Vortez review of Silver Arrow SB-E Extreme

[C64]

You must be kidding.
How can even be legit with no data to back it up?
No reviews of coolers used for data in chart.
H100 compared to air... But no data saying how it was compared. Where was it mounted? Was it drawing air in or pushing air out?

A genuine fuster cluck review.

because many of the air coolers do seem to reflect end user findings in terms of performance but yea some reviews are lackluster like many reviews out there they tend to skim over the surface

 

[doyll]

Sorry mate, data to show how they got their results would make it legit...
"seems to represent user findings" helps back up findings, but does not make it legit.

.

 

[Kaapstad]

Just used the Vortex figures 4.6ghz @1.42 vcore (I don't have a 3930 but my 3960 is close enough).

I only have prime95 which I ran until the temps settled

I was using a H100 with 4 fans on high

Room temps is a bit nippy (midnight)

I got an average core temp with prime 95 of 72C

None of the above is that scientific but I don't think I could get anywhere near what vortex managed for a H100.

 

[Helios1234]

The results are not wrong as such. I'm sure they're not making it up. The question is hoe the results are actually achieved. It's easy enough to guess where the H100 was installed but the results do look weird. As someone pointed out, there's no way the matterhorn pure is so close to the noctua when passively it's a lot poorer. And my charts will show that the matterhorn is in a much lower league. So I'll still maintain that the weird results are due to the recording methods not being thorough enough.

Testing a cooler with PWM enabled is actually flawed methodology.

 

[BrianB]

I guess most people would use it as PWM so I suppose it's worth testing it with PWM. Still should test it at full 12V though surely as well so people can see what potential the heatsink has?

Oh, nice CPU cooler round up, having a look now.

Yep, 612S will be ordered Monday, thank you for doing that round up. Nice temps for £30 cooler.

 

[Helios1234]

I guess most people would use it as PWM so I suppose it's worth testing it with PWM. Still should test it at full 12V though surely as well so people can see what potential the heatsink has?

Oh, nice CPU cooler roundup, having a look now.

It's the way the results are recorded that is flawed. The result on the chart is based on the maximum temperature recorded during the stress test. The way I did it, even for coolers with PWM fans, was to do one test at full speed and one test at 50% speed without allowing the fan speed to increase. This way, during the stress test, temperatures will gradually rise until it levels off near the end of the test and this is the result I record. By having the two tests at different fan speeds, you can extrapolate approximately what the temperatures would be at other fan speeds.

However, if you enable PWM in the test, you introduce several unknowns. The first is the profile that is set in BIOS. The reader doesn't know what it is and for all we know, it might be set to only run at 100% at really high temperatures which means the results wouldn't show the cooler's full potential. Secondly, we don't know the RPM at idle before the stress test. It may be that the fan starts at 600 RPM so as soon as you begin the test, temperatures will rise dramatically before the fan speed has a chance to adjust. As a result, the max temp is recorded at the start of the test when the fan RPM is low, when in reality the maintained temperature afterwards is lower as the fan speed increases. So while my tests show exactly how the cooler performs at set fan speeds, the results of that review are largely dependent on the PWM behaviour of the fans and that will differ from person to person. The two flaws here are that firstly we don't know how the temperatures are behaving throughout the test in response to the varying fan speed and secondly, we don't know if the results even portray the cooler's full capacity. Basically we have no idea what the relationship is between fan speed and performance.

 

[BrianB]

Good points, definitely a bad idea to test with PWM. Understood!

 

[doyll]

The data from each test may be accurate but the way it is presented is very flawed.

Comparing a cooler (water or air) that is either drawing fresh air from outside case or expelling it's exhaust directly out of case with other coolers (water or air) that are free standing in case with no duct to seperated heated exhaust air and then comparing them is comparing apples and potatoes. Calling it flawed is being kind. I call it deceptive.

 

[olivier renault]

Corsair H100 high : 46C.
Corsair H100 Low : 56C.
NH-D14 SE 2011 : 58C.

So yeah, a couple of degrees for similar amount of noise, while much better performance for more noise. That's kinda what I'd expect anyway, given that the H100 comes with 2,200 (2,500?) rpm fans. Slap on some SP120 / NF-P12 on 12V and 7V, and you'd get roughly the same performance as air coolers, if a little bit better for the H100. Same with the SA-E with its faster spinning fans. I think that review is representative.

Impressive performance from the Matterhorn Pure, hanging with the big boys and costing around 1/2.

 

[rjk]

He actually used to benchmark coolers on a test bench until I pointed to him that I review inside a case so that I can replicate real world conditions. He followed suit soon after.

