HWLabs Black Ice GTX 560 reviewed

Largest radiator I've tested so far...

http://www.abload.de/img/04tnb4.jpg

Here's a pic comparing it to the XSPC RX360, a radiator that seemed quite large when I first got it but now... well, see for yourself:

http://www.abload.de/img/055bwc.jpg

This radiator was kindly provided by Caseking, Germany.

Testing Method

Since my last radiator reviews, I've changed my testing setup and method. I got some valuable input from Skinnee, AndreaBZ and many others on how to tweak and improve the testing.

What changed?
- New fans: I'm now using Scythe Kama Flex 1900 (120mm) and Sharkoon System Fan Power (140mm). I switched to the Kama Flex because they are simply better than the ones I had previously and I switched to the Sharkoons because the Spire fans I used before were no longer available and for this test, I needed a lot of 140mm fans.
The Sharkoon fans are rated at 85 CFM @ 1500rpm but this is in open air, so I guess it's pretty much irrelevant when it comes to radiator testing.
- Better control over heat load: I have the heaters connected to a watt-meter and I manually record the wattage every five minutes.
- C/W curves: Since I now have a better idea of the actual heat loads during each test run, I adjust for the differences. The easiest way to do this is to calculate C/W curves, showing the water-ambient temperature difference per Watt of heat applied.
- More temp-sensors: I am now using three temp sensors on either side of the radiator, measuring the air in as well as air out temperatures.
- Recalibrated: All of my sensors were recalibrated, which was necessary after changing the sensor-arrangement.

Test-loop
- Laing DDC (MCP355) with EK X-Res
- 2x Inline heaters
- 4x Water-temp sensors
- 6x Air-temp sensors
- GMR flowmeter
- T-Balancer bigNG and Sensorhub for monitoring and logging temperatures amd flowrates

Each test run lasts 35 minutes, 25 minutes are warm up and the data from the last 10 minutes is all averaged out.
I don't control flowrates. The pump is always going at 12V and flow is as high as it can be in the loop. Just like the CPU blocks are tested with different flowrates due to differences in restrictiveness, I'm also testing radiators at "their own" flowrates, depending solely on their restrictiveness.
I do compare radiators of the 120-mm-format with those of 140-mm-format, even though such a comparison can never be 100% fair. The rads are being tested with different fans and (deliberately) different fan-speeds. Still, we want to know how a 560 does against our favourite 360, right? Just keep in mind that the comparison can only be seen as a rough guide and different fans would (and do) lead to different results.

Compared radiators

Having changed the test setup, I needed to retest the radiators, of course. Unfortunately, most of the radiators from my previous roundup are no longer here for testing. I retested what radiators I coul, but it was enough to cover the main cornerstones, which are:
- XSPC RX360 - Best 360 for low-speed fans.
- Black Ice 360 - Best 360 for high-speed, high-pressure fans.
- Magicool Slim 360 - Solid budget 360. Basically this one marks the low-end.
- Black Ice 420 - This was also retested and it's interesting to see the differences in performance due to the different fans use since the first test.

Now for the results:


Flow

http://www.abload.de/img/flowcfmi9cp.jpg

No real surprise here. In terms of restrictiveness, the GTX 560 pretty much marks the middle ground among the rads in this comparison.


C/W, single row of fans, push

http://www.abload.de/img/cwenre6p.jpg


C/W, double fan rows, push-pull

http://www.abload.de/img/cwdfenqgxp.jpg

Not surprisingly due to it's sheer size, the GTX 560 is way ahead of the crowd in these comparisons.

Here are all of the curves, single fan row and double fan rows (DF) together in one, slightly chaotic graph:

http://www.abload.de/img/cwvsenpihm.jpg

Here we see that even with one row of fans in push, the GTX 560 can stay ahead of contenders with fans in push-pull. On the other hand, a push-pull RX360 is actually slightly better than a GTX 560 in push at low fan speeds.
It's also interesting to see some of the other relations between different rads and fan-configs in this graph.


Data

http://www.abload.de/img/datac7eo2.jpg

IMPORTANT NOTE: Because the radiators are placed vertically on my test-bench (fans blowing horizontally), I'm not getting accurate Air out readings. I moved the sensors to a better spot after the first few runs but it still seems a bit off. I grayed out the Air out and Air capacity values becuase I suspect they are wrong. If you look at the data from the GTX 280 (original sensor placement) you can see that it just can't be right. Look at the data from the RX360 (new sensor placement) and the values make more sense, but I'm still not sure if they are correct.
Ambient temps are in a range of 4° (+/-2°) and Heat load in a range of 12.6 W (+/-6.3 W). That's not great, but I can't do anything about either.

For now, that's all. Hope you enjoyed the read.

Special thanks to Skinnee for providing samples of how he collects and compiles data.
The original, german article can be found here on DeXgo.com.

Cheers,
Shane

Really really interesting!!

Nice resting.

not surprisingly a very good radiator. it's just huge! :p:

one day i foresee a full tower radiator-in-a-case... perhaps ultra-performance delta of 0-1°C with 1000W?

nice... however that rad is just too big to be practical. another product which was born from this forum on pure greed and no application.

Thanks, HESmelaugh! Nice to see the 280 doing so well against the 360.

Quote:

---

Originally Posted by NaeKuh

nice... however that rad is just too big to be practical. another product which was born from this forum on pure greed and no application.

---

Pretty sure Hondacity would love to have 4 of these :D

Quote:

---

Originally Posted by alacheesu

Thanks, HESmelaugh! Nice to see the 280 doing so well against the 360.

---

After modelling and designing my custom case and several parts to go into it...

In my honest oipinion the GTX 280 is the best performance/practical size radiator on the market.

you can fit 2 of these rad into a tiny case and have 3 x 120 performance or when more advanced 140mm fan becomes availiable maybe even better than 3 x 120 performance.

Mark my words this is gonn abe the goto rad for SFF guys, or HTPC.

can it fit in a tj07?

Awesome work, that had to take a ton of time, very nice!!

Wow, great performing rad.

Quote:

---

Originally Posted by dubtown

can it fit in a tj07?

---

no

its longer then the monsta by more then 140mm.

It won't fit anywhere =p.

Think it could fit on the side or top of my V2110 if I wanted, and then just have the inlet/outlet be over the edge (which would give 2 cm) I could have 3 mounted on the side :D (Wtb money and GT 140mm fans!)

Quote:

---

Originally Posted by MomijiTMO

It won't fit anywhere =p.

---

which is my whole point.

it was conceived by bastards of greed without any inclination on application.

Guys dave has had this radiator for almost 4 months now, and its been sitting on his desk because he has no way to fit it.

This rad is straight custom case or bench station application only. There is no case out there, not even the ascension that could take this rad naively.

Quote:

---

Originally Posted by Mech0z

Think it could fit on the side or top of my V2110 if I wanted, and then just have the inlet/outlet be over the edge (which would give 2 cm) I could have 3 mounted on the side :D (Wtb money and GT 140mm fans!)

---

Fail.. dave has tried on a Goliath series already, and no the rad is longer then the case sideways and standing up.

You need to be more open-minded, NaeKuh. A couple of these rads could be a nice centerpiece for a glass-covered coffee table build ...


:sofa:

you could always mount it out back with a custom rad box. but i agree, ridiculously ginormous.

Quote:

---

Originally Posted by shazza

You need to be more open-minded, NaeKuh. A couple of these rads could be a nice centerpiece for a glass-covered coffee table build ...


:sofa:

---

LOL... if you had 4 they could be the LEGS of your coffee table. oh wait.. make that dinner table.

Naekuh I think the same way about ANY quad. They just don't fit in the majority of cases bar TJ07 and MM cases, none of which I have or will ever buy.

I didn't see anyone saying the same about the MORA. . ..

This stuff isn't meant to be for internal wc's.

[I'll stick with my ants. . . 320s ftw]

Nice review but does seem alittle pointless.

Quote:

---

Originally Posted by Brad8988

Nice review but does seem alittle pointless.

---

The radiator might be pointless, but the review is interesting :). Always nice to see how these things scale.

Quote:

---

Originally Posted by anzial

Pretty sure Hondacity would love to have 4 of these :D

---

hahahahah :D

If my machine wasn't always being moved I'd just set up a custom built "waterlog" box with 2 of these. I can't believe you guys are complaining about it not fitting, no1 is making you buy it. That's the one thing that detracts from cases like MM, people just buy the stuff bolt on the parts and tada its done. I get alot more satisfaction when I have to cut and mod something to put it all in or even on the outside, yeah I can't cut with waterjet or laser precision but it brings a feeling of "yeah I made it". I loved my 1st wc setup which was in an old antec case with a HC outside really was proud of it, since I made it with an ancient drill 2 plastic clamps and alot of blood and sweat, my current TJ09 setup is sleek and nice but I don't feel the same way it was just bolted together and soon to be replaced by something cheaper but something I'll appreciate more IMO since I'm going to put alot of work into it. I can see shoving this monster into one of the larger LL towers like PC7 and the rad is about 2cm to long so I would cut a hole and stick out the end tank outside the back where you can't really see it and have the tubes run out there.

