Hi Guys
Some interesting info, so I thought I'd share it with you all
Where microprocessors are concerned, heat is the mischevious imp. It holds back the ability for microprocessors-generally dubbed 'chips'-to run faster or properly process data and it tends to severely rock the overall stability of your system. Which is by no degree good. This heat can't really be used either, so a drop in temperature is required. Changing the temperature from hot to cold isn't the easiest thing to do but over time the methods have matured. Which is a damn good thing and we'll talk about what sweet cooling you can look forward to in a jiffy.
Aargh, my heated core
But first, have you ever stopped and asked yourself 'why is heat so bad'? This is the simplest and, as it may seem, perhaps the dumbest question relating to heat. But its not stupid. Sure, to chips, heat is the quintessence of Hells heaters, but apart from the obvious unfavourable effects it has on a system, what exactly does heat do technically to hinder a chip's performance? In short, its the resistance in a chips circuitry which generates the high temperatures that needs to be removed. Essentially, it comes down to what the minute circuitry in chip's is made of-metals, or conductors. The main issue with conductor-lined chip's is the power resistance levels. If you bump up the heat, the higher the temperatures, in turn, amplify the resistance levels of the nanoscopic lines of metal. It then becomes a fight for equilibrium in a disproportionate manner because, as a result of the higher resistance, more power is dissipated over the chip. And voila...this creates more heat!!! Now I'm reasonably certain I mentioned that more heat increases the resistance...
A certain amount of resistance is expected-much is due to imperfections in the conductors or traces. When the resistance jumps up, higher levels of voltage are lost over the resistor and very little current manages to pass through. Consequently, less voltage and current are being delivered to other components further down the line, causing strange things to happen, as the voltage or current [both, even] is so immeasurably low [or non-existent if the resistance is high enough] that these components are on the brink of 'I can't trigger on, captain'. When they reach their peak, whammo, you have yourself a non-functioning processor, as the resistance is far too high. Aside from the more immediate and obvious implications of heat [like, say...instant meltdown], high temperatures also intensify long term problems such as oxide breakdown and the intriguing phenomenon of electromigration.
Electromigration
Electromigration is minute circuitry playing silly buggers and rearranging itself into an unusable or under-performing state, as a result of moving conductor atoms (ions) in the circuitry. Yep metal that moves around...The reason behind this is still not completely understood, but whats widely accepted is that its the outcome of electron movement-the transferring of momentum from the electrons to the ions that make up the tracery, or microcircuitry. When conducted through a metal, the electrons meet with blemishes in the circuitry and eventually cause the conductor atoms to scatter, building up an accumulation in the direction of the electron movement and leaving behind nothing but space (known as a 'void')
As heat is made up of vibrating atoms, the more heat there is, the more atoms are going to be in the wrong place at a particular time and this increases the rate of resistance and scattering. Take a look at the electromigration samples given. Those nasty large bulges are what are known as 'hillocks' and as you can see, the metal is building up and creating a large deposit of metal. These happen because the traces are so incredibly tiny, about 60 nanometres. This build up makes the metal atoms seperate and create a void large enough that electrons can no longer reliably, if at all, cross. Seeing as alive and working components are slightly more useful than dead ones, this is why we like to keep our things chilly. Electromigration is the devil.
Chilling out in the future
There's no doubting that there are some damn incredible water cooling setups out there, but the market has been flooded with drivel. But in the near future, the cooling we may call hardcore now, will be a necessity. Water cooling is mass-market now, and just not as hardcore anymore (covers head, lol). With the fast escalation of heat in processors and the need to overclock, the hardcore enthusiasts among us have used systems which involve designs such as peltiers and other refridgeration techniques, thermoelectric designs and, of course, nitrogen in its liquid state. The majority of tomorrow's designs are far more different to these. Most cooling technologies are moving towards nano-sized.
Microchannel water cooling
Microchannels as a means of cooling integrated circuits have been theorized since 1981, when Stanford professors Dr. David Tuckerman and Dr. Fabian Pease published research proving that microchannels etched into silicon could remove heat densities as high as 1000 watts per square centimeter.
