Need suggestions Copper or Aluminum

[BillA]

Originally posted by bigben2k
Interesting indeed...

because it shows that copper is best at at least 10mm bp thickness, for most flow rates, where Aluminium is best at 10mm OR LESS, for flow rates above 3 gpm (180 gph).
. . . .

come on bigben2k, pull it together

a lower T/W (like "C/W") is better
take another gander at those graphs, apples to apples eh ?

when maskedgeek stops making assertions that he cannot substantiate
"a good waterblock will perform the same in Aluminum as in Copper.... " (in his initial post)
then he won't get banged on

he has this same thread on a number of forums where he's singing the same song
- this is the only place I'm telling him how ignorant he is
(as have others - but he hears only that portion convenient)

[bigben2k]

BillA: The way I see it, I agree that MaskedGeek has some misconceptions, and that he doesn't have accurate enough data/testing to prove that Alu would perform at par with Copper. I know better, you know better, and I would think that most people know better too.

The point is that Alu does have its place. It's not useless, it's not a waste of time and/or money, but there is a disadvantage, in that it will *never* perform as well at Copper, it's just theoretically impossible, and as testing would show, the truth.

I'm glad that you're not following him everywhere to put him down.

Anyways, this thread should get back on topic. Now that we know the truth about Al vs Cu, would anyone be willing to buy an Aluminium waterblock, knowing that its purpose is to get you watercooled, so that your computer can run quietly, but with the understanding that it will not give you the ability to overclock your computer nearly as well as a Copper block?

[Brians256]

Well, I would like to answer the question in two ways. First, I would not buy an aluminum waterblock on a high-end CPU. On a cooler CPU with less wattage (such as a small multimedia box like the Shuttle SS51) I might buy aluminum. It depends upon the price differential and also on the maintenance differential. If Aluminum is 1/2 the price.... OK. If Aluminum is much easier to maintain (because you don't have to change the water as often because of less galvanic corrosion), then yes.

Having said that, I would like to follow up by saying that I made my first three waterblocks in aluminum. First, I got the aluminum for free (scrap from the machine shop here at work). Second, I don't know how to machine very well and aluminum is MUCH more forgiving. If you are doing it yourself, learn on aluminum, then join the big boys.

Maybe when I get some free time I'll make my own copper waterblock. Too many pots on the stove right now. I certainly won't buy one if I can make it though, and I could.

As for making it a consumer product, unless you can show a considerable price differential (and I don't believe you can), stick with copper. Watercooling hobbyists are performance junkies and second-rate solutions just won't fly.

[morphling1]

You've got to ask yourself who is going for watercooling in the first place, and the answer is performance crowd looking to get the most out of cpu, and believe me because I'm not just designing block, I'm crazy about overclocking too, and I know that even 1 degree can make a difference geting theat few more MHz, also I just cant see how you would cool new t-bred with its 84mm2 of surface area overvoltage and overclocked, Al just can't spread heat quickly enough to larger area, so you would need to compensate that with very fast flow, with those cpu-s the difference would be a lot bigger then 2-5 degrees. I can't say that from real test but if I just make few simple calculations, the difference between copper and Al is very noticable.
So I see the main problem with Al , that it just can't handle very small cpu cores, It would be good for northbridges and gpu-s where all those chips are bigger and you have foot print of the block the same size as the chip so you can have very thin base thickness for very good cooling.

[maskedgeek]

i see were yoru comming from on the small dies... but still from tests iv seen aluminum blocks do the same or a LITTLE less than copper, like webmedics crystal coolers, he did silver inlayed copper copper copper inlayed aluminum and aluminum and he says they all perform the same...

[bigben2k]

I think that some people turn to watercooling for the quiet aspect, but I'll agree, there's probably fewer of those.

[maskedgeek]

ok guys as of now im gonna use aluminum and if i see a substansial(spelling?) diff between aluminum and copper ill switch over or do everything but cpu blocks in aluminum and use corosion inhibitors.....

