XSPC RS360 vs RX360, 360 Radiator Shootout Part 1

[Martinm210]

RS360 Details (http://martinsliquidlab.org/2012/03/27/xspc-rs360-radiator/)

RX360 Details (http://martinsliquidlab.org/2012/03/27/xspc-rx360-radiator/)

RS360
http://martinsliquidlab.files.wordpress.com/2012/02/xspc-rs360-001.jpg?w=614&h=284

RX360
http://martinsliquidlab.files.wordpress.com/2012/02/xspc-rx360-001.jpg?w=614&h=319

Restriction:
http://martinsliquidlab.files.wordpress.com/2012/03/xspc-rs360-pd3.png?w=614

Thermals:
http://martinsliquidlab.files.wordpress.com/2012/03/xspc-rx360-009.jpg?w=614&h=409

http://martinsliquidlab.files.wordpress.com/2012/03/xspc-rxvsrs-thermals.png?w=614

Got a couple of them done, only 11 more triples to go..
Cheers!:)
Martin

[Trox]

Would be nice to see EX360 as well

[Martinm210]

It's on the long list of 11 more radiator To Dos...:)

http://martinsliquidlab.files.wordpress.com/2012/02/360radshootout.jpg?w=614&h=739

[BeepBeep2]

With the difference in watts dissipated between them being under 10%, real air/water delta T at same heatload probably under 1c difference and the cost being about 100% more...

Why would anyone buy the RX360?

trollface.jpg

[Roadhog]

only ~210 watts dissipated for 1200rpm push/pull?

[basserdan]

For the single fan, I thought most fans were more efficient (air moved for given rpm) on radiators in pull (I want to say I learned that from you).

I guess the point here is that they're identical setups (push/pull) for each.

[Spillertwo]

Is the Black Ice SR-1 also on the long list ?
Would like to know the difference between the RX360 and this one.

If it is on the list could you check that the Black Ice is indeed about 0.9 cm thinner then the RX360 ?
For me this is also a deciding factor between the two.

[Vetalar]

oh! radiator skyscraper!!!
and great test as always!

[SpuTnicK]

It's on the long list of 11 more radiator To Dos...:)

..pic...
i wonder why the box for Phobya G-Changer v2 is so thick. Would that some how effect the thickness of the rad itself?

Martin, can also you give percent numbers for difference in max watts between rads for a set of speeds, for example, 600/800/1000/1200RPM..., or average figures for 600-1000RPM(low speed)/1000-1600RPM(medium speed)/1600-...RPM(high speed)? This may give a more clearer understanding how much is one rad better/worse to another one in terms of performance.

[Martinm210]

With the difference in watts dissipated between them being under 10%, real air/water delta T at same heatload probably under 1c difference and the cost being about 100% more...

Why would anyone buy the RX360?

trollface.jpg

Looks, because we always pay a premium for minor premium gains, and probably because I don't think people care too much about $40-$50 in rad price difference when they just poured $1200-$3000 into hardware...?


only ~210 watts dissipated for 1200rpm push/pull?

Yeah, the new bench is drawing significantly lower numbers. I wouldn't worry too much about the actual numbers as much as relative performance. The new bench is all insulated lines, enclosed to prevent room air circulation from affecting results, and internal case type results which is very different than an open air type test. Open air results without insulating anything can net quite a bit more heat. I tried one test with the top lid off just for the heck of it and got pretty close to the same result, so it's not that the ports are being overly restrictive. Anyhow, the numbers are what they are and relative to this test bench only. There is a lot of change to try and improve consistency and precision, but I'm not sure the "Watt" numbers are accurate or not. They do seem a bit low, but again most tests are run open air so not much to compare with.


For the single fan, I thought most fans were more efficient (air moved for given rpm) on radiators in pull (I want to say I learned that from you).

I guess the point here is that they're identical setups (push/pull) for each.

In shroud testing it turns out to be pretty much radiator specific. I chose push in the test method because it worked well for the bench setup. Bottom push fans are permanently fixed on a template and I just set the rad on, test, then screw the pull fans on...test.


Is the Black Ice SR-1 also on the long list ?
Would like to know the difference between the RX360 and this one.

If it is on the list could you check that the Black Ice is indeed about 0.9 cm thinner then the RX360 ?
For me this is also a deciding factor between the two.

It is on the list, I will check..:)

[Martinm210]

I also have EK's XT and XTX rads, they came after that first tower photo.

[SkItZo]

Great work as always Martin :D

[Systemlord]

It would be interesting to see how well the 360's bigger brother performs (420), there is no data for 140mm radiators.

