Archive for the ‘Fans’ Category

Fan Testing Round 12

Posted: May 7, 2013 in Fans

Welcome to round 12 fan testing:

A huge thanks goes out to cpachris and Prymus of OCN for sponsoring the fan samples!

Intro

Gentle Typhoon AP-15 1800RPM

Gentle Typhoon AP-45 2150RPM

Noctua NF-F12 1500RPM

Noiseblocker E-Loop B12-3 1900RPM

Akasa Piranha

Corsair SP120 High Performance 2350RPM

Cougar Dual-X Red LED CF-D12HB-R 1200RPM

Noiseblocker E-Loop 2400RPM B12-4

Cougar Vortex PWM CF-V12HP 800-1500RPM

Cougar Vortex HDB  CF-V12H 1200RPM

Under construction…Videos are now uploaded, but I have not extracted any chart data.

Doing a quick video review I find the following

@400FPM air flow levels

  • The two GT fans have the lowest noise level (Around 38dBA)but do have a slight whine noise quality.
  • The two Noiseblocker e-loops have the second best noise level at around 43dbA and slightly better noise quality  (more air noise and less bearing whine)
  • The Corsair SP120 has the third lowest noise level at around 44-45dbA but does have a little bit more motor noise.
  • The three Cougar fans get fourth in noise level with a varied noise quality.  The PWM flavor does exhibit more noise using voltage control than the HDB model.  The Dual-X and Vortex HDB are pretty close.  The Piranha fan is a similar noise level an quality at around 46 or so dBA, fairly smooth but not silky smooth either.
  • The loudest fan at 400FPM is the Noctua NF-F12 at around 48dBA

@500FPM air flow levels

  • The two GT fans continue to have the lowest noise level at 42-43dbA, but it is a different tone (very little air noise, mostly just a bearing sound).  Interestingly the AP-45 sample seems a bit smoother here than the AP-15
  • The Corsair SP120 is second lowest at this performance level at about 49dBA, noise quality is ok but not ultra smooth either.
  • The Noiseblocker B12-3 1900RPM model may not have super low noise level at 50dBA, but by ear it actually sounds much better.  It is a smoother sound quality and makes it seem like a good number and actually comparatively good as the GTs.  Also interestingly the noise quality of the higher speed Noiseblocker e-loop 2400RPM doesn’t seem nearly as good as the 1900RPM slower speed brother at this performance level.
  • The Cougar Vortex PWM is the only one to make it this far of the Cougars and putting out around 52dBA however noise quality is rather poor with a stronger motor tick.
  • The Noctua and Piranha also have legs to get to this air flow level, but the noise is also higher in the 52-53dBA levels and the noise quality not very good.  The Piranha also seems to have a slight motor tick at this speed.

@600FPM air flow levels

  • The two GT fans continue to have legs to get to this point and at their trademark low noise levels although not what I would call great noise quality.  There is a rather dominating bearing noise without much air noise to mask it.  Regardless, it is hard to argue with the low 46dBA noise level at these speeds.
  • The Corsair SP120 would take second lowest noise level at this point with about 54dBA.  It is considerably higher than the GTs, but still seems to be in it’s game without any unusual noise quality issues, fairly smooth for the speed.
  • The B12-4 higher speed Noiseblocker E-loop also makes it this far, but noise quality dramatically reduces with a higher whine developing. Noise level is a similar 54dBA, but it sounds louder than that by ear.
  • The Piranha also has legs to get here, but it’s the loudest of the bunch at 58dBA and also has some noise quality issues with an undulating low frequency and some motor tick.
  • The Noctua and Cougar fans are not capable of this much air flow.

@700FPM air flow levels

  • Only the GT AP-45 2150 model is capable of getting here and continues it low noise crown at only 49dBA but noise quality is continuing to suffer a little bit with a more pronounced whine.
  • Only the Noiseblocker e-loop 2450 model can make it here, but at the cost of a higher noise level 58dBA and noise quality is also poor, just flat getting loud at this point.
  • All other fans are not capable of getting this far.

 

King of “MOST AIR THROUGH THE RADIATOR”

That would be a tie between the Gentle Typhoon AP-45 (2150) and the Noiseblocker E-loop B12-4 (2400RPM) model.  I measure roughly 72hz(2160RPM) from the GT and about 77hz(2310RPM) coming from the NB.  So despite the RPM advantage the GT is still producing the same air flow through the radiator which is a tribute to it’s very pronounced and curved PQ curve.

King of “BEST NOISE LEVEL/AIR FLOW THROUGH A RADIATOR”

The Gentle Typhoons retain their crown on this criteria and really shine at very high speeds, but that differentiation does drop off as speed decrease.  The Noiseblockers e-loops also have an interesting noise quality and perform really well at speeds below about 1500RPM.  At the 400FPM and 500FPM performance levels by ear I have a hard time telling the difference between the GTs and the Noiseblocker B12-3.  The smoother noise quality of the Noiseblocker almost seems to make up the different in noise level.  The Corsair SP120 also does fairly well on noise level and consistently so throughout it’s speed range, it’s just not as smooth in noise quality as the NBs are at slower speeds, but it does hold it’s up at higher speeds.  None are even close to the GTs at really high speeds though where the GTs have upwards of a 10dbA noise level advantage.