The small size, light weight, and excellent thermal performance of the microchannel system allow tighter packing of components on the circuit board and higher reliability of individual chips as well as the entire system. By contrast, large finned heat sinks are heavy, and their mechanical leverage can crack a CPU or circuit board.
Microchannels—fine channels approximately the width of a human hair, etched into a silicon wafer—are built with a very high aspect ratio to increase their total surface area. As fluid flows through the microchannels, their large surface area enables them to cool hot spots.
There are two reasons for the efficiency of the Microchannel Heat Collector. First, the heat generated by the chip travels a relatively small distance from the transistors on the chip, where the heat is generated, to the walls of the microchannels. Second, the heat from the walls of the microchannels conducts a very small distance into the fluid before the heat energy is carried away to the radiator. As the microchannels get narrower, the walls of the channels stay cooler.
Nanotubes
Measuring between one and five nanometers in diameter -- that's roughly 1/50,000 the width of a hair -- they're nearly 100 times stronger, one-sixth as heavy, and 20% more flexible than steel. They far exceed copper's heat-retaining capacity, accounting for almost no thermal leakage, and they can carry an electrical charge at twice the speed of circuits embedded in silicon. Together, these attributes suggest an enormous potential for smaller, faster, cooler-running chips.
Indeed, some scientists believe nanotubes could lead to three-dimensional integrated circuits that would see transistors stacked not only side-by-side, but also atop and beneath each other -- a configuration researchers have never been able to achieve with silicon. Beyond chips, the proposed applications are even more spectacular, such as superefficient power cables and all-but-impenetrable armor. Trouble is, researchers haven't been able to figure out how to make nanotubes in consistent, commercially viable quantities.
Stirling Cycle cooling
Stirling cycle cooling is a concept that's a croos between standard refrigeration and Liquid Nitrogen, only its cooler, as it uses Helium instead. There are basically no moving parts, except for the piston, to compress the helium and a displacer which relocates the gas to where the heat is dissipated. It also reduces power consumption by the truckload when compared to refrigeration systems. There's even a prototype from Global Cooling www.globalcooling.nl operating on a crazily low 9W.
Pics from Global cooling, University of Notre Dame and additional info from Atomic magazine
Some interesting info, so I thought I'd share it with you all
Where microprocessors are concerned, heat is the mischevious imp. It holds back the ability for microprocessors-generally dubbed 'chips'-to run faster or properly process data and it tends to severely rock the overall stability of your system. Which is by no degree good. This heat can't really be used either, so a drop in temperature is required. Changing the temperature from hot to cold isn't the easiest thing to do but over time the methods have matured. Which is a damn good thing and we'll talk about what sweet cooling you can look forward to in a jiffy.
Aargh, my heated core
But first, have you ever stopped and asked yourself 'why is heat so bad'? This is the simplest and, as it may seem, perhaps the dumbest question relating to heat. But its not stupid. Sure, to chips, heat is the quintessence of Hells heaters, but apart from the obvious unfavourable effects it has on a system, what exactly does heat do technically to hinder a chip's performance? In short, its the resistance in a chips circuitry which generates the high temperatures that needs to be removed. Essentially, it comes down to what the minute circuitry in chip's is made of-metals, or conductors. The main issue with conductor-lined chip's is the power resistance levels. If you bump up the heat, the higher the temperatures, in turn, amplify the resistance levels of the nanoscopic lines of metal. It then becomes a fight for equilibrium in a disproportionate manner because, as a result of the higher resistance, more power is dissipated over the chip. And voila...this creates more heat!!! Now I'm reasonably certain I mentioned that more heat increases the resistance...
A certain amount of resistance is expected-much is due to imperfections in the conductors or traces. When the resistance jumps up, higher levels of voltage are lost over the resistor and very little current manages to pass through. Consequently, less voltage and current are being delivered to other components further down the line, causing strange things to happen, as the voltage or current [both, even] is so immeasurably low [or non-existent if the resistance is high enough] that these components are on the brink of 'I can't trigger on, captain'. When they reach their peak, whammo, you have yourself a non-functioning processor, as the resistance is far too high. Aside from the more immediate and obvious implications of heat [like, say...instant meltdown], high temperatures also intensify long term problems such as oxide breakdown and the intriguing phenomenon of electromigration.
