Sign in with
Sign up | Sign in
Your question

A Beginner's Guide For WaterCooling Your PC

Last response: in Systems
Share
March 28, 2007 11:20:07 AM

Everything you wanted to know about water cooling but were afraid to try.
March 28, 2007 12:30:01 PM

ummm... where are the graphs referred to in the article?
March 28, 2007 1:05:14 PM

Quote:
Our methodology is simple: push the e4300 as far as it will go with its stock air cooler, then benchmark it with the water cooling system and compare results.


So you were never going to try and push the E4300 higher under water?

I get you were showing load temp differential, but you can do more with water than just lower temps.

I second that, what graphs? 8O
Related resources
March 28, 2007 1:19:32 PM

I prefer the question.... what article???
March 28, 2007 1:38:33 PM

Quote:

So you were never going to try and push the E4300 higher under water?


Not really. This isn't an overclocking article, it's a water-cooling article for the uninitiated.

We have other articles that focus on overclocking.

Having said that, I'll probably be pushing it as far as I can, once I get some decent RAM. :) 


PS here are the benches, I'll post them here why the web guys are having trouble getting them up:

a b B Homebuilt system
March 28, 2007 1:40:58 PM

Alliterations aside, are there any add-ons that can monitor coolant flow and shut down the system if the coolant is not flowing? My primary scenario is that I turn the whole thing on and the pump does not pump for some reason - dead wire, dead pump, whatever. I would want to be able to keep the system from powering up.
March 28, 2007 1:41:29 PM

Alright, I guess my expectations exceeded the overall scope of the article.

FYI my E4300 was perfectly stable (tested) and then just gave out one day... 8O
March 28, 2007 1:42:30 PM

i enjoyed reading it, especially since i never really knew before just what was involved in a water cooling installation setup. good work :) 
March 28, 2007 1:43:01 PM

Quote:
Alliterations aside, are there any add-ons that can monitor coolant flow and shut down the system if the coolant is not flowing? My primary scenario is that I turn the whole thing on and the pump does not pump for some reason - dead wire, dead pump, whatever. I would want to be able to keep the system from powering up.


you can get an inline flow monitor.

Here at FronzenCpu.com

Aww come on choir... you know that much about hard drives/raid and you don't know about water cooling? Just kidding. :wink:
March 28, 2007 1:44:57 PM

Thanks for the great article! I've always wanted to try my hand, and this article makes me feel like I could actually pull it off. I appreciated the ability of the author to put this topic "down on the bottom shelf" so that I could grasp it!
March 28, 2007 1:48:19 PM

The Koolance system in pareticular can monitor the temperature and shut the PC down based on that; other manufacturers probably offer similar features as well.

I didn't want to speak to manufacture-specific features that much in the article though, as this is a general how-to.
March 28, 2007 1:49:11 PM

It would be nice to go into things n00bs get hung up on (or more often make mistakes with), such as mixing up tube diameters or buying an external system with a case that doesn't have holes in the back for tubes. Maybe provide a picture of a shorted ATX connector, as well, for those unfortunate souls that might find the wrong instructions online.

Also, the MB PWM is often neglected with water-cooling, and takes a beating in an overclocked system. Inverting the ATX case fan next to the CPU is generally a quick-fix, or even just adding after-market heat sinks to the regulators. Either way, it's important to watch the PWM temps on the motherboard when overclocking a water-cooled system.

There are a lot of cheap system to stay away from, such as thermaltake's bigwater, which has a short pump-life, and is NOT covered under warranty (When the pump goes, you buy a new system).

What about putting the splitter before the CPU? The GPU tends to get the hottest, and benefits the most from very cool water. What about sound levels? Can you put a db meter up to the new system versus the old?

Just my 2 cents.


