CPU cooling

[BloodRedDragon 0]

I'm looking for cpu coolers for 1 ghz p3's, something capable of keeping them cool enough for overclocking. Any recommendations?

 

K

[BladeRunner 0]

I shall stick to my advice on a copper unit.

The Thermalright SK6 with a 60mm YS-Tech fan was installed on my PIII 1Ghz and offered the best performance of all the HSF units I tried.

 

Ya honestly think all these companies would release copper if they didn't actually do the job?

[CrazyKillerMan 0]

If you want some hardcore fan noise, almost any heatsink will work. Throwing a 7000rpm delta on any heatsink will do wonders.

 

But if you want to get a new one I would recommend the SK-6 or the thermaltake dragon orb 3.

 

I have run the dragon orb for a long time now and it works great, even the slower 4500 rpm fan.

 

There are a couple heatsinks out there that work really good and are really expensive.

 

The SK-6 is a REAL solid working heatsink, though it may cost you, depending on who u buy it from.

[BloodRedDragon 0]

The problem I have is with the heatsink on my first cpu. It's cocked, and I think it's because I busted the clip, namely the light colored lever. Obviously, when the lever is down, it helps distribute the pressure evenly against the heatsink. It was even the first time I installed it, but I somehow managed to break the lever in the process of transfering to the new motherboard. Does Intel sell them seperately?

 

I have two 50mm fans attached to the case now for extra cooling. Just need to hook them up electrically. Also set-up 4-bay scsi tower and loaded it with my cdrom drives, and stuffed an ide cdrom drive in the computer, thereby replacing the two drives that were in it. Still have one drivebay left open now for another device. Ditched my pci slot wasting SB PCI64 soundcard in favor of the onboard sound of the mobo. Eventually I will get an Audigy when I can add it to my budget.

 

K

 

PS->As a machinist, I have to vouch for alluminum as a better heatsink. The only way copper can be a good heatsink is with a well cooled case. The problem is with copper being denser material, and therefore maintaining a temperature. The copper heatsink will draw the heat off the cpu faster when it's cold, but once it's hot the cooling effectiveness will drop because it can't dissapate the heat into the air fast enough. Alluminum is poor at maintaining temperature because of its low density, but it dissapates heat way better than copper. As a result, the cpu temperature will probably climb faster with an alluminum heatsink, but will stay reasonably cool for a longer operating duration than with a copper heatsink due to the copper will eventually heat up and lose it's cooling effectiveness. Still, it's somewhat dependent on case cooling. If you can keep the the copper heatsink cool, it might not be so bad. I'd like to see a test done on heatsinks after running them for 12 hours or more while the cpu is busy running Quake 3. That would be a more realistic test in keeping with what you really need the heatsink for. Also, a copper heatsink with alluminum fins could be worse due to the fact that the manufacturing process that attaches the alluminum to the copper needs to maintain good surface to surface contact between the two materials. Who needs to worry about another surface contact point issue when we already have one with the cpu and heatsink?

[clutch 1]

This is what I use and it's awesome:

 

http://www.vantec.com.tw/cck6035d.html

[clutch 1]

OK, as for heat dissipation and what metal is best. I just spoke to our Engineering Manager, and he dug up a book named "Marks' Standard Handbook for Mechanical Engineers". In this book in section 4-81, it lists certain values for the thermal conductivies of metals. For a given temp range of 70*-700*F, Aluminum is 130 and Copper is 232 (higher is better). Of course, it is using symbols that I can't reproduce here, but I can scan it in if it's THAT big of a deal. Also, in temperature conductance and resistivity, here's the order and value of some metals you might have heard of:

 

1. Silver 9.8 Ohms cir mil/ft

2. Copper (drawn) 10.37 Ohms cir mil/ft

3. Gold 14.7 Ohms cir mil/ft

4. Aluminum 17.01 Ohms cir mil/ft

 

These were listed in order from best to worst of this particular group of metals. I would imagine that those people told APK that aluminum was better because it's a lot cheaper to work with, and can be cheaply anodized to protect from oxidation. It is also considerably lighter for the same amount of surface area that's used. So, there you have it. Do what you wish with the info, but I will probably still get copper myself since this cheap Vantec really kicks a$$.