Ironically, he now gets sponsored by Corsair...

being sponsored by corsair doesn't magically make an hour long video about a piece of computer hardware interesting though.

 

[doyll]

being sponsored by corsair doesnt make an hour long video about a piece of computer hardware interesting though.

Video is for entertainment or quick overview of a product. For in depth analysis I need a written review.

Maybe I'm just old school but I want to be be able to look at specs, graphs, etc. and a/b compare them. Can't do that with a video.

 

[doyll]

Helios shoulda been doing this review.

Yes indeed!

 

[rjk]

Video is for entertainment or quick overview of a product.

I wouldn't go as far as calling a TTL video 'entertaining'

and there is certainly nothing quick about them either.

 

[doyll]

I wouldn't go as far as calling a TTL video 'entertaining'
and there is certainly nothing quick about them either.

Not in the least bit entertaining.
An hour long review is not anything but plain old simple BOARING.

We can read, analyze, compare, cross reference, etc. a very comprehensive written review in half that time... often even less.

 

[Nelly]

I didn't like to mention, but I noticed this going on a long while ago >> http://www.vortez.net/articles_pages/gigabyte_z77x_up5th_review,10.html

It looks like the LinX has simply been downloaded off the xtreme systems website from 2009, the linpacks are vastly out of date, it should be outputting double that.

Latest version >> LinX 0.6.4 - new Linpack (11.0.0.004) Click Here

 

[Helios1234]

being sponsored by corsair doesn't magically make an hour long video about a piece of computer hardware interesting though.

Oh yeah, I agree totally. I was pointing out to those that believed TTL is always honest and "doesn't get paid to review" actually gets sponsored by Corsair which to me is ironic.

Nelly, as for the version of LinX that Mishima uses, I think it's because he's been using it in his reviews for quite a while and hasn't updated it. It's probably to prevent newer hardware reviewed from being stressed differently to previous ones tested. Constantly updating to the latest version is probably not the best idea when trying to test different hardware under the same conditions although it might have been wise to update before moving to his Z77 test bench.

 

[mishima]

Was just looking over their review:
http://www.vortez.net/articles_pages/thermalright_silver_arrow_sb_e_extreme_review,8.html

Couple things jump out. H100 shows 4.6-6.7c cooler stock and 12-12.6c cooler overclocked than K2 and D14.

How can this be when other reviews show D14, K2, and Phanteks all within a couple of degrees of H100.

Test system, coolers compared and methodology:


But when I searched Vortez reviews I found D14 in 500r, but can't find reviews of other coolers SA SB-E Extreme is compared with in this review.

Hi doyll,

The SB-E Extreme review was one of mine. The reason the other coolers aren't found within the same case is we had reviewed these coolers already so rather than do specific reviews they were simply benchmarked and added to a new set of graphs since the platform has changed.

Each cooler is tested as they are, out of the box... so a PWM will be tested as exactly that. The passive benchmark test then reveals how effective the heatsink design is.

I spent an entire week getting that current lineup of coolers tested. There are many more to come!

 

[Helios1234]

The problem is, not everyone uses PWM because a lot of those coolers perform pretty quietly at full speed where their cooling potential is at their best. Simply testing with PWM enabled isn't a great measure of cooling performance because you're PWM settings may differ to someone's else. Its largely dependent on your motherboard and your ambient case temps. By showing performance at different fan speeds, at least readers get a better idea what the cooler can deliver at various settings.

Passive benchmarks are just one way of showing efficiency. Coolers will respond differently to different fan sizes or fan speeds.

You would have a more valid review if you simply disabled PWM.

 

[doyll]

Hi doyll,

The SB-E Extreme review was one of mine. The reason the other coolers aren't found within the same case is we had reviewed these coolers already so rather than do specific reviews they were simply benchmarked and added to a new set of graphs since the platform has changed.

If platform changed how can they be compared on a single graph? My understanding of comparison testing is all are tested on same platform, conditions, temps, humidity, etc.

Each cooler is tested as they are, out of the box... so a PWM will be tested as exactly that. The passive benchmark test then reveals how effective the heatsink design is.

You stated this in review, I quoted it, now you state it again. If platforms changed even passive testing will give different results due to different air movement. I would like to see the reviews on each cooler to see just how close the conditions and platforms were. Any variance in test perimeters change results.

I spent an entire week getting that current lineup of coolers tested. There are many more to come!

Can you please give us links to the reviews of each of these coolers?

 

[Helios1234]

I think what Mishima means is that the reviews of each cooler has already been done but when he moved to the new platform, rather than re-reviewing each cooler with pictures etc, he just redid the tests. But that doesn't prevent the testing from being lacking.