Thanks.
March 28, 2007 1:49:13 PM

Quote:
FYI my E4300 was perfectly stable (tested) and then just gave out one day... 8O


Ouch. Sorry to hear that, mate.
March 28, 2007 1:51:33 PM

i had always just been too worried that something was going to go wrong at some point during/after installation (primarily leaks), so i had never seriously pursued it before, its always looked interesting though
March 28, 2007 1:52:52 PM

Quote:
The Koolance system in pareticular can monitor the temperature and shut the PC down based on that; other manufacturers probably offer similar features as well.


Just FYI, though, if I'm not mistaken you need a serial-port to take advantage of this feature. You can get a USB-to-Serial converter for cheap online, but it's good to know ahead of time ;) 
March 28, 2007 1:53:51 PM

Nice guide. Can you do a beginners' guide to liquid nitrogen cooling now, please?
March 28, 2007 1:54:58 PM

All good points. There are a million things I could have done or added, but that's not what I was trying to accomplish.

This is a primer and sometimes you have to draw a line, although you are knowledgable enough that I think it's safe to assume you aren't the target audience for this one. :) 
March 28, 2007 1:55:48 PM

Quote:
FYI my E4300 was perfectly stable (tested) and then just gave out one day... 8O


Ouch. Sorry to hear that, mate.

It happens. It took 3 days of tweaking... it was a good accomplishment 379*9 :) 

--Choir

I feel ya. had my first rig sprung a leak randomly and nearly shorted out my board (it was my first install and the clamps were cheap 8O ). Water sounds scary but you can mitigate nearly every risk and this article goes a long way to alleviate some of those problem areas.

Whizard

The CPU has stricter heat tolerances than the GPU (which operates generally at higher temps, afaik) so you spend as much effort as you can cooling the CPU, hence the splitter after the CPU block not before.
March 28, 2007 1:56:46 PM

Quote:
Nice guide. Can you do a beginners' guide to liquid nitrogen cooling now, please?


8O

Hell yeah!
March 28, 2007 1:57:00 PM

Quote:

Just FYI, though, if I'm not mistaken you need a serial-port to take advantage of this feature.


No, it works directly with the MB's power switch. The system doesn't need to be hooked to the motherboard any other way.

There is a serial looking cable between the system and it's custom bracket - but the bracket's tiny board only hooks up to molex power, nothing else on the mobo.
March 28, 2007 2:20:18 PM

Quote:
Specific heat capacity is the other important physical property, which refers to the amount of energy it takes to heat a substance by one degree. The specific heat capacity of liquid water is about four times that of air, which means it takes four times the amount of energy to heat water than it does to heat air. Once again, water's ability to soak up much more heat energy without increasing its own temperature is a great advantage over air-cooling.


I didn't dig out a reference to check the specific heat values, but I think it should be pointed out that the vast majority of water's specific-heat advantage is due to it's density rather than it's specific (by mass) heat capacity. Even if water had half the specific heat capacity of air, it's density is, what, about 1000 times greater, so it would still more readily absorb the heat. The conclusion is the same, though I think the emphasis should be that the cooling contribution of the conductivity of water (25x air's) is paltry in comparison to heat capacity (and then forced convection by the pump). It would be more interesting to see a specific heating capacity with respect to volume at standard pressure. My guess is this would put water at 4000x the heat capacity of a given volume of air.
March 28, 2007 2:23:02 PM

Quote:

Just FYI, though, if I'm not mistaken you need a serial-port to take advantage of this feature.


No, it works directly with the MB's power switch. The system doesn't need to be hooked to the motherboard any other way.

There is a serial looking cable between the system and it's custom bracket - but the bracket's tiny board only hooks up to molex power, nothing else on the mobo.

Ah. Cool.

Good primer. I'm looking forward to the next iteration.

P.S. It's nice to see the author active in the forums.


Oh, yeah: UV-responsive tubes with some cold-cathodes look awesome.
March 28, 2007 2:45:30 PM

Quote:

Oh, yeah: UV-responsive tubes with some cold-cathodes look awesome.


Hell yes! Wish I'd have had some around for the article.

