Check this homemade WB out

[Daveburt714]

nateman: I applaud your innovative ideas, and machining skill bro!

I really don't doubt that your block works quite well...
That being said, some independent testing would help your cause.

The whole idea that a pump or other failure wouldn't end in catastrophy would be comforting to most OC'rs...

If the design works as well as you claim it does, I'd suggest you take all your info/drawings/shematics and put them in a letter and send it to yourself via certified mail (then don't open it when you get it back),
almost as good as a patent if things go to court...

Granted, you have some proof of concept just by these threads, but that may help you cover your butt...

Keep up the good work man!
Thinking outside the box is how things get done, and you seem to be doing that quite well....

[nateman_doo]

My test data (the stuff you're quoting from skinneelabs.com) is at 4.2GHz, 1.52v on a different CPU than yours. You're at 2.66/2.8GHz (depending on Turbo) and 1.2v. It's totally and completely incomparable.

It is the same CPU, a 920, just different stepping. I read you tested at 4.2GHz, but I figured you started with a baseline test at stock speeds, and the other graphs were at the 4.2 overclock. Apologies for the confusion.

[Silverhawk]

I think a more interesting test is to see if the block + pump alone would fare better than the heatkiller with a single 1x120mm rad.

If the results are better/equal, then this block already shows promise because it allows you to "water cool" without needing so much room and tubing.

[nateman_doo]

I think the members (including me) here are against his actions of adjusting the test results in his block's favor and trying to fool people into thinking it's better.

Looking at the fact that he gets worse temperatures at stock clock than Skinee does at 4.2GHz USING THE SAME BLOCK says it all, doesn't it?

I did NO SUCH DELIBERATE ruse to FOOL anyone. Lets put that to bed RIGHT there on the spot.

I used a double radiator with 2 crappy fans. His test used TWO TRIPLE rads with SIX FANS on each.

[nateman_doo]

I think you've misunderstood me. I'm basically saying your tests aren't valid. Don't feel bad. I don't trust any manufacturers tests on their own products. I'm trying to get you to commit to sending your block to someone like Vapor or Skinnee for some real world testing.

I've said it already, if you're as confident in your blocks performance as it sounds like you are then you should be jumping at the chance to get these guys to test it.

Thanks... I was saying in a nice way thanks for putting me out there. But you are right. No one (here) will likely ever take my word for it.

[shazza]

It's good when people try new things.

It's also okay for others to question and comment on the test protocol - let's just make sure we do it in a friendly and constructive manner. I've deleted a few of the negative posts - no more bashing of other forums - please keep your comments to the technical and testing elements.

I applaud Nateman's creativity, but my major question is with using a baseline for comparison that shows much higher temps than what most of us see with a stock i7 water cooled setup.

A couple of questions / thoughts for additional info and testing -

1) What was the exact configuration of fans for the tests - location and RPM?

2) Assume this was done with the case side panel off - what are the results in a closed case?

I can see where this could be an attractive solution for someone not wanting to do a full water cooled system - something along the lines of the Corsair water-cooled CPU cooler, but with the advantage of the copper heatsink. Of course, cost will also play a part in commercial viability, as I'd think that much copper is on the expensive side.

I think it would also be good to show what the experimental error of the testing is - 5 remounts seems to be a typical number used by CPU block testers. Again - absolutely nothing wrong with showing the results from a single test, as long as it's clear that it is just preliminary info, and it's not used as basis for marketing a product.

[nateman_doo]

It's good when people try new things.

It's also okay for others to question and comment on the test protocol - let's just make sure we do it in a friendly and constructive manner. I've deleted a few of the negative posts - no more bashing of other forums - please keep your comments to the technical and testing elements.

I applaud Nateman's creativity, but my major question is with using a baseline for comparison that shows much higher temps than what most of us see with a stock i7 water cooled setup.

A couple of questions / thoughts for additional info and testing -

1) What was the exact configuration of fans for the tests - location and RPM?

2) Assume this was done with the case side panel off - what are the results in a closed case?

I can see where this could be an attractive solution for someone not wanting to do a full water cooled system - something along the lines of the Corsair water-cooled CPU cooler, but with the advantage of the copper heatsink. Of course, cost will also play a part in commercial viability, as I'd think that much copper is on the expensive side.