 

[clutch 1]

Umm, I stated that copper is not only a better conductor of heat, but has less resistance to the transfer of heat as well. It will also disipate heat quicker due to the other 2 properties. And, that comes from this handy-dandy book on the mechanical properties of metals and other such lovely information.

 

Also, as for aluminum being used in engine components, the REAL reasons are:

 

A. It's lighter

B. It's cheaper to work with

C. It's deformation with heat is close enough to iron (though not in the case of one particular Cadillac engine and the Quad 4 from Oldsmobile) to be used in most cases.

 

There is no way in hell that anyone would want a copper cylinder head or engine block. However, they DID make copper radiators a while ago, but went to aluminum units with plastic tanks because they were CHEAPER to make (as I pointed out above). Why do I know this? I was a service manager, and to go with my MCSE and MCP+I certs, I also have this zippy ASE cert for automotive parts (retail sucks tho, and computers are a lot more fun).

[clutch 1]

Oh, and on those sites you list for all of these tests, I have seen at least as many showing that copper was the way to go. Personally, I believe in what I have seen, and copper kicks aluminum's a$$. I have worked with both metals in the automotive industry (which was rather nice that you brought that up, thanks ) and copper was always the way to go for radiators and heater cores (which their sole job is to get rid of heat as fast as possible and in large volumes), but aluminum is cheaper to work with and in the end it came down to profit margins.

[BladeRunner 0]

Well, I guess I could use some of my many free web spaces and then give you a link to it.

There I'll just put the results of my tests

 

When I was cooling my PIII 1Ghz nothing can top the Thermalright SK6 with a Delta fan.

The Delta fan was simply too noisey, so I replaced it with a 60MM YS-TECH fan instead.

The result, a slight raise in CPU temperature but still cooler than any of the other units I tried.

I tried a lot and I mean a lot of coolers/HSF's.

Due to then having no real case cooling I was getting up to some really high temperatures, the SK6 was the only unit that gave me enough cooling to keep things nice and safe.

 

I'm back to aluminium now, stock cooling for my P4 (Which by the way is excellent) but if I can source me a copper HSF for my P4 I'll get it.

[clutch 1]

Welp, I am not sure what APK is driving at here, but at good old Tom's hardware, THEY even tell you that copper is better:

 

Quote:

At first glance it is obvious that silver and copper have the best conductivity. As silver is too expensive for a massive heat sink, the manufacturers at best use copper. Using aluminum is cheaper but the conductivity is by far not as good as copper. That means silver and copper spread the heat very quickly. These two materials offer the best quality for a high cooling effect in a heat sink/fan system.

 

 

Also, this testing was done a year ago, and copper handling has improved dramatically since then. So, what am I missing here? Year old benchmarks with out-of-date heatsinks and design issues that were probably the limiting factor, and not the material itself. If you are a diehard Aluminum fan, then that's great. But I got my copper cooler (with Delta fan) for about $25, which is a helluva lot less than some of the others, and cools my P3 933@1084 to 29*C at idle, and 41*C under full load when gaming for an hour or so. After seeing performance like this, I will continue to use copper until something better comes along at a comparable price point.

 

[three 0]

as far as i know about copper, it can't be extruded like aluminum can there for has to be machined from a copper block, that's why you see most copper hs out there are manufactured in plates then attached onto a copper plate, this saves on the costs. the company i work for had 5 machines that were used to manufacture from one single copper block, but my boss decided against it due to the costs involved.

 

as for clutch's remark about copper for the automobile industry, i agree..and also like to add that weight was probably another consideration.

[clutch 1]

With the switch to plastic tanks from all copper units, that was a pretty good weight loss indeed. Unfortunately the plastic tanks tended to leak a lot in refurbished/remanufactured units, but the industry was going forward with them anyway.

[three 0]

well alec i hope you weren't refering to me cause i never said you said that...anyways..on a side note...i always thought the best heatsinks were the copper/aluminum combo? the ones with the copper inlay and the aluminum fins... they were always the winners on hs roundups...

[clutch 1]

First, calm down. I, like three it would seem, am getting the impression that you are irritated. If you are irritated, fine, be so with me. Somehow though, I think it will only get worse after you read this...