Makes the PC look like it's alive with the Predator's glowing neon blood flowing through it's veins... :) 
March 28, 2007 2:48:53 PM

Eh. Not a bad article....but the first picture on the second page is wrong.....I can't remember the last time I used a heatsink that the fan didn't push air into the heatsink.
March 28, 2007 2:53:20 PM

Wouldn't have to worry about leaks with the MB mounted upside down. I know heat rises - and it would be a problem with passive cooling - but with active cooling such as this liquid based system - I would think it wouldn't matter - and if a leak should happen to appear - it won't drip on anything of consequence.
Also - I would like to see some advanced cooling techniques - and maybe include the hard drives in the cooling scenario.
Alternative liquids, and gases even - I read somewhere about a system that uses steam to carry away heat - and someone here mentioned liquid nitrogen - seems there would be some concern about lower than ambient temps creating condensation - would have to monitor humidity levels as well -
What is the limit of cooling - the colder the system the greater the performance? To what point?

How about building a system into a small freezer or fridge? If the whole system is immersed in cold - would there be any condensation problems?
March 28, 2007 3:09:19 PM

Quote:
I can't remember the last time I used a heatsink that the fan didn't push air into the heatsink.


You're right of course, the diagram is more of a simple schematic illustrating that ideally the CPU heated air is pushed out of the case. It would have been a bit more difficult to illustrate properly but I probably could have put more time into it.
March 28, 2007 3:10:24 PM

The fridge or freezer suggestion comes up frequently. You shouldn't have to worry about condensation, because the computer is warmer than it's surroundings, and the fridge/freezer dehumidifies the air anyway. However, the compressor on a fridge/freezer is not capable of handling a 200+Watt continuous load. It's designed to cool things once and then keep them cool. Think of how long it takes for your fridge to cool down a large turkey that just came out of the oven, or just a big pot of boiling water. You'd burn out the compressor quickly.

You could set up an insulated box cooled by an air conditioner unit, which is designed for nearly continuous operation, and you can easily get one capable of pumping 1000+Watts of heat. But then, in the summer time, wouldn't you just prefer to air condition the entire office or whatever room you're in?

If you could find the link to the steam cooling, I'd be very interested. I imagine the working fluid is kept at a very low pressure so that the boiling point of water is below the operating temperature of a CPU.
March 28, 2007 3:41:46 PM

Heres the link to the article:
http://news.zdnet.com/2100-9596_22-6168908.html

As for the fridge/freezer scheme - I see the point in the constant cooling needed for an active heat source - I have a friend who is going to school for HVAC - I will inquire as to requirements and specifications for such a system - also good point about air cond in the summer - but I'm talking about REALLY cooling this thing off - as in below freezing if possible. I don't keep my 'stat that low.
March 28, 2007 3:47:47 PM

Quote:
i had always just been too worried that something was going to go wrong at some point during/after installation (primarily leaks), so i had never seriously pursued it before, its always looked interesting though

Those worried about leaks, check this product out.

http://www.fluidxp.com/

It is a NON-CONDUCTIVE alternative for the liquid component of your liquid cooled system. I personally purchsed it and am running it on my system. C2D E6600 & Nautilus 500 and I idle at 25C as in this article. I have a very minor overclock for now (running at 2.7GHz) and max temp at full load is 32C. I will overclock more this weekend, I just haven't had more time to play around with it (plus my RAM is needs replacing). Check out the reviews on the fluidxp page in the left hand column.

Quote:
Eh. Not a bad article....but the first picture on the second page is wrong.....I can't remember the last time I used a heatsink that the fan didn't push air into the heatsink.

I agree, all the heatsinks I've ever seen the air was pushed downward over the heatsink rather than pulled up.
March 28, 2007 3:53:29 PM

Sounds just like a heatpipe, only the working fluid is water instead of... well, I don't know what's in a "standard" heatpipe.
March 28, 2007 4:00:00 PM

Guys, this picture in the article has problems:



I think most CPU HSF design as we know it blows air toward the heat sink, not suck from it.

and then from the article (CPU temperature benchmark section):

Quote:
This doesn't seem like a big deal at all when we consider that the stock cooler is a low-end part and that a premium aftermarket air cooler would probably be more effective


Where is the comparison for the "premium aftermarket air cooler" ?

then there is the question of locating the graphs the article mentioned. Am I missing something?