[Waterlogged]

It's on the long list of 11 more radiator To Dos...:)

snip

I'd like to see the bottom 2 in that stack get tested next please. :hrhr:


The Alphacool and Phobya you can just toss in the bin right now, that should cut the work load down a little* :ROTF:


*Yes, I've been drinking some haterade again.;)

[Makymaco]

Martin, your are the best :up:


I'd like to see the bottom 2 in that stack get tested next please.
I'd like the rads tested with different characteristics tested first : [thick/high fpi] [thin/high fpi] [thin/moderate/fpi]

[Martinm210]

These fans are actually testing fairly low on the CFM/RPM on my fan test rig which is partly responsible for what appears to be lower than normal watts dissipated numbers. 2200RPM Titan RPM on the graph is about equal to 1800GT-15s or 700RPM Titan is closer to 500RPM GT-15, etc. I guess it doesn't really matter and I could always make a second chart to convert to GT RPMs maybe even a 3rd chart for delta 38mm numbers which is even more CFM/RPM, but that's part of why the numbers are low.

I generally test them in the order received plus or minus, but I can see I'll probably need to do some grouping by the time this is all said and done with as many lines as there will be. This will probably be the last testing I do for several months though, the ole half finished mini-buggy builds in the garage are calling my name again..:)

[RCG_Bex]

No Kaze 3000RPM testing? :(

[Scubar]

I'd like to see the bottom 2 in that stack get tested next please. :hrhr:


The Alphacool and Phobya you can just toss in the bin right now, that should cut the work load down a little* :ROTF:


*Yes, I've been drinking some haterade again.;)

Why do you say that ?

I have the Phobya G changer 240 and 480 and both have been excellent for me. They have been pretty good in reviews also from skinnee

[Waterlogged]

Why do you say that ?

I have the Phobya G changer 240 and 480 and both have been excellent for me. They have been pretty good in reviews also from skinnee

I thought my little disclaimer would have given it away. ;)

Phobya and Alphacool are owned by AquaTuning and I have serious issues with the ownership of said company (see right hand link in sig for reminder).

The Phobya should match up closely with the old Fester X-Changer as I'm fairly certain it's the same thing. Fester dumped them pretty fast once word finally got out that they were made in China like all the rest and set out on their failed adventure to develop and manufacture one in Germany.

[Scubar]

Oh right the Aquatuning thing. It is a good rad for the money. Im moving my rig to a diff case soon anyway so shoving some Coolgate rads in there instead.

[Martinm210]

No Kaze 3000RPM testing? :(

No, in an effort to both make testing less painful and to allow a tendline between all points collected I decided to stick with one set of fans this time. It also helps with consistency because the lower push fans are a permanent fixture in the acylic template so I am always reading rpm from the same exact fan so if there is any fan sample variation..each rad is tested with the same exact thing.

It doesn't test high speeds as well with this but it is hitting across the most used rpm range and most rad tuning lines cross within the range being tested..ie a rad doing well at 2200rpm will continue that trend up through 3000.

[tiborrr]

One thing to note: I would suggest you test fan at different voltage, not RPM speed. The RPMs are quite different on the same voltage when the fans run on ~ 7FPI or 20FPI radiators (giving unfair advantage to high FPI radiators).

[meanmoe]

One thing to note: I would suggest you test fan at different voltage, not RPM speed. The RPMs are quite different on the same voltage when the fans run on ~ 7FPI or 20FPI radiators (giving unfair advantage to high FPI radiators).

I'm not sure that makes sense unless he's testing specific fans. Don't we undervolt for lower fan speed.
If he did it by voltage then the results would be skewed depending upon the fan that he used. Also, since there is no direct correlation between fan voltage and radiator performance, I think that the results would be less meaningful.

He covered a full range though and that's what's important. It could be that your point is more subtle and I'm missing it.

[Martinm210]

One thing to note: I would suggest you test fan at different voltage, not RPM speed. The RPMs are quite different on the same voltage when the fans run on ~ 7FPI or 20FPI radiators (giving unfair advantage to high FPI radiators).
I am well aware of the voltage vs rpm thing, but you have to weigh the priorities for what is being tested (radiator) and what is being held constant. Is it the electrical power of the arbitrary fan selected or noise of the radiator?

I think noise of the radiator and noise correlates more closely to rpm than it does voltage. Nobody really cares how much power is consumed and very few run fans at their full 12v or control any one fixed voltage. In addition voltage control is generally not something most users know when controlling fans so if you speak tests result in terms of voltage you are speaking in a language few understand and can relate to since it is specific to that one fan using voltage which few have the tools to measure.

Also fan rpm is something I can log throughout the test and average fluctuations on where voltage I can not. Any time you can average and log results real time and average later the better the test. In this case I am logging RPM and it does fluctuate depending on heat and other factors, it's just not as constant as you might like, but no big deal when the result is averaged and the watts dissipated is plotted relative to that actual RPM averaged.

