Zalman 7000Al Cu or 7000Cu?

Cooling Processors quietly

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
segask
Posts: 9
Joined: Tue Apr 13, 2004 8:21 pm

Zalman 7000Al Cu or 7000Cu?

Post by segask » Tue Apr 13, 2004 8:39 pm

I read SPCR's review of the 7000 and while it shows the AlCu and Cu versions performing identically, the review speculates that the Cu would probably do better in a hotter environment or a hotter runing cpu than was used in the test setup.

I noticed in all the threads on these forums that everyone recommends the AlCu and not the more expensive Cu version. Which version would be better for a mobile Barton 2500? I would run the fan in silent 5V mode when websurfing and using WinTV tuner. I would like to overclock a little bit when running the fan at 12V full speed. It will be in an Antec Sonata.

Do you guys recommend the AlCu version or the Cu?

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Tue Apr 13, 2004 9:06 pm

Most people recommend the AlCu because it's cheaper, it's lighter (socket attachment issue), and because it performs within ~5% of the all-cu version.

Still, I'd go with the all-cu for my box. The CPU is the biggest heat dissipator, so even 5% improvement in efficiency is worth it to me..

DGK
Posts: 70
Joined: Sun May 18, 2003 3:59 pm

Post by DGK » Tue Apr 13, 2004 11:25 pm

I have a Athlon 64 3200+ which I use a 7000 AlCu on. I keep the fan on it at its lowest setting and I have been very happy with its performance.
My recomendation is to go with the AlCu because of its much lighter weight, lower cost and nearly the same performance as the Cu version.

Ralf Hutter
SPCR Reviewer
Posts: 8636
Joined: Sat Nov 23, 2002 6:33 am
Location: Sunny SoCal

Post by Ralf Hutter » Wed Apr 14, 2004 5:38 am

I'd also go with the ALCu. Unless your CPU temps are already running way too hot there's no reason to hang a heavy lump of copper off your mobo, just to possibly save a few degrees.

segask
Posts: 9
Joined: Tue Apr 13, 2004 8:21 pm

Post by segask » Wed Apr 14, 2004 9:41 pm

thanks for the replies guys :)

so its 3 votes for AlCu and 1 for all Cu version.


...question about the heavier weight of the Cu version: is there a danger of it actually cracking an athlon xp motherboard where it mounts around the cpu socket? I won't be using it in a lan case or anything like that. The rig I will put the 7000 in will stay put in one place on the floor next to the desk, but can these big heavyweight heatsinks like the zalmans and thermalrights cause damage simply because of their weight?

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Wed Apr 14, 2004 9:49 pm

Unlikely, unless you move your computer around.. vigorously.. a lot. That said, the Zalman things attach to the plastic socket in a rather unconventional manner. They don't use the 4 hole mount technique of a lot of the heavier heatsinks (the thermalright sp94, for example).

loren_brothers
Posts: 230
Joined: Thu Mar 25, 2004 7:11 pm
Location: Friday Harbor, San Juan Island, WA, USA....Just West of B.F.Egypt

Post by loren_brothers » Wed Apr 14, 2004 9:56 pm

...question about the heavier weight of the Cu version: is there a danger of it actually cracking an athlon xp motherboard where it mounts around the cpu socket? I won't be using it in a lan case or anything like that. The rig I will put the 7000 in will stay put in one place on the floor next to the desk, but can these big heavyweight heatsinks like the zalmans and thermalrights cause damage simply because of their weight?
The MB manufacturers don't make those specs just because they are sitting around being bored. :lol:
The possiblity always exists. You have to realize that it is almost 2 lbs. you will have hanging there! If nothing else...it will give it "stretch marks". :lol:

But seriously: any violent change in aspect in the position of the case could result in 'motherboard disfigurement'. But if zero mass exceleration is the rule of the day then you should be fine. Just be careful when you do the spring house cleaning!

Despite what wumpas had to say; it would only take a good jolt to crack something. And once (not "alot") is all its gonna take to ruin your day. 2 lbs is a lot of weight to have hanging off (at 2" perbendicular to the stress bearing plane) of a piece of compressed composite. If you check you will see that the Cu version will exceed the recommended weights by at least 70%, and that is far beyond any engineering "safety" tolerances usually built into product specs.

