Tom's Hardware > Forum > CPU & Components > CPUs > Logical, Cost Effective cooling

Logical, Cost Effective cooling

Forum CPU & Components : CPUs - Logical, Cost Effective cooling

Tom's Hardware: Over 1.4 million members in 6 different countries available to answer all your high-tech questions. Sign up now! Its free!

News

Ask the community Add a reply

Bottom Search this thread

I thus far have seen on this site and several others several articles having to do with water cooling (bulky) Liquid Nitrogen cooling (Expensive) and several other exotic methods to cool CPU's. I have also seen and read articles saying AMD cpu's are notoriously impossible to overclock air cooled. I am starting this post to hopefully allow for free exchange of Ideas.
I will start:
Parts Needed:
1 Dynatron 60mm Copper skivved fin heatsink
1 Volcano 9 Fan, set to max (gain ten degrees Farenheit of heat and use a variable Volcano 7 if quiteness is an issue)
1 Azteka 60mm to 80mm Fan Adapter
Result on a cheap, K7VMM board, with one stick of 256 2100 DDR
XP 1800 running at 105 degrees Farenheit (42 Degrees Celsius, roughly) UNDER LOAD (loading Windows 2000 Professional and assorted other software, then checked)

Add a reply

A good heat sinking fan is fine for a AMD CPU of course it might depend on how much you want to over clock it!

A cheap way to cool is open your case and put a box type fan blowing into the case.

the Prisoner

I'm not a number, I'm a free man! :mad:

Reply to the_Prisoner

Well, one thing that has been always in my head was the follow: coolers usually has the fan on top, blowing or sucking. But why don't put it on one side and even another one on the other side?

AFAIK only one HSF uses this method (I will look for it). But it seems to me more powerfull AND can direct the hot air to the fan located at rear top of your case. That way the HSF also helps to improve the airflow.

Another thing is to make a HSF that goes up at 45º, I mean not like a cube. Something like this:

.....[----] <- Your favorite fan
.....------
.../ HSF /
...------
------------ (this line is your mobo :smile: )

Then add some liquid cooling, like the late pipe designs (OTES, GeForce FX). The design at 45% should benefit from gravity to make liquid run better and perform better.

Just two ideas from a no expert THG member.

Still looking for a good online retailer in Spain :frown:

Reply to baldurga

Quote :

Well, one thing that has been always in my head was the follow: coolers usually has the fan on top, blowing or sucking. But why don't put it on one side and even another one on the other side?

I've often wondered about this myself for a long time. My thoughts brought me to a heat sink design like a triangle with the points lopped off.
__
/ _ \
\|_|/
\_/

The orange sections are where two small fans are mounted, blowing air over the cooler. The green section is where the die contact is made. The blue section is a small fan that sucks air across the cooler.

Have the thing oriented so that the ehxaust goes straight up into the power supply's intake. Have 'rows' of pillars that make channels to guide the air from the fans, and have these channels 'collide' right straight over where the CPU is situated.

I don't know why, but a design like this just sounds really good to me for some reason. I guess it's all the turbulence where the three air streams intersect that sounds good to me. Well, that and that the three fans would move the air pretty quickly, even if they are small fans at a relatively low RPM. I wish I had a degree in aerodynamics so that I could tell if I'm just nuts or what. Heh heh.

But ultimately you would think that you want the air going from the CPU and then right straight out the exhaust. You don't want that hot air just dilly-dallying around in the case, and especially not blown downwords onto the motherboard. You want the hot air to exit so that new cool air can come in. You want the CPU to suck from the direction of a new cool air source. This whole top-mounted CPU thing just seems a bit silly to me.

Then again, maybe I'm missing something.

<A HREF="http://forumz.tomshardware.com/community/modules.php?name=Forums&file=viewtopic&p=45775#45775" target="_new">Join the THGC LAN Party!</A>Edited by slvr_phoenix on 12/13/02 11:55 AM.