I think it would also be good to show what the experimental error of the testing is - 5 remounts seems to be a typical number used by CPU block testers. Again - absolutely nothing wrong with showing the results from a single test, as long as it's clear that it is just preliminary info, and it's not used as basis for marketing a product.

Thank you. Cool heads have prevailed.

You bring up a good point about seating of the block a few times. always room for error there. I seated the HK block twice because my first tests were SO off I knew something was wrong. Then I re-seated it and all was fine.

I don't recall the push or pull configuration of the fans. I was not concerned with the actual "results" of the test per say. I was only concerned with the comparison of each. If the temps were 90°C vs 99°C, I still would not have cared. I was just looking for the results of which block worked better with the same stuff. So I just replaced the block, and changed nothing just to see the difference in testing.

I only did the tests twice just for myself to see, but I did not pay attention to any other numbers but the CPU temps (since thats all I was interested in). If i was looking at the other numbers I wouldn't have went public until I could explain. I saw them and just figured my blocks fan was the reason behind the better temps. The GPU temps are unexplainable.

The reason for the crappy results for the baseline comparison is that radiator used with very low CFM coolermaster fans. I don't know the exact RPM.

Some folks here seem to think that I deliberately screwed with the test. That is completely false. I am not infallible, but I would never screw with anything deliberately in my favor. I have enough personal intregety enough not to do that.

If my block didn't work I would have sucked it up and moved on. I failed with another block that I was working on and didn't doctor any results to make it appear as if it would work. I am just going back to the drawing board.

I PM'd skinnee yesterday, but I have not heard anything back yet.

I am confident that if you take 2 identical motherboards & chips running side by side with one having my block, and the other having the HK LT block similar results will be seen.

[shazza]

If i was looking at the other numbers I wouldn't have went public until I could explain.

That's understandable, and shows it's helpful to get input from others - you'll sort it out as you go forward.

If the temps were 90°C vs 99°C, I still would not have cared. I was just looking for the results of which block worked better with the same stuff. So I just replaced the block, and changed nothing just to see the difference in testing.

Again - makes sense for what your focus was. But, it only shows the performance difference at one extreme. The heatkiller (and other CPU water blocks) were not designed to give optimal performance with the radiator / fan setup you used. That's what others were trying to say earlier in the thread - because the heat transfer from your radiator/fan combo was so inefficient, just adding the extra waterblock and fan was going to give a big improvement. Could a similar improvement be achieved by just using better fans on the radiator? And perhaps a fan on the mobo / chipset?

The reason most CPU waterblock tests are conducted with at least one triple radiator - and often two, is to make sure the radiator/fan combination is not the limiting factor. Essentially trying to take one of the major variables out of the equation, to get to the point of how well each waterblock does.

Hopefully, one of the things your results do show people is how important it is to still have some airflow in the case, even with lots of high end radiators.

It might also be worth it to do some research on the first waterblocks - I believe there were many different configurations tried by the pioneers in this hobby. Keep up the experimenting - and don't get discouraged by the debate.

[nateman_doo]

Thank you for the advise. Very good points.

The reason most CPU waterblock tests are conducted with at least one triple radiator - and often two, is to make sure the radiator/fan combination is not the limiting factor.

People sometimes just slap together a water cooled system. There is a likelihood that people will have a radiator that is the limiting factor and that is where this block would shine vs the HK block which people may buy the HK block thinking it will cure all. I think if for nothing my test shows you shouldn't use a double rad with low cfm fans on an i7.

I do have 2 triple rads, & 2 pumps, and access to another EVGA X58 Micro board with 920, in which I could bench test 2 identical setups at the same time using the same desktop space so they share ambient temps. Basically a mirror of each system a few feet from each other.

No worries. I do not get discouraged easily

[crazydiamond]

im still w/ ya on this one nate i know the heatkiller (or any block for that matter) isnt perfect at transferring heat from the block to the water, so i have no doubt this concept could help remove heat from the cpu. i just wanted to say though, that i wouldnt run the test with 2 different cpu's because depending on leakage etc. i think they could possibly show different temps w/ the save volts/same cooling

[zeroibis]

I just hope that this will get some more people to try. Who knows what could be done with wc/air+heat pipes in the future. I have always wondered why there were no blocks like this maybe it will get some of the companies to try some new things.