 

 

Second, here is the quote about automotive components that you made:

 

Quote:

Copper is a good conductor, good at drawing heat, but due to its density (much greater than aluminum)? It HOLDS HEAT WAY LONGER! This is why aluminum is used in engine block heads also, not only lighter weight, but GREAT HEAT DISSIPATION!

 

Now, this is where I based my comment on there not being any efficient way to make any large solid components out of copper in vehicles due to machining complexity and weight. Also, it's deformation and expansion properties would probably not fall inline with that of iron or aluminum. Now, your heat dissipation statement is incorrect in that copper has been used for YEARS for cooling structures in automotive and industrial applications, but aluminum is cheaper, (and Tom's Hardware, in that very same article you linked to, even stated that Copper is obviously better at this) lighter, and easier to work with. Copper is a fairly soft material that can lose shape easily, hence a lot of the aluminum shrouds that you see on CPU coolers.

 

Now, if you *must* see what peeps do on high end cooling systems, check out another review on that lovely site:

 

http://www6.tomshardware.com/cpu/00q4/001221/vapochill-04.html

 

Note all of the copper tubing and components (just like the ole fridge at home, eh?) that have been used. Do you think they would use it because it is *slow* at exchanging heat? Aluminum fins are normally used because copper ones bend WAY too easy and close up, thereby rendering airflow in affected regions null.

 

Also, as for the laws of physics, I already stated from a well respected engineering "bible" that copper was clearly superior to aluminum in heat dissipation. But in any case, let the site listings begin:

 

http://www.hardocp.com/reviews/cooling/roundup1101/

 

If you note, almost all of the major high-end players at *least* have copper base plates (thanks to Alpha kicking that off; they KNEW that copper was a better conductor for heat but it's very soft to work with for the large fins their designs call for) and have revamped how the copper connects with other materials (either more copper, or aluminum fins/pins).

 

Here's a site I found doind a search for Copper vs Aluminum on Google that managed to get 3 HSF units from the same company that appeared to use very similar (if not the same) dies for production, but a couple had copper parts. Wanna guess who wins?

 

http://www.overclockersclub.com/volcanoshootout.shtml

 

As for the winner of LAST YEAR'S roundup at Tom's? Well, funny you should ask, guess what material they are going to?

 

http://www.swiftnets.com/mc462.htm

 

Here's a nifty quote from them which I am sure you will enjoy:

 

Quote:

The 9 in2 copper base is largely responsible for the MC462's uncanny ability to rapidly spread AMD's and P4's highly concentrated heat source through the pins. It is also responsible for this heat sink's hefty weight: 20oz (560g) without fan.

 

So, now we have the primary reason to use it in the first place (the RAPIDLY spreading of heat part) and a major reason NOT to use it for quite a while now (the weight part). But "weight" (punny, I know ), there's more!

 

Quote:

Crash test procedure: Barebone equipped with an Asus A7V motherboard, our favorite Duron 650, and the MC462 heat sink without fan. The only protection between processor and heat sink was a piece of paper masking tape (only because we didn't want to damage the heat sink's soft copper surface!).

 

Here's another one from a Google search, this time showing all the elements I spoke of already:

 

http://www.overclockers.com/tips188/

 

Now, here's a quote of interest from this site:

 

Quote:

So there it is. If we only care about performance, silver is the winner. If it's all about cost, aluminum is the choice. The best price/performance is clearly copper, but it's a bit harder to work into useful shapes.

 

If you look at the last email sent in at the very bottom, and Mr.Darnell wrote in claiming that the reason you see aluminum fins on automotive radiators is because it can dissipate heat to air faster than copper, yet that's not entirely correct. We already know that copper can dissipate heat quicker regardless of the surrounding medium due to the wonderful chart on that page (I saw a few from that engineering book I mentioned earlier), so why would he have this opinion? Simple, you can get more than enough surface area using Aluminum to offset its performance delta with Copper because you can get away with thinner fins. Thinner fins means more fins, and more fins means more surface area. This is why you see those Aluminum fins, and this goes along with how I mentioned that Copper bends really easily. Thin Copper wouldn't be all that great in an application where large bugs come crashing into it, now would it?