:wink: Beware of shear factors! :wink:

silvervarg
Posts: 1283
Joined: Wed Sep 03, 2003 1:35 am
Location: Sweden, Linkoping

Post by silvervarg » Thu Apr 15, 2004 12:51 am

The MB manufacturers don't make those specs just because they are sitting around being bored.
Max CPU heatsink weight values are set as they are to shipping reasons.
If you stay under the limit you have a very good chance that the computer survives UPS shipment and handling.
If you exceed that weight just be a little carefull when moving the computer to the LAN party.

The 7000Cu version should in theory perform better than the AlCu version. For some resons it seems to perform the same as the AlCu version in reviews. My conclusion is that the AlCu version is at least a lot more cost effective, so that would be my choise if I went for a Zalman 7000 cooler.

NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

It *does* use the four holes on the Athlon XP

Post by NeilBlanchard » Thu Apr 15, 2004 2:56 am

Hello:
segask wrote:thanks for the replies guys :)

so its 3 votes for AlCu and 1 for all Cu version.


...question about the heavier weight of the Cu version: is there a danger of it actually cracking an athlon xp motherboard where it mounts around the cpu socket? I won't be using it in a lan case or anything like that. The rig I will put the 7000 in will stay put in one place on the floor next to the desk, but can these big heavyweight heatsinks like the zalmans and thermalrights cause damage simply because of their weight?
Another vote for the Al/Cu version!

BTW, it comes with various mounting brackets -- on the Athlon XP, either one requires that you have the 4 holes for mounting. It uses the two shallow 'V' aluminum brackets (one is blue and the other is clear anodized). On an Athlon 64, it uses the backplate and the two brass knurled cylinders through the two holes, and (I'm guessing) that with a P4 it uses the stock frame with the 2.5" aluminum yokes.

FWIW, the Al/Cu version performed slightly *better* in Mike C's review! :o Some of us surmised that maybe the heavier copper unit *pulls* itself away from good contact with the CPU *because* of the extra weight?

burcakb
Posts: 1443
Joined: Tue Mar 09, 2004 9:05 am
Location: Turkey

Post by burcakb » Thu Apr 15, 2004 4:48 am

I have the Cu version and I'm wishing I had the AlCu. The weight is INCREDIBLE !!. I was thinking, if whats a few 100g? WRONG. IT IS HEAVY.

As for performance, I run my stock Barton 2500+ @ 1.5V and with the Zalman @ 1400 rpm, my CPU temp is 49C when folding.

PhilC
Posts: 19
Joined: Fri Apr 02, 2004 8:47 am
Location: San Diego

Post by PhilC » Thu Apr 15, 2004 8:50 am

I would go with the ALCU because of the CU's weight issue. Also there is the fit issue with either of these coolers. The Zalman website has compatibility info for mobos etc. It's important to keep in mind there are two factors concerning fit: 1. Does it fit on the mobo. 2. Will it fit in the case.(Verticle clearance under the PSU) I believe both are the same size, the difference being weight only.

I have the ALCU on an Intel D875PBZ board in a Sonata and it clears the bottom of the PSU by about a half inch.

PhilC

loren_brothers
Posts: 230
Joined: Thu Mar 25, 2004 7:11 pm
Location: Friday Harbor, San Juan Island, WA, USA....Just West of B.F.Egypt

Post by loren_brothers » Thu Apr 15, 2004 10:15 am

silvervarg stated:
Max CPU heatsink weight values are set as they are to shipping reasons
I think not! MB's are not shipped from the factory with Heatsink/fan combos in place. I have never purchased a MB that had one in place. When drawing up the mechanical specs for a MB the engineers look at and test failure rates incurred by various weights applied to the composite board. They do this with a stress test machine under tightly controlled conditions. They set the maximum allowed tolerances based on the point where deflection in the board occurs. IE: if the rating is 450g then the board has shown consistant evidence of deflection/distortion at 540g [450g + 90g (20% industry safety margin)= 540g]. They don't strap a 900g weight to it and ship it via UPS or FEDEX to Hoboken NJ to test it! :?

It is also interesting to note that Zalman prominantly displays warnings at its site about the weight issue. They also are not doing this because they are bored. They do it so that the consumer won't be able to come back and claim damages after they incur damage to their MBs.

but back to the point: considering the performance differential between the two.... go with the AlCu.