Geez, Where is the people who edit the article? It really looks like a mess.
March 28, 2007 4:08:49 PM

Quote:
Guys, this picture in the article has problems:


Yeah, my response to that on the first post on this page. Look up! :) 

Quote:
Where is the comparison for the "premium aftermarket air cooler" ?


Isn't one. This isn't a premium cooler vs. water cooling comparison, it's a water cooling primer. Would be nice to add everything I could ever want in the most comprehensive article of all time, but that's where I drew the line.

Quote:
then there is the question of locating the graphs the article mentioned. Am I missing something?


?? That I can't answer, I've emailed the web guys to fix it a while ago. In the interim I've posted the graphs on the first page of this forum post.
March 28, 2007 4:12:51 PM

Hahaha... the guy's right here in the forums. I can understand not reading every post of an 8-page thread, but really it wouldn't have taken much work to figure out that most of the things you're complaining about have already been addressed. You're demanding very high standards at the same time as you demonstrate that your actions don't stand up to the same standards.

Edit: Changed wording from "...that you don't stand up to the same standards".
March 28, 2007 4:13:50 PM

Thank for the graphs cleeve. I could see the numbers in the text, but the visuals help. :D 
March 28, 2007 4:41:42 PM

Quote:
Guys, this picture in the article has problems:

This doesn't seem like a big deal at all when we consider that the stock cooler is a low-end part and that a premium aftermarket air cooler would probably be more effective


Where is the comparison for the "premium aftermarket air cooler" ?

then there is the question of locating the graphs the article mentioned. Am I missing something?

Geez, Where is the people who edit the article? It really looks like a mess.

Amen. This looks like about a third of an article. Why would someone spend $500 on a cooler when their PC works fine at stock speeds? Why would you compare a $500 cooling system to stock coolers? How many people care about temperature only, and not temperature as it relates to overclocking, etc? I don't think I've ever seen a sig line that says 'e4300 stock 33C at 100% load'.

Was this article paid for by Koolance? Most of the section reads like the Koolance installation guide, only with fewer pictures:

http://www.koolance.com/support/files/manual_exos2_100d(eng).zip

Why is there no comparison of decent air cooled systems, or of other water cooled systems, or a demonstration of why someone would spend $500 on a funny looking box that sits on top of their computer (ie overclocking)?

Why talk about performance charts and not include them? Why does it look like half of a review? Did Koolance set a deadline of the end of March for their advertainment to get published?
March 28, 2007 4:47:41 PM

Once again, this is a primer for the uninitiated, not a how-to overclocking article. It's also not a premium air cooling vs. water cooling article. I know many of you would prefer those articles, but this isn't them. Don't know how to make that any clearer to you.

If you don't like it fine, your perogative. But accusing me of writing a paid advertisement is over the line.

If I want to see an article benchmarking a Hauppage PVR capture card, I don't read an X1950 PRO vs. 7900 GS article and bitch that the writer was paid off by Ati and Nvidia... :roll:

If you guys would like to see a premium air cooling vs. water article, let me know and I'll consider it for the future. :) 
March 28, 2007 5:03:11 PM

Does anyone actually read through a thread before posting? :roll:
March 28, 2007 5:21:32 PM

Quote:
Does anyone actually read through a thread before posting? :roll:


Yeah, sorry to hear about your dead 4300.
March 28, 2007 5:36:55 PM

I liked the article. Very nice and easy for a starter guide.

I would only add a few things thing:

You never mentioned if the water blocks you got were aluminium or copper water blocks. Mixing those can have some nasty effect on corrosion. Talking about corrosion, you never metioned algae, which may grow in the tubes and corrode the water blocks.