[dejanh]

Nate, I posted in the EVGA thread to explain the reasoning behind needing to use multiple (in this case at the very least more than dual) radiators. Using multiple radiators removes radiator heat dissipation capacity as the limiting variable out of the test and allows you to properly test block vs. block, without other variables affecting the test. When you are at the heat load capacity a block that allows dissipation by acting like a heatsink itself will naturally produce better results. It is guaranteed that it will produce better results in fact. However, what needs to be demonstrated is that at peak efficiency (i.e., lowest water to air delta) your block still performs a few degrees better than a comparable top of the line waterblock. I also do not suggest using an HK LT for this test, especially not lapped.

As I said though in the EVGA forums, very good work. It's great to see somebody bring something like this to fruition

[One_Hertz]

So to sum up:

1) The temps shown for the heatkiller waterblock are overwhelmingly higher than they should be with the stated ambient temp and CPU. Vapor's testing shows the same temps with roughly double the heat output from the CPU.
2) All other temps in the system are much higher during HK waterblock testing, part of which there is no reasonable explanation for as adding a fan on the CPU will not cool the GPU by 18C.

Why is this even taken seriously?

edit: and a dual 120mm rad should have absolutely no problem dissipating heat from a single STOCK i7 920 if the fans are 1000rpm+. We are looking at sub 5C air to water delta here...

[Shocker003]

We should support those that try to bring better solutions to our WC hobby and not bashing him from the start. Well done nateman_doo on your new block. If you intend selling these blocks, i will be placing my order sofort. Positive aspect of your block will be in case of pump failure the cpu woun´t overheat and one can still use the system until your new pump arrives.

[jumper2high]

I did NO SUCH DELIBERATE ruse to FOOL anyone. Lets put that to bed RIGHT there on the spot.

I used a double radiator with 2 crappy fans. His test used TWO TRIPLE rads with SIX FANS on each.

70*C on stock clock is bad for single fan air cooling!!!

[Alexandr0s]

So to sum up:

1) The temps shown for the heatkiller waterblock are overwhelmingly higher than they should be with the stated ambient temp and CPU. Vapor's testing shows the same temps with roughly double the heat output from the CPU.
2) All other temps in the system are much higher during HK waterblock testing, part of which there is no reasonable explanation for as adding a fan on the CPU will not cool the GPU by 18C.

Why is this even taken seriously?

edit: and a dual 120mm rad should have absolutely no problem dissipating heat from a single STOCK i7 920 if the fans are 1000rpm+. We are looking at sub 5C air to water delta here...

Because Nateman already said he did not want to fool anyone with this test. If he indeed falsified his own results, he did so without knowing.

Nateman, one of my suggestions would be to use that fan on both setups, and see what it's effect is.

[nateman_doo]

70*C on stock clock is bad for single fan air cooling!!!

The actual temps mean nothing to me in this test. I already stated I was just looking for a temperature difference, not the actual temps themselves. I could have hooked up multiple radiators, my Bong, my chiller, anything I wanted. i wanted WORST case scenario and I think I executed that perfectly. I think the only way I could have done worst would have been a single 120 radiator.

Because Nateman already said he did not want to fool anyone with this test. If he indeed falsified his own results, he did so without knowing.

Nateman, one of my suggestions would be to use that fan on both setups, and see what it's effect is.

When you say a fan on both setups, you mean a fan blowing on the HK block?

Why is this even taken seriously?

Because it is a fully functioning waterblock... that works without water.
Thanks btw.

[One_Hertz]

Because Nateman already said he did not want to fool anyone with this test. If he indeed falsified his own results, he did so without knowing

I am not saying he did it on purpose. I am just saying something is horribly messed up with the heatkiller block comparison.

[KILLER_K]

I like the idea behind it. Some people are scared of something being different these days. How well it works or doesn't work is always a factor when looking at something. To the people that say {BS}, what if it turns out his test was correct? If it turns out you was right, then he simply starts over on a different design then. I think it is good he even shared it and he is taking a bashing here. But yet still holds his cool from all the flames here.