[morphling1]

Better use anodization, that way you can have some nice colored block that will appeal, and forget about that galvanic corrosion

[maskedgeek]

Originally posted by morphling1
Better use anodization, that way you can have some nice colored block that will appeal, and forget about that galvanic corrosion

yeah thats exactly what im thinking, black and the corosion i saw my corosion i did to a popcan block with no corosion inhibitors, it wasnt bad but it looked like salt build up for some reason, so anodizing it will stop corosion rite? and if i build a complete system all aluminum i wont have corosion rite?

[morphling1]

Yes, and yes of course

[maskedgeek]

ok just making sure cuz steel will corode and rust with just water, itll do it anyway but it does it with water, just making sure on this, would it be possible to annodize a heatercore also? or would it be too much to do? i really dont know anything about anodizing or anything i gotta figure out if i can get it done around here...

[morphling1]

The heater core would be a little trickier I mean inside, the interior should be completly filled with electrolite, and if you have some air pocket that means that part won't be anodized...

[jaydee]

Ok I am going to jump in here and try to throw out some logical thoughts about using AL over Copper or Copper over AL.

First of all I have run high wattages and have seen others run high wattages on T-Birds and XP with great success on AL blocks. Now I am not going to tell my temps as some fools claim they are wrong but whatever. I don't care about that. I know I have done it and others that have personally seen me do it know it is possible. If I can get a 305mhz OC at 2.20Vcore on a T-Bird1400 with an AL block then it is more than capable of being used. BUT that was with the T-Bird core which is larger than the XP and the T-Bred. The point morphling1 brought up about AL not distributing the heat fast enough I don't quite buy (or maybe I do), my theory is AL doesn't absorb the heat fast enough. Which I guess maybe the same things in a different view and I will aggree. Especially with smaller core surface area's. BUT, when AMD comes out with the heat spreader this may all change again and AL may gain a little ground back in that area. I totally disagree that 1C will get you any more mhz as I have been OC'ing since 25mhz DX comps where out and 1C never made any difference. In fact 3C never made any difference. But 4C on up can and I have definatly noticed it in the Durons 600/700/800/850/900/950/1000 CPU's I have used and the T-Bird 1400 and the XP1600+ in the last few years. 4-5C seems to be the spot that I noticed you can gain some extra mhz. Of course it isn't enough to make any difference in anything exept stats junkies and mad onion wannabee's (not that there is anything wrong with that). I used to be a OC crazy fool to and wanted every last mhz I could pull out of the system. But then I got to thinking WTF am I really trying to gain here. 100mhz on a 1400mhz chip really doesn't do anything noticable. The faster CPU's get the less OC'ing is worth the effort IMO. But that is my thinking and people like morphling1 still like to do the heavy OC'ing which is just fine! I have tried to say that different designs will react different in Copper and AL and the temp swings will be higher or lower. No one ever reads that from me I guess and just continues with the "you are a idiot" BS and try to twist everything around. I have noticed that larger wider channles seem to have less of a temp swing from AL to Cu than smaller tighter channels do. And that would backup the theory that AL cannot transfer the heat fast enough as AL cannot take advantage of the smaller channles as the heat will not get that far, but Copper can take advantage of the smaller channels and spread the heat to those channels in time for the channles to be usefull. This is why I belive morphling1's multi channel spiral has such a huge temp swing from AL to Cu and my large wide channel e design has less of a temp swing. It just makes since and I think that cannot be argued or disagreed upon.

So if you are going to make a AL block make a large channel one with lots and lots of open area for water. I have a center inlet duel outlet design that would be the only one I would use that I would claim could perform well enough to go into the performance range of blocks up ther with Copper. My first test really impressed me and I have already redesigned it to make it more efficent. Being it has a big center inlet right over the core of the CPU the AL wouldn't really need to transfer the heat anywhere else but up and that is my theory is to why it works well. Not to mention you are dumping a LOT of water right over the core. That definatly helps. If I could make this block in copper it would even perform that much better but that is a lot of milling as there is a bunch of pins milled into it.