Other than looking at what the maximum performance is of one particular fan selected, I fail to see how voltage is even really useful. The fan selected is really fairly arbitrary except for the CFM/RPM differences in which you could correct for if you know what those corrections are.

Anyhow...RPM is where its at IMO...I've thought about this a lot when doing my shroud testing and establishing a new test procedure for my V2 bench. Voltage is really only good for seeing what a pump or fans maximum performance is. It makes perfect sense with pumps since many people run them at fixed 12V, but pretty much everyone is dialing fans down, and it's trying to test at like noise levels and RPM seems to make the most sense. I can still see some value when testing fans specifically as you may want to know a fan's maximum performance, but it a whole different focus when testing radiators.

If we really wanted to get anal, you could run another series of tests to evaluate and correlate RPM to dbA and then compare performance of watts dissipated relative to dbA, but that would be even more work and yet another test that also would convey the information in a set of units people wouldn't understand other than which is better (It wouldn't tell them scale or what they can expect on their system). When it comes to radiators and fan operation, RPM is the universal language most understand and I think best fitting of noise as the priority.

I am using a fixed pumping power on this round though, so rads with higher restriction will naturally receive slightly lower flow rates as they would in a real system, but regarding the fan, I'm sticking with RPM as the common X axis. It's not quite as good as dbA, but it's the next best thing I've figured out that is practical and useful.

[meanmoe]

If we really wanted to get anal, you could run another series of tests to evaluate and correlate RPM to dbA and then compare performance of watts dissipated relative to dbA, but that would be even more work

And it would be wasted IMO because it would be too fan specific. As you know better than most due to your extensive (and impressive) testing on the matter, every fan sound different at different rotational speed. So I agree with you that rotational speed is the most relevant. Plus, there are general rules of thumb with respect to fans. Under 1000rpm is silent-quiet and over 2000 is for those that do not care about noise. For me, I'll focus in on your curves at the 700rpm to 1200rpm range, because I know I can find an equivalent fan in that range that fits my ear.

To use voltage is to dumb down the meaning of the results to a simple relative comparison. It would be specific to the Ultra Kaze and would not offer as much benefit for buyers. This could benefit the EK rads slightly (based on another test) maybe, but it should not cause high FPI rads to shine over low FPI rads - not to the informed reader anyway, e.g. those that don't *require* higher is better be stated on a plot.

An informed reader would understand that the higher is better applies only within the desired fan speed range. Maybe this it what tiborr was getting at?

Just my opinion (and your test), but I like the call

[basserdan]

Why would anyone buy the RX360?trollface.jpg

Because you can sit next to it and just tell it a story about a faint breeze and it will keep your voltage soaked quad core under 60c during stress runs.

[UrbanSmooth]

My three RX360s handle my system quite nicely. Rarely do I see temps over 50C in the CPU and GPUs. I've been running this system for over a year now.

[meanmoe]

@martin
So after thinking about this a little further, the money chart should be tied to air flow, right, and then accompanied with radiator fan testing plots. So in terms of proper analysis, I don't think that the entire design space can be put into one plot. What do you think? And of course, this is just my opinion. I'm not the one bustin' ass to do this work. So thanks for the free test and testing data - however you choose to do it is much appreciated.

[Martinm210]

@martin
So after thinking about this a little further, the money chart should be tied to air flow, right, and then accompanied with radiator fan testing plots. So in terms of proper analysis, I don't think that the entire design space can be put into one plot. What do you think? And of course, this is just my opinion. I'm not the one bustin' ass to do this work. So thanks for the free test and testing data - however you choose to do it is much appreciated.

I was thinking a very basic watts/dollar across the same RPM range. I'll pick a few vendors and shop for the best price, then add in the fan cost, and then use something like 10C delta to plot watts/dollar. Haven't tried it yet, but at least that's sort of what I was envisioning. Obviously the higher speed RPM the better and the thinner value rads will shine, but that's what I was thinking.

I'll work up an example, it will probably make more sense then.

[Systemlord]

I was thinking a very basic watts/dollar across the same RPM range. I'll pick a few vendors and shop for the best price, then add in the fan cost, and then use something like 10C delta to plot watts/dollar. Haven't tried it yet, but at least that's sort of what I was envisioning. Obviously the higher speed RPM the better and the thinner value rads will shine, but that's what I was thinking.

I'll work up an example, it will probably make more sense then.

You probably don't hear this enough so I'll go ahead and say it, the world appreciates all of the work you have done over the years providing invaluable information for us water coolers. A BIG Thank You for all you have contributed to the water cooling enthusiasts crowd! :up: