Printable View
?? So you consider Lee/RoboTech an idiot? He made the copper die tests and he also said that they don't seem to reflect real world CPUs which have IHS installed. So I really don't understand your comment or your analogy. Could you elaborate? (lol, now I uderstand: you own Storm...)Quote:
Originally Posted by creidiki
A better analogy: 10 people tell the world is spehrical, 1 person tells it is flat but he also tells that his observation seems to be incorrect. Yes, I believe all of them.
#1 is a VERY bad way to measure the performance of a waterblock considering how finicky overclocking is. you would need to use the same CPU and alot of CPU's degrade their overclock over time. Its just not even close to being reliable, and as far from scientific as you can get.Quote:
Originally Posted by Paapaa
#2 is in theory a VERY good way, but the problem is that you need a 840EE at 5ghz (high volts too) and a HIGHLY accurate and calibrated RTD probe based system with atleast .01C resolution/accuracy. 0.001C resolution is preferable. And to be frank, i dont know of any publicly available products that give 0.001 resolution.
With a heat load of around 250-300 watts from an overclocked/overvlolted 840EE it would make the increase a bit larger than any other waterblock so measuring the dT would be eaiser, but still pretty dificult. Also, when you get up to around 3GPM for the testing, things would get even more dificult due to the dT decreasing
But in all seriousness, if you have a spare 100,000 dollars to spend on a setup like that (or however much it would cost as I have no real idea) then go for it as it would provide some VERY intersting data.
Of course it has to be done with the same CPU. But that way we would get actual data which should show the real difference between blocks. And who cares about "scientific" if the results don't have any practical meaning at all - as with Lee's results. If Apogee and Storm gave identical overclock on 10 different processors and Lee's 14mmx14mm copper simulator gives huge differences, do you still believe Storm is the superior block and everyone should "try to avoid" Apogee? Unfortunately we only have results from one single test (made by Swiftech and won by Apogee).Quote:
Originally Posted by MaxxxRacer
And the bottom line is again: if the differences (to one way or another) are that small, why would you recommend members to buy one block over another? I'm starting to believe that all who seem to reject all the real CPU data actually own Storm... I might be wrong.Quote:
Originally Posted by MaxxxRacer
The biggest problem with using overclocking is that its not reliable. Not even close. You could for some wierd reason get better temps with a Maze1 than a MP-05/Storm/Whatever just based on the oddity of the CPU and the mounting system.Quote:
Originally Posted by Paapaa
Lets put things into perspective here. At normal flowrates (1.2gpm) the water does not heat up much per pass due to water high heat capacity. Because of this, even with a 10C delta in die temps, with my current equipment, I would most likely not be able to tell the difference. My equipment has a 0.1C resolution with accuracy worse than .1C. As pH (Derek) found out, the delta T for his 70watt heat load (not sure the GPM but i think it was 1.5gpm) was around .03C or so. so you see its VERY hard to meausre this, even if the difference between two waterblocks is large.Quote:
Originally Posted by Paapaa
i can see the test here using the 14mmx14mm setup as most people here seem to remove at some point thier IHS to get a little(or alot in some cases) more overclock. you know, get a little more of an xtreme result:D
please take this as my personal opinion from what little i have read here in the forums and my still minimal knowledge in watercooling. just seems like at a forum like this where people are looking for that extra through any means the test Maxxxracer does is more relevant to the people here.
if three or four water blocks perform very closely with the cpu's IHS on but begin to show much different results with it removed. then it would be nice to know which block performs the best with the IHS removed. maybe if the funds and Maxxx's willingness are there, then there could be a test rig that would simulate a cpu with the IHS and another with the cpu naked. this way we could see both results and make our own decisions which way we will go in our purchasing choices.
in a nutshell, a waterblock that may not seem superior in one test will show it's worth in the other allowing the word to be used. especially when it comes to what i've seen as it is increasingly difficult to improve on watercooling blocks to any significant degree most would consider.
I agree 100%: if you remove IHS, then Storm is most likely the best block. The same applies to old Athlon XP which doesn't have IHS. But how many users really remove IHS ruining the 3 year warranty and possibly killing the processor? Is that really a common practice? It may also become more difficult if the IHSs will be non-removable in the future like in Pentiums - but that is just speculation.Quote:
Originally Posted by omga14
And nevertheless this should be mentioned in the sticky!! My suggetion:
"If your processor has IHS, all the waterblocks will give you identical performance and you should buy the cheapest. If you remove IHS (and ruin your warranty) you will see some differences between various blocks."
haha, alot of the guys here don't really worry to much about removal of their IHS's. either way what you say is fair enough with me:D most(i won't dare say all) who do remove their IHS off the cpu are pretty well experienced with handling the procedure and know the penalties involved if they scew it up.
either way i can respect your arguement.
no offense but the first part of that statement is pretty ignorant and one that I will remedy in the not too distant future.Quote:
Originally Posted by Paapaa
No offense taken, it really was meant to be provocative. It would be good if someone could shed a little light on this issue - IMO it is not clear at all.Quote:
Originally Posted by MaxxxRacer
I am in the process of doing a waterblock review/roundup with an IHS capped winchester 3000 based system.. check my test rig worklog if u want some more info.
Hi, y'all. I've been reading this thread with a bit of excitement because i'm going to be water cooling my next rig, and don't want to spend more than I have to. I read all of those reviews posted a while back of the apogee, including the french one (no translation needed, since I happen to be a native speaker of french :) ) They compared the storm and the apogee on two pentium 4 procs (with the IHS on!), and found that the Apogee preformed marginally better :eek: .
http://www.cooling-masters.com/image...s/results1.png
http://www.cooling-masters.com/image...s/results2.png
translation: eau="water"; valeurs mesurees="measured values"; ecarts calcules= "calculated differences" [in temperature]; ventilos="fans" don't get freaked out by the comma's, european people use them instead of periods for decimals, so 35.1 is writen 35,1 by a european. P4 EE results:
http://www.cooling-masters.com/image...s/results3.png
They go on to say that the storm preforms much better for procs without an IHS (or with the IHS removed) because it provides intense cooling power concentrated in the center, over the core, where as the apogee should preform better on proc's with IHS, and especially on dual core procs, where the heat load is spread out much more evenly over the surface of the proc assembly, instead of just over the core. that actually makes sense to me, but all the reviews i've seen of the apogee are with single core proc's. Could some one do a test of the apogee against the storm, and also against a block in the apogee's price range (such as the RBX) to see which does better on a Dual core proc with an IHS. I really don't care how the storm preforms on other proc's because i'll be building my rig with one of the soon-to-be-released conroe proc's. :D
I actually hope the apogee does better, because wouldn't it be awesome to have a block for 5/8 the price of a storm that gave the same or better preformance?
EDIT: there is a disclaimer in the review that states that the comparisons were between the 2 swiftech kits, APEX and APEX ultra, not a pure comparison between the apogee and the storm.
Which completely invalidates any comparison between the blocks from that data.
Especially considering the ambient temps were .5-1c higher for the stomr tests.
true...but still, it would be nice to see a pure comparison between the two, with a dual core IHS proc. On a side note, for an SLI watercooling setup, is it better to put the GPU blocks in series, with one getting hotter water than the other, or in parralell, with each getting half the flow rate? Also, will I still be able to get good cooling on my 2 GPU's if I use a storm water block? those max out at 2 gpm, where as the apogee maxes out a 3gpm. I will be using a D5 pump.
series.
No it doesn't: the only difference in the setups was the block. How does it make the data invalid?Quote:
Originally Posted by creidiki
How did you make that conclusion? The Air-column shows higher values for Apogee, not for Storm! And they don't actually tell (if I understood correctly) what Air1 and Air2 means. It really seems to be the temperature of air that exits the radiator, not the temp of air that enters - you can see it from the picture on another page. You can also see also that ambient was held constant between 22-23 C (told on that another page).Quote:
Originally Posted by creidiki
For some reason the water got significantly hotter with Apogee than Storm. Maybe this is because more heat was transferred to water by Apogee - maybe due to faster flow? This would also explain the better CPU temperatures. It is clear that Apogee was the better block in that test - no matter how much you hate to see that result :)
Indeed that does not invalidate the results as they were using deltas. If they didnt use deltaT's it would invalidate the testing, but since they did use deltaT's its fine.Quote:
Originally Posted by Paapaa
The air temp was at the intake of the fans. Remember that fanas blow towards the 4 (or sometimes 3) support legs where the wires are routed. Here is a pic of their setup. Air1 and Air2 are the temps for each fan intake. Its best to have intakes on both fans so that you get more even handed results.Quote:
Originally Posted by Paapaa
Intel suggest 3-4 thermocouples per fan (bread probes not the big honking ones used in this test), but that isnt realistic for 99% of testers who dont have the data collection units that are rather cost prohibitive, especialy for high accuracy ones. (For instance, the high accuracy RTD unit from fluke costs 15k)
http://www.cooling-masters.com/image...ges/sonde3.jpg
Ok, I though the fans were pulling air from radiator. Did they mention that the fans were pushing air into radiator?Quote:
Originally Posted by MaxxxRacer
Nevertheless, the ambients were thus higher for Apogee and yet the CPU temps were slighly lower/equal. Had the ambient temperatures been equal, Apogee would've been even slightly better.