So now lets get into the price area. Lets say you are going with the standard .5" thick 2" wide 3" long base with a 1/4" top.
AL prices:
$5.62 a foot for .5"x2" material.
$2.75 a foot for .25"x2" material
------------------------------------
$8.37.
Now you can make 4 blocks per 1 foot of material at the given specs above. That6 comes out to $2.10 per block just for the base and the top unmilled.

Cu prices:
$18.55 a foot for .5"x2" material
$9.42 a foot for .25x2" material
---------------------
$27.97 per foot or $7.00 for one block for the base and the top unmilled.

Now that is for the block material, where the hard part comes in is the milling costs. As anyone with any milling experience know Copper is considerably harder to mill than AL and the tooling costs are about tripple because you can use standard HSS endmills for AL which are about $5 for a double sided one and you can make about 5 decent blocks with one double sided one. But with Copper you really should use the carbide endmills which run $25 for a cheap one sided one that you can get about 5 blocks out of one (depending on the complexity of the block ect..).

So lets break that down a bit more. $25 divided by 5= $5 a block. So take the $7 for material and add the $5 for tooling and get $12 for the milled (basic) block.

And take $5 and devide that by 5= $1 a block and add that to the $2.10 and get $3.10 for a milled block.

Now you have to add in the rest of the variable such as YOUR time, power consuption, machine wear, shipping for tooling and materials, the barbs, the mounting hardware, ect..

So basically you got a $8.90 difference just for the material and the milling. The mounting hardware depending on what you want to use can be from $2.50 from standard bolts, nuts, washers, springs on up to whatever.

People may think wow only $8.90 difference, but you have to consider the wear on the mill and it takes about 2-3 times longer to mill a block in Copper than AL. That adds up big time in equipment wear and the power bill.

Also one other MAJOR cost to AL is you have to get it anodized!!! This is not cheap and definatly closes the gap of the $8.90 difference. This alone may make Copper the better choice even price wise when making blokcs to sell.

Also note these are just basic numbers they will be more complicated!!!

So in the end with having to anodize AL the price difference in making a Copper over AL really is not that far apart!!! So in conclusion if you have a mill that can handle copper with no problem then do them in Copper!

I argue that AL is worthy and I belive it is as I seen it do incredible things, but on a manufacturing/marketing level I would have to lean towards Copper for performance blocks.

The only reason I really use AL only is because my mill cannot handle extended running times as it has a non cooled high tourque DC motor that gets real hot after about 30mins. 30mins is not long enough to mill a good copper block. Unfortunatly I cannot find a better motor as the way it is mounted just will not fit a normal AC motor as they are to wide. Also I Find AL to work just fine and if I can get a 305mhz OC with it I see no reason to change, at least not yet.

I was once going to attempt to sell blocks and have been crunching numbers the last few weeks ,while I had nothing better to do in the hotel room on the road trips for work, and decfided to drop the idea all together. It just isn't worth it to me. I would make about $5 profit a block (providing they sold) and that is not worth it, not even close. That means I would be making about $2.50 an hour. To hell with that! I make $17.50 an hour at work. So in short I am not going to bother with it. Just isn't worth it, not to mention all the other BS you have to put up with with these accuracy fools (not to say that they are real fools BTW as they are not) who have nothing better to do than worry about how far off you are on you results. Just isn't worth the hastle even if they find it to perform well.

Another idea I though about was selling DIY water block making kits. You would basically supply everything (monting hardware, bottom, top, barbs, ect...) but you would would not mill the block. All you would do is drill the mounting holes and you would let the customer put his own design in the block. You could sell them in both AL and Copper. Not sure if they would sell but who knows!

But don't listen to my Pansy Ass! I know nothing!