Electromigration
Electromigration is minute circuitry playing silly buggers and rearranging itself into an unusable or under-performing state, as a result of moving conductor atoms (ions) in the circuitry. Yep metal that moves around...The reason behind this is still not completely understood, but whats widely accepted is that its the outcome of electron movement-the transferring of momentum from the electrons to the ions that make up the tracery, or microcircuitry. When conducted through a metal, the electrons meet with blemishes in the circuitry and eventually cause the conductor atoms to scatter, building up an accumulation in the direction of the electron movement and leaving behind nothing but space (known as a 'void')
As heat is made up of vibrating atoms, the more heat there is, the more atoms are going to be in the wrong place at a particular time and this increases the rate of resistance and scattering. Take a look at the electromigration samples given. Those nasty large bulges are what are known as 'hillocks' and as you can see, the metal is building up and creating a large deposit of metal. These happen because the traces are so incredibly tiny, about 60 nanometres. This build up makes the metal atoms seperate and create a void large enough that electrons can no longer reliably, if at all, cross. Seeing as alive and working components are slightly more useful than dead ones, this is why we like to keep our things chilly. Electromigration is the devil.
Chilling out in the future
There's no doubting that there are some damn incredible water cooling setups out there, but the market has been flooded with drivel. But in the near future, the cooling we may call hardcore now, will be a necessity. Water cooling is mass-market now, and just not as hardcore anymore (covers head, lol). With the fast escalation of heat in processors and the need to overclock, the hardcore enthusiasts among us have used systems which involve designs such as peltiers and other refridgeration techniques, thermoelectric designs and, of course, nitrogen in its liquid state. The majority of tomorrow's designs are far more different to these. Most cooling technologies are moving towards nano-sized.
Microchannel water cooling
Microchannels as a means of cooling integrated circuits have been theorized since 1981, when Stanford professors Dr. David Tuckerman and Dr. Fabian Pease published research proving that microchannels etched into silicon could remove heat densities as high as 1000 watts per square centimeter.
The small size, light weight, and excellent thermal performance of the microchannel system allow tighter packing of components on the circuit board and higher reliability of individual chips as well as the entire system. By contrast, large finned heat sinks are heavy, and their mechanical leverage can crack a CPU or circuit board.
Microchannels—fine channels approximately the width of a human hair, etched into a silicon wafer—are built with a very high aspect ratio to increase their total surface area. As fluid flows through the microchannels, their large surface area enables them to cool hot spots.
There are two reasons for the efficiency of the Microchannel Heat Collector. First, the heat generated by the chip travels a relatively small distance from the transistors on the chip, where the heat is generated, to the walls of the microchannels. Second, the heat from the walls of the microchannels conducts a very small distance into the fluid before the heat energy is carried away to the radiator. As the microchannels get narrower, the walls of the channels stay cooler.
Nanotubes
Measuring between one and five nanometers in diameter -- that's roughly 1/50,000 the width of a hair -- they're nearly 100 times stronger, one-sixth as heavy, and 20% more flexible than steel. They far exceed copper's heat-retaining capacity, accounting for almost no thermal leakage, and they can carry an electrical charge at twice the speed of circuits embedded in silicon. Together, these attributes suggest an enormous potential for smaller, faster, cooler-running chips.
Indeed, some scientists believe nanotubes could lead to three-dimensional integrated circuits that would see transistors stacked not only side-by-side, but also atop and beneath each other -- a configuration researchers have never been able to achieve with silicon. Beyond chips, the proposed applications are even more spectacular, such as superefficient power cables and all-but-impenetrable armor. Trouble is, researchers haven't been able to figure out how to make nanotubes in consistent, commercially viable quantities.
Stirling Cycle cooling
Stirling cycle cooling is a concept that's a croos between standard refrigeration and Liquid Nitrogen, only its cooler, as it uses Helium instead. There are basically no moving parts, except for the piston, to compress the helium and a displacer which relocates the gas to where the heat is dissipated. It also reduces power consumption by the truckload when compared to refrigeration systems. There's even a prototype from Global Cooling www.globalcooling.nl operating on a crazily low 9W.
Pics from Global cooling, University of Notre Dame and additional info from Atomic magazine