Also, would be nice adding something on the coolant like water, distilled water or other coolants, and what are the benefits or disadvantages of each. Also, if using biocide along with the coolant is good.

I know it's a lot of info, but hope you can address these and others on your second article on liquid cooling. :)  (btw, if you could address the 1/2" tubing would be nice - a lot of wc veterans do suggest using these along with 3/8" water blocks, so you can get a tigher lock on the tubing :D  )
March 28, 2007 5:47:16 PM

Quote:
If you guys would like to see a premium air cooling vs. water article, let me know and I'll consider it for the future. :) 


I would like to see that.

To be honest, he does make a good point: 99% of the people looking to water-cool are going to want to overclock. Processors stay pretty cool at stock speeds, even with the cheap (not inexpensive; cheap ;)  ) air solutions.

I understand that you were looking to give a n00b "how-to" on setting up water-cooling, but I think the point of water-cooling is to achieve an overclock beyond what air-cooling can provide. Given that, even n00bs interested in water-cooling would want some overclock results more than what were given. The only other reason to water-cool is sound.

I know, we've been over this, but I still believe the target audience would be interested in either peak overclocks or sound levels; neither of which were really fully addressed.

I'm not trashing the article, but I did want to point out that I think it's missing the point of water-cooling in the first place; even from a n00b's perspective. It would've been nice to see a little more done in the form of overclocking or sound tests, just to see how much heat this thing was really capable of carrying away.

All of this is just to reinforce the request for an overclockers' article for water-cooling with the Koolance :) 

(Disclaimer: It's easy to nit-pick, and that's what I'm doing. If it had been overclocked well, people would complain that it's not a better processor, and so-on and so-forth. You can't please everyone, but I just wanted to point out where I thought the article lacked a bit. A 40-page article might have been a bit too much, anyway ;)  ).
Anonymous
a b B Homebuilt system
March 28, 2007 5:48:35 PM

With the mobo mounted in a horizontal position with the componentry face-down, a liquid heat exchanger on the CPU/GPU would work just fine. Heat only "rises" in "free" convection.

The water in the liquid cooling system is the "working fluid." That working fluid could be COMPRESSED air to get its mass density up so it would work better. But the amount you would have to compress it would require rigid plumbing and pressure fittings. Lots of downsides.

In heat transfer, heat transfer = mass flow x specific heat X temp difference. Water (liquids in general) work far better than gases due the the mass alone.

Interesting, it seemsl like the steps to get this system going resemble 'burping' a car radiator system - not coincidental methinks.

As an over-view, I thought the article was EXCELLENT. The text, pix, and graphics were great...nothing read like a promotional for a particular manufacturer. Activity of this sort is common in any industry; technical fellows who write industry reference books/guidelines/trade rag articles tend to work for companies that also sell stuff into the industry they are doing the piece on. If I was the manufacturer, though, I'd still point potential customers to this site, you'd be stupid not to use all the free press you can get. But the author didn't slam other manufacturers or solutions, and there was none of the glitzy photo studio pix you might see on the outside of the product box - I think if the manufacturer was going for the "adver-edu-tainment" angle, there would be at least one booth girl in the pix...

Let's quit acting like somehow enthusiast sites don't cater to promoting an interest in new/better/faster hardware in general - it would be naive to otherwise thinks so. To wit, every review article that may mention an manufacturers name would then be suspect. Pretty much would have to shut the site down to avoid the appearance of manufacturer influence....wouldn't want that would we?
March 28, 2007 5:54:07 PM

Quote:
Sounds just like a heatpipe, only the working fluid is water instead of... well, I don't know what's in a "standard" heatpipe.


Heat pipes amaze me.

Still, I tend to think that they should lose their effectiveness once they reach a certain temperature, which makes me avoid them (I could be wrong).

I've always assumed that once your heat pipes reach a certain temperature at the radiator, the fluid won't liquify, and your heat-pipes cease to function (or function well). I'm probably wrong.