[bigben2k]

Actually, since you're already so far ahead of us blockmaker wannabes, why don't you try selling an Alu block making service? I mean, let the customer give you the design (you're not responsible for the performance that way), plus, if he doesn't give you a detailed measurement drawing, then you can charge for that service too.

You never know, you might even get DangerDen to ask you for some prototypes.

I really like your cost analysis. It would be nice to see the power consumption for Al vs Cu. The bit cost has to be a fair amount too.

I don't believe that an Al block NEEDS to be anodized, as long as the other components are Aluminium too. You could even put together an Aluminium watercooling kit.

1-Alu prototyping service, with optional draft svc.
2-Alu wc kit (sell it on demand only, see how it goes).

[pHaestus]

these accuracy fools

If only there were engineering professors with this attitude! My GPA would have skyrocketed!

Prototyping for others could get messy as the rights to make blocks would be compromised. Not the sort of thing you can really patent after all, and whoever codes the CNC would have possibly as good a claim as who came up with the idea.

[bigben2k]

Then add an NDA to the deal.

Block designs that are to remain private, should stay so.

The problem is that if JD starts selling blocks, he won't be able to use a design that he's done for someone else, at least not exactly the way he did it. This NDA would have to include an honest statement by which JD can be upfront and have it read: "I may sell some blocks someday, and if I do, I might use your design to inspire me, but I won't reproduce your design exactly".

Of course the NDA is optional...

(NDA=Non-Disclosure Agreement, aka copyright, technology sharing agreement, or whatever other terms there are out there.)

Alternatively, this agreement could be extended to a licensing agreement, by which JD pays the requestor a fee, either per block, or as a one-time fee, if he ever decides to sell a block with it.

[jaydee]

I apoligize for using the term "these accuracy fools " I didn't mean it in a bad way although it really did sound like it. As much as I disagree with precision testing I still do not think you all are real fools.

I just don't know about the anodizing. I think it would be a must as a marketing ploy if nothing else as everyone thinks it is a must even if it isn't. I don't think I will get into selling water cooling stuff. Just not enough money in it. That may sound selfish but I have to maintain a living for myself and my daughter aswell.

[bigben2k]

Originally posted by jaydee116
I apoligize for using the term "these accuracy fools " I didn't mean it in a bad way although it really did sound like it. As much as I disagree with precision testing I still do not think you all are real fools.

I just don't know about the anodizing. I think it would be a must as a marketing ploy if nothing else as everyone thinks it is a must even if it isn't. I don't think I will get into selling water cooling stuff. Just not enough money in it. That may sound selfish but I have to maintain a living for myself and my daughter aswell.

I know what you mean.

I've got two step kids (teenagers) and one of them is both physically and mentally handicapped. They have a deadbeat dad, so my dreams of starting a business are on hold, for a little while.

[jaydee]

Originally posted by bigben2k
I know what you mean.

I've got two step kids (teenagers) and one of them is both physically and mentally handicapped. They have a deadbeat dad, so my dreams of starting a business are on hold, for a little while.

I am looking into other areas for bussines aswell. But for now my regular job will have to do. As for water cooling stuff I also want to see how things play out with future CPU's. I am pretty certain they will not sell anything that cannot be air cooled, but I am also concerned about mounting and the possibility of then putting a stop to overclocking which they can do at anytime if they feel the need to. The future just isn't clear enough to just jump in and try to make living off of water cooling. I also think at some point strait water cooling will slowly be phased out in favor of active cooling depending on the future CPU's. And real extream OC'er is probably already getting into phase changing stuff as it really doesn't appear to be as hard as people think. My engineer budy does this stuff for a living and said he could show me how to do it in a few minutes. Of course he does this for a living, but the more things catch on the more they are used and older stuff is phased out. Things are just to uncertain at this point.
I don't know about you all but I know enough about business to not get involved with something that is not even close to predictable.

