Computer Cooling

[thermcool]

Hello Guys,
I would like to invite discussion on the current problems that we are facing in the cooing of the microprocessor for the laptops/desktop computers.

The number of microprocessors in the computer and their heat dissipating tendency is increasing year by year according to the predictions of Moore (Founder of Intel) who forcast that number of the transistors on the chip will double every 18 months.

At present the heat disipated by microprocessors of desktop computer is around 80-100 W max while for laptops is 20-30 W max. The areas of the chips are as small as 1-4 cm2 depending on the configuration and processing capability of the equipments.

There are number of problems that are currently being faced by the thermal engineers for the thermal control of the chipsets that includes local hot spots on the microprocessor (due to the unequal heat distribution on the chip), multiple heat source cooling (CPU, memory chips, display chips), rise in the heat flux and heat load of the processors each year etc.

Companies like Intel are now also looking at the methods like vapour compression refrigerators, Thermo electrical coolers which unlike the two phase heat pipe technology are active devices (means they consume power to operate) thus energy inefficient particularly for the devices like laptops that runs on the battery.

what is the future
two phase cooling (we need quite a bit enhancement in this)
active single phase pumped cooling
or mix of both ??

 

[BigInch]

My laptop fan is sooooo noisy I sware I'm getting a hearing problem. A constant ringing in my left ear. The fan side. Well... either that or its a brain tumor that goes away 1 hour after I turn this off. I can actually lower the volume on the TV 5 settings when I shut down. Its not only going to be a heat problem. Watch the noise generation too.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[25362]

Have you considered swapping the fan out for a lower CFM unit ?

See

http://www.cpemma.co.uk/

 

[BigInch]

If it wasn't a tightly built laptop that runs so hot it keeps my coffee warm, I would have already subjected it to major surgury. I don't like opening laptops, but in this case it may be necessary.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[ivymike]

Alrighty, let's try to stay on topic...
How about single-phase passive convection cooling, with a primary fluid other than air? Could you spread the heat load over a larger area via direct submersion of the CPU and local electronics in (an electrically non-conductive) fluid before transferring the heat to air?

2-phase would work better, of course, but you need to have a phase change in the right temperature range...

 

[KiwiMace]

You mentioned the two phase heat pipe - this is sealed and can both increase the heat transfer area and utilise latent heat to reduce size requirements. Wicked heatpipes should also work in laptops, where the inclination is not always sure.

 

[ko99]

thermcool, I don't know what the future will be but I welcome the discussion. Just some thoughts...

With 100-200W chips, the max thermal loss from the source to (eventually) air needs to be about 0.1 to 0.4 C/W. There are quite a few ways do do this. Most are pretty bulky and expensive or noisy. It's ironic that while the performance/size ratio of the electronics is dramatically shrinking, it causes the cooling (and power) solutions to dramatically grow, partly defeating the product potential.

Another obstacle, particularly at such high heat fluxes, is the interface loss. Some technologies may bring the rest of the thermal path under 0.1 C/W, but if the interface is stuck at 0.2 C/W, we're not solving the problem.


ko (

http://www.ecooling.biz)

 

[EdStainless]

On all CPUs the current trend (at last) is much more work/watt. I have seen current generation AMD machines running with the total draw at the wall <100W. Granted, it uses integrated graphics, but at least it is in the right direction. I have a dual CPU box with a big graphics card that is a couple of years old. At full load it draws about 600W from the wall. The only problem now is that graphics cards are drawing more power.

Back too the question. I see notebooks staying with passive cooling. Heatpipes will be the order of the day.

For desktops, there is great potential in vapor compression. Right now the favored approach is a compination of heat pipes and large, low speed fans. This is quiet, but takes up space.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[thermcool]

Thanks Guys for your replies,
BigInch, you are right acoustics is actually one of the major concern in the design of any thermal control system, You see whatever cooling means we device using two phase cooling or single phase cooling, in the end we have to dissipate the heat to the atmosphere using air cooling (fin heat exchanger). As the heat load for the CPUs increasing, high air flows are needed to dissipate the heat from the same or smaller size of the heat exchanger. This is giving rise to major acoustic problems. particularly, in the Laptops due to the space constraints this problem even complicates. Device like heat pipes, vapour chamber or even single phase cooling systems serves better to spread the heat from the high flux source or transfer it to remotely located sink but untimately all the heat has to be dissipate by air cooling.

ivymike,
May be direct submersion of the CPU is not acceptable to the electronic companies, it can lead to major electrical related problems. The major leap to something related to this is the integrated heat sink on the CPU like the microchannel machined on the CPU face. It works well, but the use of the microchannel along with the pumped system increases the cost of the cooling device immensely. Cost is one the major factor that also decides the acceptable solution.

CinciMace
Of course, wick heat pipe will continue to provide solutions to the current laptops but how long, Apple is already looking at the pumped loop solution for the laptop cooling. As the heat pipes finds a major limitation when multiple and concentrated hot sources are involved. The laboratory test done by researches claims very good potential for the heat pipes to control the high heat fluxes but unfortunately such a potential is not repeatable in mass production until we can look at the production side problems of sensitive product like heat pipe (to NCGs, impurities, material incompatibilities)

ko99
you are right, Interface resistance in any thermal product set a limiting performance. What do you think about the Nano-tubes (being grown on the chip and thermal solution), I heard last year some companies are coming up with such novel techniques to improve the interface heat transfer. Other means like diamond coating, graphite coating, using high conductive thermal paste (degradation of the these pastes with time is the major problem) etc.

the product performance is definately defeated due to the limitation on the thermal performance of the cooling devices.

EdStainless
The graphics processing units are definately adding and magnifying the total heat dissipated by the system. Devices like vapour chamber are also performaing very well as a heat transformer and spreader for the high flux server application.


Lately, some of the novel means including piezo fans, high conductive heat pipe material/coating with graphite, integrated heat pipe single unit systems with CPU to decrease the interface resistance has been sucessfully use along with the single phase/two phase counter part and claims to improve the thermal performance of the cooling devices.

My personal opinion is that piezo fans and Thermo electrical coolers might help out in future to improve and achieve the required thermal resistances. These device cant achieve the pupose along but when use as a secondary cooling device with the heat pipes or pumped system unit, they does play a part. For example, number of piezo fans when use inside the laptop at different dead spots create some force circulation that helps to elimiate the dead hot spots inside the system. similarly when TEC use at the interface decrease the consequences of the hot spots as well thermal resistance at the interface.

 

[tbedford]

hello everyone, i am a refrigeration mechanic, not a computer engineer...but i am sitting here with my laptop heating my legs just the same.

could we think simple and outside the box...

can you relocate heat sources to the vertical and backside of the monitor? we could even form cooling fins on the composite cover. It may even catch on as cool/trendy!!

 

[thermcool]

hi tbedford,
Thats actually very interesting idea and had been brought to practical use by number of laptop models in the past. However, there are number of limitation encountered in that design, the primemost limitation is the thermal resistance that is encountered at the hinge (heat pipe to heat pipe hinge) when the heat from the microprocessor in the main encolsure is transferred to the back side of the display.

Also, the display screen thickness is very small to include any two phase cooling system in it (e.g. vapour chamber), Although the use of thin vapour chambers ,ade from liqht materials seems potential but need thorugh investigation from material compatibility and choice of working fluid.

Besides, taking into consideration the requirement for the light weight for the laptops the use of the material like copper is restricted as a heat dissipating plate (condenser at the back side of the display)