Does anyone have any good write-ups on heat-pipes in CPU-coolers (I've found plenty on HVAC units).
March 28, 2007 5:59:44 PM

So you included the liquid cooled power supply but you didn't liquid cool any of the hard drives?
March 28, 2007 6:02:50 PM

Quote:
Once again, this is a primer for the uninitiated, not a how-to overclocking article. It's also not a premium air cooling vs. water cooling article. I know many of you would prefer those articles, but this isn't them. Don't know how to make that any clearer to you.

If you don't like it fine, your perogative. But accusing me of writing a paid advertisement is over the line.

If I want to see an article benchmarking a Hauppage PVR capture card, I don't read an X1950 PRO vs. 7900 GS article and bitch that the writer was paid off by Ati and Nvidia... :roll:

If you guys would like to see a premium air cooling vs. water article, let me know and I'll consider it for the future. :) 


Don't worry too much Cleeve, it seems like it has become popular around here to pick apart every article that comes out because it doesn't cover every posible angle any reader could have though of.

I mean couldn't you compare each of the processors currently available too see which had the biggest drop in temps by watercooling vs. stock cooler vs scythe infinity, and the best overclock you could acheive with each on air, and then water? :roll: Just a little more effort Cleeve. :roll:
March 28, 2007 6:06:19 PM

I don't have a technical article for you to look at, sorry.

I'm sure heatpipes are optimised for a specific temperature range, but I'll bet that the design temperature range of a heatpipe used in a CPU HSF is significantly larger than the range of temperatures you'll experience. Regarding the concern over running out of liquid, it should be considered that as more liquid turns to vapour, the pressure inside the tube rises. With rising pressure, you increase the boiling point of the fluid, so that the range of temperatures for which you should still have liquid in the tube is larger than you might otherwise.

The biggest drawback of a heatpipe is that the hot vapour is held very close to the cool liquid in the wick. This is why heatpipes' performance is proportional to tube diameter / length (or some similar expression). One solution is the loop heatpiple (LHP), which has the hot vapour taking one path to the radiator while the returning cool liquid follows a wick along a separate path. LHPs are more efficient, but significantly more expensive. It seems the best air coolers can keep a very hot CPU well within design temperatures even at significant overclocks, so it may be for now that the standard heatpipe is good enough. I've only heard of LHPs being applied in notebooks (source: Master's student working with LHPs) and satellites.
March 28, 2007 6:15:32 PM

Quote:
Once again, this is a primer for the uninitiated, not a how-to overclocking article. It's also not a premium air cooling vs. water cooling article. I know many of you would prefer those articles, but this isn't them. Don't know how to make that any clearer to you.

If you don't like it fine, your perogative. But accusing me of writing a paid advertisement is over the line.

If I want to see an article benchmarking a Hauppage PVR capture card, I don't read an X1950 PRO vs. 7900 GS article and bitch that the writer was paid off by Ati and Nvidia... :roll:

If you guys would like to see a premium air cooling vs. water article, let me know and I'll consider it for the future. :) 


I saw that you were defending it as a primer, but what is in this article that isn't in an installation manual? I have never installed a computer watercooling system or seen one up close so I was very interested in this article when I saw the title. After I read through it I felt cheated. I have gained greater understanding of watercooling principles from reading the forums here and on other sites than from this article. The installation instructions and pics were decent, but the manual for the product has better ones. The stats showed some impressive numbers but were without context, and therefore meaningless. Yes, a fancy watercooling setup beats a stock cooler. Tell us something we didn't know. You started to with the one paragraph comparison to an aftermarket GPU cooler.

To make this a more useful 'primer', talk about the basics of watercooling in greater detail. A little more research and accurate info on the thermal properties would be good, like a previous poster pointed out. You say there are three options, built into case, in case, and out of case. What about preconfigured kits vs piecemeal setups? What about potential pitfalls other than leaks and serial configuration? Pump failure, the importance of thermal safeties, battery effects, fully passive silent setups? Flushing air pockets? Types of waterblocks? Decreased/increased load on the PSU? What maintenance is required?