Reply to slvr_phoenix

Or in other words, more logically laied out in both color and orientation (since the other one was effectively oriented upside-down and had orange 'hot' intake and blue 'cool' exhaust, which was kinda stupid of me) design a heat sink that has an air flow like this:
| = exhaust fan (aimed straight at PSU)
. = processor
/\ = 2 intake fans angled against each other

<A HREF="http://forumz.tomshardware.com/community/modules.php?name=Forums&file=viewtopic&p=45775#45775" target="_new">Join the THGC LAN Party!</A>

Reply to slvr_phoenix

You guys might be surprised at how much work has already been done in this area. Take a look at this doc from back in 2000: <A HREF="http://www.intel.com/technology/itj/q32000/articles/art_4.htm" target="_new">http://www.intel.com/technology/itj/q32000/articles/art_4.htm</A>. Given that heat is an increasing problem, you might imagine that a lot of extra effort has gone into thermal solutions in the meantime. They've got some pretty slick things in the labs (which of course I can't talk about). We'll have to wait to see which ones end up hitting the market.

* Not speaking for Intel Corp *

Reply to sonoran

Hey man, Raystonn, the other Intel employee I told you about is back here. Now we got 2 Intel Employees in the THGC, heh.

--
The THGC Photo Album project,<A HREF="http://forumz.tomshardware.com/modules.php?name=Forums&file=faq¬found=1&code=1" target="_new">CLICK HERE!</A>

Reply to Eden

Nice Idea. I actually at my place of business where I work have already tested a few of these ideas jury-rigged or proto typed. Here are some problems you get:
1. Fans on the sides of the heatsinks only don't move enough air and current designs cause a push and pull fighting effect
2. Reason air goes straight down is that you are putting an air stream STRAIGHT AT THE DIE IN DOING SO. Lets be realistic here, most chips (at least XP'S, can't speak for Intel chips as have not used one in a long while) are set to run at 100 degrees Farenheit or higher. The hottest air in the case ain't that hot. Here is another principle of Physics: If you bring a fluid (air is considered a fluid) in to contac with a solid with more thermal energy some of that thermal energy will be transfered into the fluid, and continue to be transfered until the two temp.s are equal.
My idea was instead of changing something that could work far better and already exists to some exotic new design do the best with what is available. Example is if you want quite create an adapter to move all 70-100 cfm (cubic feet air per minute) of some nice 90mm fan running at about 2000 RPM (quiet) onto the chip. The basic concept is that if you can move between 75-100 cfm on to a small heatsink directly on top of the chip you can run an XP chip at 100-105 degrees farenheit unless it is overclocked. This is proven by the combo I already mentioned, but my improvement suggestios is the 90mm 'cause the Volcano9 at max flow is loud and the Volcano7 doesn't move enough air.

Reply to ksleeper

I didn't really see anything new in there. Unless I missed something, all of what they talked about is already available for purchase from one vendor or another.

The thing is, one has to wonder if the designs that these are all based on are even sound in the first place. I mean all of the true 'air' based solutions keep the hot air trapped in the case, and depending on how the heat sink fan is mounted, it might even keep the hot air on the motherboard. A considerable portion of the heat then basically ends up being sent back into the CPU or sent off to other components inside of the case. And as those other components heat up as well, they work towards increasing the ambient temperature of the air inside of the case, which hinders an air cooled CPU's ability to get rid of its heat.

I think the simple matter is, the solution of the future doesn't just have to be the most efficient heat sink in and of itself, but also has to be designed to keep the air flow through the case as efficient as possible for pulling cool air in and sending hot air out. The air cooling solution of the future basically has to look at the whole air flow inside of the PC, not just the air flow right at the CPU.