PiSan
Posts: 42
Joined: Sat Jan 17, 2004 5:06 pm
Location: New Jersey
Contact:

Post by PiSan » Thu Apr 15, 2004 12:18 pm

I would also like to recommend the AlCu version. It's fairly heavy in it's own right and I would be uneasy adding another 100 grams to my processor. With my 2.6 at a high overclock I get about 50c at full load. That's with poor case airflow mind you.
The potential increase in the Cu version doesn't justify the added cost and weight IMO.

segask
Posts: 9
Joined: Tue Apr 13, 2004 8:21 pm

Re: It *does* use the four holes on the Athlon XP

Post by segask » Thu Apr 15, 2004 8:22 pm

NeilBlanchard wrote:Hello:



FWIW, the Al/Cu version performed slightly *better* in Mike C's review! :o Some of us surmised that maybe the heavier copper unit *pulls* itself away from good contact with the CPU *because* of the extra weight?
:o interesting theory. Can that kind of thing happpen over a period of time? For example, you put a big copper zalman or thermalright heatsink on and maybe after several months its immense weight causes it to pull away from the chip and your cpu temps start creeping up. Has anyone ever heard stories of this happening with the big zalmans or thermalrights?

IsaacKuo
Posts: 1705
Joined: Fri Jan 23, 2004 7:50 am
Location: Baton Rouge, Louisiana

Post by IsaacKuo » Thu Apr 15, 2004 9:09 pm

Doesn't he test with the motherboard lying flat as in a horizontal case layout?

burcakb
Posts: 1443
Joined: Tue Mar 09, 2004 9:05 am
Location: Turkey

Re: It *does* use the four holes on the Athlon XP

Post by burcakb » Thu Apr 15, 2004 11:35 pm

segask wrote: :o interesting theory. Can that kind of thing happpen over a period of time? For example, you put a big copper zalman or thermalright heatsink on and maybe after several months its immense weight causes it to pull away from the chip and your cpu temps start creeping up. Has anyone ever heard stories of this happening with the big zalmans or thermalrights?
that depends on the elasticity of metal and plastic used. All material have a limit where any force applied below the limit does no permanent deformation and the material springs back into shape. Any force level above the limit causing deformation will make the deformation permanent (plastic deformation)

The limit for any material is dependent mostly on its substance (ie type of metal etc) and its purity. Inpure = more likely to undergo deformation. Furthermore impurities have a nasty habit of "walking" through substance called creep that amplifies the problem. And guess what, creep is accelerated by, you guessed it, heat.

Comparing the steel holding clip, the plastic socket, the aluminium brackets and the composite PCB, the lowest elasticity coefficient is with the PCB i think. So the weight would deform the PCB before the metal became lax enought for the HS to fall off or lose contact. The time frame for any of this happening is beyond me. But I would guess it to be several years before getting even slightly worried about it.

This was just my "materials 101" course that I didn't have to remember for the last 12 years :)

IsaacKuo
Posts: 1705
Joined: Fri Jan 23, 2004 7:50 am
Location: Baton Rouge, Louisiana

Post by IsaacKuo » Fri Apr 16, 2004 4:10 am

I don't know the time frame, but I think pretty much every socket 478 motherboard with a CPU/heatsink installed has a deformed PCB. When I installed my Scythe Heatlane Zen, I let the motherboard "rest" heatsink-less for a day before installing it, just to let it flatten back out. I'm sure it's curved again by now.

Farnsworth
Posts: 50
Joined: Fri Apr 16, 2004 4:11 am
Location: Belgium

Post by Farnsworth » Fri Apr 16, 2004 4:15 am

And what about the Al only version of the heatsink?