[gone_fishin]

jaydee116, I am trying to follow your logic.

quote
" I totally disagree that 1C will get you any more mhz as I have been OC'ing since 25mhz DX comps where out and 1C never made any difference. In fact 3C never made any difference. But 4C on up can and I have definatly noticed it in the Durons 600/700/800/850/900/950/1000 CPU's I have used and the T-Bird 1400 and the XP1600+ in the last few years. 4-5C "
end quote


If 3C never made a difference for you but 4C has, then this is a 1C threshold that you are describing that once crossed, you believe will give a better overclock. My question is how do you know you will never be already at this threshold and 1C more is all you need to go higher? If in fact you are sitting at this threshold with an aluminum block, then a copper block would put you over it and get the better overclock.

[myv65]

Originally posted by jaydee116
I still do not think you all are real fools.

Nah, we're "theoretical fools".

Sorry JD, couldn't resist.

[jaydee]

Originally posted by gone_fishin
jaydee116, I am trying to follow your logic.

If 3C never made a difference for you but 4C has, then this is a 1C threshold that you are describing that once crossed, you believe will give a better overclock. My question is how do you know you will never be already at this threshold and 1C more is all you need to go higher? If in fact you are sitting at this threshold with an aluminum block, then a copper block would put you over it and get the better overclock.

Good point, I never looked at it this way! I was looking at it from starting at a set temp. For example 40C. 39C, 38C,37C, and 36C would pose no gain in mhz but 35C would. So yeah from that perspective I stand corrected. I never used different water blocks to change temps though. I always used other methods so maybe that is why I wasn't looking at it from that angle.

I find that cooling the water itself as oposed to making a better block seems to work out rather well for overclocking porposes. If I was to get back into more extream overclocking I would not bother with trying to make a better block that gaines a C or 2, I would work on cooling the water instead. Seems much more logical to me than wasting money on dozens of blocks to get a few C cooler when cooling the water will drop you many many C with the same block.

Maybe I am off base there?

[gone_fishin]

Originally posted by jaydee116
If 3C never made a difference for you but 4C has, then this is a 1C threshold that you are describing that once crossed, you believe will give a better overclock. My question is how do you know you will never be already at this threshold and 1C more is all you need to go higher? If in fact you are sitting at this threshold with an aluminum block, then a copper block would put you over it and get the better overclock.

Good point, I never looked at it this way! I was looking at it from starting at a set temp. For example 40C. 39C, 38C,37C, and 36C would pose no gain in mhz but 35C would. So yeah from that perspective I stand corrected. I never used different water blocks to change temps though. I always used other methods so maybe that is why I wasn't looking at it from that angle.

I find that cooling the water itself as oposed to making a better block seems to work out rather well for overclocking porposes. If I was to get back into more extream overclocking I would not bother with trying to make a better block that gaines a C or 2, I would work on cooling the water instead. Seems much more logical to me than wasting money on dozens of blocks to get a few C cooler when cooling the water will drop you many many C with the same block.

Maybe I am off base there? [/b][/quote]

No, you're on base there if your intent is for making something for your personal machine but the research and effort could pay off in the marketplace if you want to sell a lot to the public.

[Brians256]

From a practical standpoint, I quite agree with you, Jaydee. Agonizing over a 1C differential is nearly a waste of time. However, the reason that many of us DO spend considerable energy upon it is for two reasons: 1) we like to understand all the ramifications and intricacies (the theory, if you will), and 2) we are raving mad enthusiasts.

Despite the urge to gain that last degree of cooling, active cooling is indeed the real answer, and phase change is probably the best bet for now.


I still want a copper waterblock though.

[jaydee]

Originally posted by gone_fishin


No, you're on base there if your intent is for making something for your personal machine but the research and effort could pay off in the marketplace if you want to sell a lot to the public. [/b]

Well I agree to a point. But would it not be worth looking into constructing a water chiller for the marketplace? Seems to me(and I may be wrong) that the extrream overclocker would bite on such a device. I guess the biggest problem is making it reasonable priced.