they didnt mentinon it, but that is indeed how its setup.. So I would assume thats how they did it unless they thought it would be funny to take the picture like that and set it up differently.. but the french are funny like that. ;)
Ok well i've been reading this post for a little while now and it has my interest. So interested that i signed up for an account.
i am fooling around with the idea of water for my rig, because my opty is running a little warm and i'm looking to overclock it more
i've looked around and found that swiftech overall is a well respected company and makes a damn fine kit.
I've been looking into the apex ultra alot, and i've also been looking into to some different blocks, i thought the apogee block was good and was gonna buy it.
Now i will be the first one to say i do not know much about water cooling at all...
http://www.overclockersonline.com/in...num=360&pnum=7
that seems to hint that the apogee block CAN be better than the storm.
I've seen a lot of talk about that block not being too good. Why? All the tests i have seen show it to be a solid performer in low and high flow settings.
Could someone help me out here?
A few possible reasons:Quote:
Originally Posted by Dtess17
1. Apogee is cheap, Storm is expensive -> how could Apogee be better?
2. Apogee is manufactured in China, Storm is manufactured in USA -> how could something manufactured in China be better than something manufactured in USA?
3. Apogee uses "old fashioned" technology, Storm uses high-tech-jet-impingement-system -> how could old technology beat a new innovation?
This is all about impressions, not about facts. All the real CPU tests so far (including the one you posted) show that apogee is equally good/slightly better than Storm. Yet these forums have so many Storm fanboys bashing Apogee. Go figure...
IMO the difference is totally, absolutely insignificant between most waterblocks out there. Buy the cheapest/most beautiful and buy a large radiator. Radiator has soooo much more influence on temps than the design of the block.
Thanks for your reply. I think i might just get the apex ultra then, regardless of block seems to be a damn fine system. I was just wondering if it was worth it to buy a different block, but i guess not.Quote:
Originally Posted by Paapaa
and to the possible reasons if people are really thinking along those lines.
1. More expessive does not always = better. Scythe Ninja is a lot less expensive than the Zalman 9500 yet they perform on par with a slight edge to the ninja in some cases.
2. Although i love the USA, and like trying to buy things from this country, let's face facts. Many of the products we use everyday are from a another country. I'd say most of our computer parts are.
3. Old tech new tech. blah blah. Sometimes things are a step backwards.
Intel has changed over to what many consider to be a old tech for the new procs. does that mean they are no good?
now my next question if i may, i understand that the rad size has a lot to do with the performace of a given system.
if i only plan to cool my processor at this point (oppty 170) what temp differences would i see from a single 120 dual 120 or a tri 120?
Also overall how much heat does each new component add to the loop? say i add my gpu's can a 2x 120 rad still handle that?
quite a lot from single to double, not much from double to triple , though you will be able to slow down the fans a lot on the triple.
you should really make a separate thread for thhsi tho :)
A few questions:
1. I'm moving from stock fan to a maze gpu block on a 7800gt soon. I guess I need ramsinks, these as good as any? - http://www.dangerdenstore.com/produc...&cat=57&page=1
2. do I need active air on those sinks? I have two 120mm case fans (input and exhaust), a 120mm active on my ram and a 90mm active on pwmic.
3. and who can tell me anything about this radiator - http://www.dangerdenstore.com/produc...&cat=14&page=1 - is it a "noisy rad" or "silent" type. does it even compare to the ones Max suggested? This is my current rad - I'm moving to the DDC+ pump and Storm CPU water block - currently on Mag3 pond pump and DD TDX cpu block
Thanks
Chad
1/2. they work ok without but youd want the swiftech mcr14 sinks for best indirect ariflow performance. ive got those as well.
3. noisy, cores need strong fans
.Quote:
Originally Posted by creidiki
Part of the reason I'm asking all these questions, is that my son is inheriting the components that I change out and building his own system, so I’m spoiling myself with better components.
Just wondering if anyone has any experience with the dd provided dual 120mm heater core linked in my previous link. Would it be considered a noisy (crank some cfm to it), or a quite – does fine with moderate cfm. Or do the w/c custom dual 120mm rads out perform the old car heater cores?
Once again, here’s the rad .. http://www.dangerdenstore.com/produc...&cat=14&page=1
Thanks
Chad