It's too bad the pinnacle 24 is a few mm too short to fit this in the front panel. Is XSPC's height measurement on it accurate?

--Matt

Quote:

---

Originally Posted by shazza

The radiator might be pointless, but the review is interesting :). Always nice to see how these things scale.

---

Lol ya i didnt mean the review was pointless.

Quote:

---

Originally Posted by mattkosem

It's too bad the pinnacle 24 is a few mm too short to fit this in the front panel. Is XSPC's height measurement on it accurate?

--Matt

---

My thoughts exactly.

Quote:

---

Originally Posted by tool_462

My thoughts exactly.

---

Pardon me, I meant HWLabs. If only they could have made the end tanks 2mm smaller on each end... I see 602mm of front panel real estate in the Pinnacle with non-precision measuring instruments.

--Matt

Naekuh is right and closer to 6 months.
It is a measured 23.81"x6"x2"
It will not fit in any case that I've found EXCEPT the MM new tower( Ascention24? or is it Pinnacle24) and that it will fit in under the top but just and you'd need to have the top custom drilled for it.
I mentioned this to Ben at MM and he says he can do a top for it.
The other is one Darthbeavis mentioned, an acrylic case from Danger Den but I hate acrylic..:rofl:
When I hit the lottery I'll pick up the MM case and finally be able to use this.

A guy over at OCN had Danger Den custom laser a case for him to hold this rad, its pretty cool. Heres the worklog with pics if anyone's interested.
http://www.overclock.net/case-mod-wo...completed.html

Thanks for the comments, everyone.

The way I see it, this radiator is definitely not meant for internal use inside a case. I guess external rads aren't a big thing with you guys?
One thing I mentioned in the conclusion in the german article is that I think HWLabs should include a set of stand-offs so you can mount the rad to a wall or a set of feet so it can safely stand upright or something like that. I guess a radiator this size was made "because we can" more than anything else.

BTW: I measured 15,5 x 5,3 x 60,7 cm, what ever that converts to in inches.

I'm already testing some more "oversized" rads to add to this comparison, so stay tuned. :)

Someone pointed out that I had made a pretty stupid mistake: I had the first two C/W graphs switched...
They are in the correct order now.

It would have been nice to test it against a monsta or other 140mm fan radiators, nice review though. Some feet would be nice, but if it was external it would ruin the look inside the case.

Quote:

---

Originally Posted by PeterStoba

It would have been nice to test it against a monsta or other 140mm fan radiators, nice review though. Some feet would be nice, but if it was external it would ruin the look inside the case.

---

The only issue is there aint really other 140mm rads out there yet, sounds like Thermochill should have one out soonish

Btw do the high FPI have any impact on how the flow is inside the rad (Does the tubing inside have to be smaller?) just wonder why its so restrictive compared to RX, PA and TFC as its almost as thick, I know it aint a problem, but would be nice if it was just a freeflow rad :)

I updated the first post with a data sheet. Please read the note below it as well.

Quote:

---

Originally Posted by Mech0z

The only issue is there aint really other 140mm rads out there yet, sounds like Thermochill should have one out soonish

Btw do the high FPI have any impact on how the flow is inside the rad (Does the tubing inside have to be smaller?) just wonder why its so restrictive compared to RX, PA and TFC as its almost as thick, I know it aint a problem, but would be nice if it was just a freeflow rad :)

---

More 140-rads coming up soon. ;)

I don't think the tubes in the radiator need to be smaller for higher FPI but it makes sense that they would be. After all, thicker tubes take up more space that then cannot be used for fins.

I want one of these... but absolutely no room anywhere.... if only it fit in the bottom of the tj07

imagine 2 of these internal :P:

maybe I could make a rad box... although that detracts from the simplicity of just one case for everything

if i remember well, there is a thread started by movieman, regarding how to fit a 560 inside a lian li V2000 (http://www.xtremesystems.org/forums/...d.php?t=212025). the solution mentioned there was pretty good imo.

Quote:

---

Originally Posted by HESmelaugh

I updated the first post with a data sheet. Please read the note below it as well.




More 140-rads coming up soon. ;)

I don't think the tubes in the radiator need to be smaller for higher FPI but it makes sense that they would be. After all, thicker tubes take up more space that then cannot be used for fins.

---

THANKS HES!!!