You can exclude options, but tell us why you are excluding them. You had a one line blurb about kits being better than piecing together a system, but you didn't back it up or explain that you were not going to mention piecemeal setups from that point on. Or say that the EXOS is self-flushing.

You then include benchmarks that show dramatic improvement, but no useful info. Why include benches at all? If you are going to tease us with temp comparisons then dazzle us with pretty charts and comparisons of different configurations, please.

I'm sure all you got from this was the free cooler, but Koolance should be paying you. This article will probably sell quite a few EXOS-2's, but will it rightly sell them? Are they the best value compared to good air cooling or any other brand of water cooling? I understand that an article comparing a dozen water and air cooling systems would be a massive undertaking, but showing the stats of one other alternative would help put the EXOS into perspective. You also quote some impressive stats about the EXOS, but you don't say whether they are above or below average for water coolers.

An article talking about more options than just the Koolance EXOS would probably have worked better for a primer, as would an article that detailed all options and then briefly showed the EXOS as an example instead of emphasizing it. After reading this article, a novice would only know about the EXOS, without understanding so most of the possible pitfalls of a water cooling system, or of the other options and configurations out there. An intermediate user just sees things that are left out or improperly explained.
March 28, 2007 6:24:03 PM

Quote:

I saw that you were defending it as a primer, but what is in this article that isn't in an installation manual?


Let's see:

- a discussion of water cooling benefits and detriments on the PC
- a listing and definition of the different types of water cooling systems available (integrated, internal, external)
- a description of the basic workings of water cooling
- a discussion of the multiple components most commonly liquid cooled and why
- benchmarks of stock vs water cooling temps


Looks like the only information that is similar to the manual is planning and installation. Gee! How could a reasonable water cooling primer talk about that! RELEASE THE LAWYERS! CLEARLY I AM PLAGIARIZING! :roll:

I drew a line at what I was covering and what I wasn't covering and you don't like it. Once again, that's fine. But if you're using this as evidence I'm working for Koolance, then it's just as viable that your posts are evidence that you're working for a competing air cooling heatsink manufacturer... :wink:
March 28, 2007 6:25:02 PM

(in the same tone as the post above)

You make excellent suggestions, but the tone of your writing is such that the author is put on the defensive, so he is less likely to develop your ideas. Why make suggestions at all if you do it in a way that attacks the other party?
March 28, 2007 6:26:07 PM

Wikipedia has a good article about heatpipes, and explains how they can work over a broader range than the working fluid at atmospheric pressure.

http://en.wikipedia.org/wiki/Heatpipe

An interesting comparison would be heatpipes v watercooling. Heatpipes have many orders of magnitude greater thermal capacity in their operating range, but water coolers have much more effective dissipation with their external radiators.
March 28, 2007 6:27:54 PM

Quote:
I don't have a technical article for you to look at, sorry.

I'm sure heatpipes are optimised for a specific temperature range, but I'll bet that the design temperature range of a heatpipe used in a CPU HSF is significantly larger than the range of temperatures you'll experience. Regarding the concern over running out of liquid, it should be considered that as more liquid turns to vapour, the pressure inside the tube rises. With rising pressure, you increase the boiling point of the fluid, so that the range of temperatures for which you should still have liquid in the tube is larger than you might otherwise.

The biggest drawback of a heatpipe is that the hot vapour is held very close to the cool liquid in the wick. This is why heatpipes' performance is proportional to tube diameter / length (or some similar expression). One solution is the loop heatpiple (LHP), which has the hot vapour taking one path to the radiator while the returning cool liquid follows a wick along a separate path. LHPs are more efficient, but significantly more expensive. It seems the best air coolers can keep a very hot CPU well within design temperatures even at significant overclocks, so it may be for now that the standard heatpipe is good enough. I've only heard of LHPs being applied in notebooks (source: Master's student working with LHPs) and satellites.


Ah. Thanks.

The pressure thing makes sense, and seems to be what I'm missing.
!