It's very light, but how well does it cool? :?

silvervarg
Posts: 1283
Joined: Wed Sep 03, 2003 1:35 am
Location: Sweden, Linkoping

Post by silvervarg » Fri Apr 16, 2004 5:20 am

Loren_brothers:
I think not! MB's are not shipped from the factory with Heatsink/fan combos in place. I have never purchased a MB that had one in place. When drawing up the mechanical specs for a MB the engineers look at and test failure rates incurred by various weights applied to the composite board. They do this with a stress test machine under tightly controlled conditions. They set the maximum allowed tolerances based on the point where deflection in the board occurs. IE: if the rating is 450g then the board has shown consistant evidence of deflection/distortion at 540g [450g + 90g (20% industry safety margin)= 540g].
This would depend on the motherboard used, the mounting method of the heatsink, if a shim is used, the height of heatsink and the orientation of the computer when mounted (and a few more factors).
You must mount the motherboard to a case or other semi-hard surface to do any of these tests accurately. Ofcourse there is no problem to ship the heatsink separately, as is often the case for us home builders. The specifications are not made for home buildt computers, it is intended for shops that build the entire computer and send to the customer. This is the way the majority of computers are sold.
As far as warranty goes home builders are smoked anyway since the CPU warranty is usually void anyway if you mount the CPU cooler yourself. This won't stop most of us anyway, but it gives a clear indication that we are not the intended audience for the weight limits anyway.
They do give us a clue to be slightly more carefull when mounting a heavier and taller heatsink.

loren_brothers
Posts: 230
Joined: Thu Mar 25, 2004 7:11 pm
Location: Friday Harbor, San Juan Island, WA, USA....Just West of B.F.Egypt

Post by loren_brothers » Fri Apr 16, 2004 10:41 am

silvervarg:

It would be interesting to know where you come up with the basis for your statements. It just doesn't 'hold water' from everything that I've seen. And I've actually seen these tests being made. (Not to mention everything I learned as being educated as a mechanical engineer many, many, many moons ago) Those specs are determined solely from a sheet of pre-drilled composite long before any manufacture or distribution ever occurs. There is NO heatsink, fan, or anything but a metal rod, attached to the planned mounting holes, used in these tests. Force is applied to this rod to simulate the force applied by the "theoretical" heatsink/fan. The only variable in this situation is force vs. deflection of the composite. A "shipping force" variable would be impossible to quantify, much less enact upon the testing environment.

As far as warranty goes home builders are smoked anyway since the CPU warranty is usually void anyway if you mount the CPU cooler yourself.
Also, toss in some sort of background for the "void warranty" statement too.
(edit: found out what you were referring to. AMD's direct sale PIB warranty is voided if the supplied heatsink is removed, and reseller purchased OEM products are only warranteed by the reseller not by AMD)


The specifications are not made for home buildt computers, it is intended for shops that build the entire computer and send to the customer
I personally think that this basis for your premise, and hence your entire arguement, is basically flawed and unsupportable.


BUT, if you can prove (by fact, not opinion) your arguement I will very happily "eat crow" (a big, fat, juicy one even) in front of everyone here in this forum. But I doubt that my dietary habits are destined to change anytime in the near future.

Don't get upset at this.... I am just "calling your bluff" as us Americans like to say.... it is not personal in any way (I would even 'call' Ralf or Mike on this one). I just bet that my 2 pair (education, experience) will beat your queen high showing (opinion). :lol:

Enough said on my part. :D
Last edited by loren_brothers on Fri Apr 16, 2004 11:43 pm, edited 1 time in total.

Apoc
Posts: 11
Joined: Fri Apr 16, 2004 7:47 am
Location: Sweden

Post by Apoc » Fri Apr 16, 2004 1:30 pm

Acording to the swedish HW site nordichardware.se the Cu version cools slightly better under heavy load.
http://www.nordichardware.se/Recensione ... rivelse=43
It's in swedish, but the graphs should be understandable for everyone. The blue bar is average temp, while the silver one is maximum temp.

Just FYI. I'm still going for the AlCu to save the poor mobo.

The reviewer also give both models very good noise level ratings (based on listening only).

wumpus
Patron of SPCR
Posts: 946
Joined: Sat Sep 06, 2003 9:57 pm
Location: Berkeley, CA, USA
Contact:

Post by wumpus » Fri Apr 16, 2004 2:20 pm

So, under 100% load: 4c at medium fan speed and 5-6c at low fan speeds. That's actually a more decisive victory for the all-Cu than I would have expected. .

yermolovd
Posts: 61
Joined: Wed Dec 24, 2003 7:13 pm
Location: Kitchener, ON, Canada
Contact:

Post by yermolovd » Sat Apr 17, 2004 11:29 am

I vote AlCu though, AlCu is already quite heavy! It cools very well at 5V it handles my AXP1800+(Palomino)@50°C idling, about 60 on load.

Post Reply