And that's the problem. Current heat sink development is always oriented specifically at the exact device that you are cooling. It would be fine if that was the only device present, but it is a horrible idea when numerous devices are all sitting there together. The RAM is next to the processor. The graphics card isn't too far away. And just beyond that are the hard drives. Which, consequently, many hard drives are cooled by the air intake, meaning that the air to cool the CPU is the heated exhaust from the hard drive cooling.

So far the logic in keeping the most important components cooled the best just isn't being done by air circulation solutions unless you really look long and hard or do your own mods. The only two solutions that really give thought to the circulation of the entire PC are water cooling systems and heat pipes. (And it could be argued that a heat pipe is actually just a very specialized water cooler.) They both in effect have 'radiators' placed at the very edge of the case (sometimes even outside of the case) with a dedicated air flow to pull that hot air away from the case. So the heat never gets recirculated to other components or back to the CPU.

So unless the CPU cooler gets redesigned to place nicer with the case's air flow, the future will only be open to water coolers for hotter CPUs.

Granted, the ideas for lowing the temps in the CPU itself is a good idea too. However, since Intel seems bent on making sure that each part of the CPU is continually utilized (with technology like Hyper Threading) there will be very little that can actually be done to lower temps in the CPU when running demanding software.

So really, more thought has to go into PC air flow than is currently being given if future PCs are going to get even hotter.

One thing I do like, the newest GeForce uses a cooling solution that takes up the PCI slot below the card entirely in order to suck air in through that unused slot and send the exhaust right back out of that unused slot. At least this keeps the video card from adding dramatically to the ambient temperature inside of the case. This was a good idea in my opinion. Hopefully more video cards will do this in the future.

<A HREF="http://forumz.tomshardware.com/community/modules.php?name=Forums&file=viewtopic&p=45775#45775" target="_new">Join the THGC LAN Party!</A>

Reply to slvr_phoenix

Each thing in its logical place. Where to start...
Water Cooling and Radiator cooling(heat pipes): Currently too bulky, expensive, and labor intensive for end user. Also, it is just as easy to fry a piece of electronic hardware with liquid as it is with heat.
Here is the idea: have a couple blowers just above the serial/sound/ps2/ports ect. part of the motherboard hookups on the back of the case to suck hot air out (Currenty you can get blowers in card slots only, at least without chopping metal). KEEP the air cooling on top of the cpu at high speeds, as it currently IS the most heat sensitive piece of hardware in the case. Also have a couple case fans (one on the power supply is already semi standard and cheap) pulling in cool air from out side to cool the rest of the case. No one is denying that keeping hot air in a case is a bad idea, but moving air has never been the problem, to a user willing to do a little research. The problem is the massive amounts of heat the cpu itself generates that need to be removed, and unless one wants a bulky water cooled or expensive Freon cooling system the best options for that is to allow for the fact that wherever you have a processing chip (be it NorthBridge, Cpu, or Gpu) heat will be generated and that heat should go in the air of the case wich can be flowed throughout the power supply (and through blowers in the future, hopefully) but in the meantime will allow the case to run slightly cooler.

Reply to ksleeper

Well good overclocking results depends on good research and purchaces.

The best thing is to get a good mobo, some fast ram and a CPU thats known to overclock well.
i.e. a XP1600+ 'AGOIA', TbredA XP1700+ or a TbredB XP2400+

a cheap board and cheap ram will really hold you back.

[Rik_] I wonder how many people have made their own phasechange system?
[LHGPooBaa] I get phasechange whenever i eat a hot chillie :lol:

Reply to lhgpoobaa

Ask the community Add a reply

Tom's Hardware > Forum > CPU & Components > CPUs > Logical, Cost Effective cooling

Go to:

Help |
Forum rules

There are 709 identified and unidentified users. To see the list of identified users, Click here.

Page generated in 0.886 seconds

Please mind

You are about to answer a thread that has been inactive for more than 6 months.
If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.

Add a reply Cancel

Become a fan

Logical, Cost Effective cooling

Forum CPU & Components : CPUs - Logical, Cost Effective cooling

Please mind