically with very little clearance between them – made much worse when using bigger, full-height drives over 100gb.
The result of this can be catastrophic; if the drive is used hard for any length of time, heat buildup reaches a point where the drive will cease to function, or worse, seize up solid.
There is no hard and fast solution to the problem, as the variations in case size and configuration are so many, however common sense and ingenuity should be the yardstick when addressing it; hard drives don’t have to be at the front of the case, and nobody’s stopping you from installing an extra fan to provide airflow where it’s needed.
I hope some of this dissertation has been helpful to you, and maybe given you an insight into an area of computers that rarely, if ever, is looked at in any depth.
I wish you well.
JB.