Quote:

---

Originally Posted by Mech0z

The only issue is there aint really other 140mm rads out there yet, sounds like Thermochill should have one out soonish

Btw do the high FPI have any impact on how the flow is inside the rad (Does the tubing inside have to be smaller?) just wonder why its so restrictive compared to RX, PA and TFC as its almost as thick, I know it aint a problem, but would be nice if it was just a freeflow rad :)

---

they use smaller tubes not because of the FPI but because it reduces laminar flow inside the tubes and I believe the tubes are thinner.

martin illustrates this nicely

http://martin.skinneelabs.com/img/HW...beCompared.png

How about you take a few of them and slap them together side by side, then use some MB standoffs and attach your MB directly to them. You could then mount the pump and res directly to them on one of the fans and there you go, a new case, no cutting needed.

I'd love to see the rad mounted to the back of this guy:

http://lian-li.com/v2/tw/product/upl.../x2000q003.jpg

Seems like it might actually work "fit" properly on the back. Too bad the damn case is $600.

I thought the $400 dollars I spent on my case was bad....:eek:

wow... LiLi is also on crack...

I could see a use for this. Whenever i finally get out of this apartment and into my own house, hopefully it'll have a basement window that i can stuff 2 of these into :)

Corrective Update: I found a tiny rounding error in my spreadsheets and corrected it. I uploaded the new datasheet in the first post. The C/W values have changed but only very marginally. There's no visible difference to the graphs but I just wanted to correct this error for completions' sake.

Quote:

---

Originally Posted by [XC] Synthetickiller

I'd love to see the rad mounted to the back of this guy.

---

damn I never even thought of that, thanks man :up:


The new Corsair case should fit them

http://corsair.com/products/800d/800dgallery2.jpg

relocate the PSU to the top, strip out the bottom area and mount two of these babies back to back :eek:

Quote:

---

Originally Posted by b@llz0r

The new Corsair case should fit them

---

Wow, that's a sweet looking case.

Nice test.

How would the GTX560 compare against the TFC480.

Are we looking at roughly 30-35% better performance based on 1500 rpm fans?

Quote:

---

Originally Posted by alacheesu

Wow, that's a sweet looking case.

---

Indeed.

Quote:

---

Originally Posted by Economist

Nice test.

How would the GTX560 compare against the TFC480.

Are we looking at roughly 30-35% better performance based on 1500 rpm fans?

---

I have no idea. I don't know if my testing is similar enough to Martin's and accurate enough for a C/W comparison. If I'm not mistaken, he did extensive testing on the TFC 480.
Of course, I'd love to get my hands on a TFC 480 for testing, but so far I have not had any luck trying to get samples from TFC... or even buying one of their rads. No one in Switzerland sells them and someone I tried to get one second hand from conned me. :(

But I have a nice little surprise coming up soon. ;)

is there a way or a need to try to "standardize" the tests?
i know you've been doing great work... so has Martinm and Skinnee... just think it would be great to cross-reference with ease, in nub language...

Quote:

---

Originally Posted by HESmelaugh

Indeed.



I have no idea. I don't know if my testing is similar enough to Martin's and accurate enough for a C/W comparison. If I'm not mistaken, he did extensive testing on the TFC 480.
Of course, I'd love to get my hands on a TFC 480 for testing, but so far I have not had any luck trying to get samples from TFC... or even buying one of their rads. No one in Switzerland sells them and someone I tried to get one second hand from conned me. :(

But I have a nice little surprise coming up soon. ;)

---

No problem.
I can up with the performance gain using Andrea spreadsheet and cross referencing your results. Sort of back of the envelope stuff. But nothing beats real world testing. Look forward to the surprise test.

does the corsair case fit one of those internally?

that would be damn sexy....


might replace... tj07...

nope never happening until mobos themselves grow out of the tj07

And another corrective update:

As Martin found out, you get skewed results when calculating the delta using only "water out" temperatures when the flowrate is dynamic. This is exactly what I had been doing. Many thanks to Martin for catching this one. :up:
I corrected all of my calculations to use the average of water in and water out. The actual differences in the results are practically nonexistant though. While the C/W-values obviously all increase a little bit, in relation to each other, there are no noticable differences. The differences in flowrates just aren't that big.
Here's a direct comparison of "old" vs. "new" method:

Old:
http://www.abload.de/img/cwendkqt.jpg

New:
http://www.abload.de/img/cwenre6p.jpg

As you can see, the difference lies mainly in the scaling of the Y-axis.

Still, the error needed to be corrected, and I'm going to replace all the graphs and of course the data sheet, with new ones.
(EDIT: New graphs are now in place)