It just seems most of you could have went about this a easier way. Yes we all need opinions and help but there is a right and a wrong way to give advice and help others. Thanks for the pictures and info. Wish you the best of luck on it. Keep on doing the project and stay focused.

[Mudgey]

I really like the idea of making the blocks bigger and use a larger heatsink with more cooling area, however I have the same HK block, the same pump and the exact same WC layout and route with a Y piece before the pump. I get 78c loaded at 4ghz and 1.25v with a single XSPC RS120 rad and a 1200rpm fan. My ambient is approx 22c (68F). My vcore is approx 0.04v higher (1.24v) and the chip is running at 4ghz also.

I would say that something has been overlooked and there is a problem somewhere in the system that is resulting in these overly high temps with that block.

[shazza]

Hey folks - I'm going to keep deleting posts that are flames. That doesn't mean you can't disagree or offer constructive advice as to better testing procedures. Whether nateman_doo wants to take the advice, or publish his future results, is totally up to him.

And, if you're here in support of nateman_doo, that's fine - but please don't cross the line of trying to get an XS vs EVGA forum flame fest going on - many of us participate in both forums and there's no need to make broad generalizations.

[nateman_doo]

I really like the idea of making the blocks bigger and use a larger heatsink with more cooling area, however I have the same HK block, the same pump and the exact same WC layout and route with a Y piece before the pump. I get 78c loaded at 4ghz and 1.25v with a single XSPC RS120 rad and a 1200rpm fan. My ambient is approx 22c (68F). My vcore is approx 0.04v higher (1.24v) and the chip is running at 4ghz also.

I would say that something has been overlooked and there is a problem somewhere in the system that is resulting in these overly high temps with that block.

It is a 3 year old swiftech radiator. I could agree with you that the performance of the HK was questionable, but I used the same radiator for both. I would say perhaps it wasn't seated on the CPU well, but If you look at the pictures in my original post, you can see that there was good contact with the CPU. I took that picture in case the seating of the CPU came into question.

This is also uncured Artic Silver Alumina btw. I know people hate the stuff, but I had a tube of it and as long as everything was the same I am not concerned with the overall results, just the difference. I changed no bios settings at all. Just walked away from it and closed the door behind me so it wasn't touched, just checking in periodically to make sure the AC was still cycling.

[NaeKuh]

Hey folks - I'm going to keep deleting posts that are flames. That doesn't mean you can't disagree or offer constructive advice as to better testing procedures. Whether nateman_doo wants to take the advice, or publish his future results, is totally up to him.

And, if you're here in support of nateman_doo, that's fine - but please don't cross the line of trying to get an XS vs EVGA forum flame fest going on - many of us participate in both forums and there's no need to make broad generalizations.

For the greater good of the forum...

Naekuh has commited Seppuku...

[ryan92084]

its a great idea for a build where rad size is very limited (as others have stated this block just gives bonus radiator space) but it's merits as a block in a build with appropriate raddage has yet to be seen. In the second scenario i can see it being useful as a backup but not much else.

it reminds me of nanofluid (the milky looking kind) with it being useful (to a lesser degree)in a similar situation and that too received this type of response. Most of the builders here severely over rad their builds so they don't immediately grasp the use of such a block.

[Church]

Fin area is too small to actually serve as viable backup for hot overclocked modern cpu. Compare those fins to top air cooler's or how much fin area is in yer average LC rad .. i'd say it gives arround 0.5x120 rad cooling performance at best. And just like with as small air coolers or as small rads it can cool enough only with monstrous airflow and stock clocks (which in heat amount dissipated sometimes is just 1/3rd of what overvolted/overclocked same cpu might bring) = eg. in servers where there are similar sized radiators (especially in 1U & 2U ones), but where also no-one cares about noise, so fans can be really amped up ..
So, as tests shown and others noted, it helps with unadequate rad situation, and is almost irrevelant in liquid cooled builds of most. With normally functioning LC loop with adequate rad area optimising waterblock insides for better heat transfer and less flow resistance will bring better cooling then spending time on implementing such 'cruthes'. Want more cooling? Use more/bigger rads. It's simplier and what LC really is about unlike air cooling where you are limited how big you can grow air cooler.