 

The *other* reason I mentioned why they haven't been in production until recently; the material is *soft* and difficult to work with. Here's a little comparison for you; do you own a mountain bike? I do, a Cannondale CAAD2 which is all aluminum. At one time, aluminum bikes were unheard of because they were difficult to build due to the high degree of training needed to handle the tubing. However, the airline industry had a major shake up in the mid-eighties, and almost overnight big-name makers (well, Cannondale always made aluminum bikes, but they became a helluva lot cheaper) like Trek, Paramount/Schwinn, Giant, etc starting releasing Aluminum frames in their top-end lines. Did the laws of physics change to accomadate this new ability to create these frames? Nope. Highly skilled workers from the airline industry made it to tubing fabrication plants and were able to make more frames with less defects (and less warranty claims). Over a couple of years, the pricing came down as the manufacturing process was streamlined. That's what you are seeing here; the manufacturers want to take advantage of copper, but the first runs may have not been the absolute best.

 

I have shown you that the element is clearly superior from an engineering manual, Tom's hardware claimed it was clearly superior (a quote from your link), major manufacturers are using it in water cooling and peltiers, and the automotive industry used it for years as a large volume cooling system components until they realized they could do something almost (yes, *almost*) as good with aluminum and plastic, but cheaper. Now, is this enough? I really, really think I covered this topic, and was trying to avoid doing this. But, since you seem so, um, passionate about it I thought I would fully describe my reasoning.

[clutch 1]

Dude, did you read my post?

[clutch 1]

Are you nuts? You are the reigning king of eternally long posts, and now YOU don't have the time to read one that's full of all kinds of information that YOU asked for? Whatever man, I answered your question in spades, and now you are just commenting for the hell of it. If you are too busy to read before you respond, then I must have overestimated you. Topic over.

[three 0]

there will probably never be an end to this debate and personally i think those tests aren't fair when comparing materials simply due to the fact that the design of the heatsink itself plays a very important role. what would be interesting to see and end this debate is to see 2 heatsinks made of the exact same design with the exact same fans on the exact same testbed, with only one exception, the material, one being pure aluminum while the other being pure copper. that would definetly end this debate.

[enio 0]

bah, best thing to do is to "reinforce" coolers u buy. Them itself or like i did, i drilled two big holes on the side of my tower and put two gigantic fans i assembled myself(one 220v other is with adaptor, posibility to change fans speed) and let me tell u that from then i call my rig "IGLOO" in true sense of the word !!! Temps like 23c/24c full load around 30 are no longer a wish(hehehe overclockers would apreciate this).

[BloodRedDragon 0]

Quote:

there will probably never be an end to this debate and personally i think those tests aren't fair when comparing materials simply due to the fact that the design of the heatsink itself plays a very important role. what would be interesting to see and end this debate is to see 2 heatsinks made of the exact same design with the exact same fans on the exact same testbed, with only one exception, the material, one being pure aluminum while the other being pure copper. that would definetly end this debate.

This is very true. The variations in the design of the heatsinks and the fans used make it impossible to truely compare the materials. Also, your idea is good, but it probably wouldn't end the debate. It's likely one design would not effect both materials equally, and that could unfairly bias the results toward one type of material.

K

[EM 0]

Quote:

No! Was WAY WAY too long, & I hit those links & they are WAY Slow... get me a page with a graph!

* I am busy on something here. I cannot spend alot of time reading this evening... got a graph of anything like that page at Tom's Hardware page does, showing what I have been asking all nite?

APK

After reading through this whole LONG discussion, this particular question came to my mind. Why is it that Mr. AlecStaar has no time to read the presented evidence yet he has seemingly unlimited time to write LONG posts in response to the evidence presented? Just wondering.

As for myself, I use a water cooling setup with a copper radiator and copper waterblocks. My reasons for using copper in the waterblocks is to help in the prevention of corrosion with dissimilar metals and the radiator was already copper. As to which is better? I have no idea, just buy whatever works best for you. I like the watercooling because it is MUCH quieter on my dual system. Maybe a couple of Dragon Orb 3's with the 7500 rpm fans would cool better but the freaking noise would drive me even more insane than I already am.