BEST FAN 300FPM (~1000 RPM)

At this point I would call all the fans relatively the same.  The Cougars, Noiseblocker, Noctua, GTs, Piranha, are all performing in that 36-39dBA and subjectively hard to tell much difference.  This and slower speeds is where it just doesn’t make sense to spend a ton of money on fans.  Cheaper value fans perform just as well as the expensive ones and the Gentle Typhoons also don’t really separate themselves much.  I actually prefer the sleeved and HDB bearing type fans a little better in noise quality at these slower speeds.  I would probably call it a tie between the Noiseblocker B12-3, Noctua NF-F12, and Cougar HDB as they seem to have a slightly smoother sound quality at these slower speeds.

BEST FAN 400FPM & 500FPM (<1500 RPM)

This is where it gets kind of interesting to me.  I kind of like the Noiseblocker e-loop B12-3 at these points although the GTs are good too and do have a slight noise level advantage.  I just like the smoother sound quality of the noise blocker better and I also prefer the looks of the NB fan better.  I’m going to call it a tie between the GTs and the Noise blocker E-loop B12-3.

BEST FAN 600FPM & 700FPM (>1500 RPM)

While the noise blocker is really nice at slower speeds, the Gentle Typhoon rules the high speed area.  I wouldn’t call either one silent or perfect in noise quality, they do produce a different tone which seems slightly more whiny with bearing noise than others, but the noise level is remarkably low in comparison.  The Corsair SP120 is actually not too bad at higher speeds as well, but it does have a slightly lower maximum where the GT AP-45 is doing just a bit better.

 

Overall.  I was a bit surprised by the Noiseblocker e-loop B12-3 at the mid speed range, very good noise quality makes them competitive with the GTs.  The cougar fans seem to do well at very slow speeds, but they are a bit limited in higher speeds and didn’t do as well as I hope based on forum comments.  The Noctua NF-F12s do well at very slow speeds, but very poor at the higher speed relative to the forum comments.  I don’t understand why they are so often so highly recommended except I know of a few reviews done out there without proper tools that are likely spreading bad information as usual.  The corsair SP120 fan also surprised me a bit.  It did seem to do a little better in noise quality than the H100i fans, but it’s not as strong a performer in noise level as many forum recommendations seem to convey.

 

As always, I would highly encourage you to “Listen” to the videos and be your own judge.  It is very clear to me that the noise level showing in the meter is not at all the full story.  You really need to listen to the audio in the videos and match up air flow numbers to gain a good perspective.  I would also caution that this “Air Flow” comparison is also just approximate.  Comparing thermally would be more accurate, but I know from radiator testing that to do rad thermals accurate, you are looking at a 1 hour test per data point.  I was able to capture upwards of 10 air flow data points per fan in about 10 minutes of flow bench testing.  It took me about 2-3 hours of video recording followed by another 4-5 hours of processing/uploading.  Trying to capture 10 data points per fan over 10 fans thermally would have literally taken 100 hours or more, and yes I’m too lazy for that.  This is the best I’m willing to do and as always, it is just one test.   I would encourage you to take it as such and utilize other resources to gain additional perspectives and test data.

My one and only suggestion about other testing is to look carefully at how they measure noise and performance.  If noise is not measured while the fan is mounted to either a radiator or a heatsink, then the data is pretty much useless.  Noise MUST be measured when mounted in the used condition.

Thanks for looking!
Martin

 

Welcome to my round 11 fan testing.  This is a fairly small round of fans from the kits I previously tested.  Rather than do my normal written form, I’m trying to do this more video based.

Before I do that, I would first like to thank my parts sponsors, without their support this test wouldn’t have happened:

Logo-FrozenCPU

Swiftech_logo_white_backgrounds

Kit Fans Intro

This video does some physical comparisons of the fans and gives you a good close up look of the fan, sleeving, build quality, etc compared with my previous best performing fan the Gentle Typhoon.

Fan Test Rig Description

This video is just a quick overview of the flow bench and meters used in the fan testing to follow.

Individual Fan Tests

The following videos are of the actual test run on each fan recorded with audio and stepping through 50FPM air flow results.  You can now easily adjust two or more fan videos to like air flow numbers and pause them both, then switch back and fort for a direct apples to apples air flow comparison.

Larkooler Kit Fan

Corsair H100i SP120 Kit Fan

Swiftech H220 Kit Fan

XSPC 750 Kit Fan

Servo Nidec Gentle Typhoon AP-15

Extracted Results

These were pulled from the video, by isolating a looped region where air flow was close to the 50FPM increment.  This provides the resulting detail read on the meters and a calculated RPM.  On the right are some subjective noise quality comments I added as I reviewed and extracted the results.

R11-FanTesting-Detail

Summary Radiator Noise Level vs Radiator Air Flow

This is the “Meat & Potatoes” result.  While I wish I could measure noise quality in a good quantitative way, that’s really not possible.  The next best thing is to compare noise levels when mounted to a radiator at like air flows through that radiator.  It takes into account the fans pressure capabilities and puts it in a more real world condition.  It’s not perfect, but the best thing I’ve been able to come up with to simplify radiator noise performance.  Fans that extend further right are capable of higher air flow maximum results at 12V.  Fans with lines lower on the Y axis are producing more air flow per noise level.

R11-FanTesting-Summaryl

No real surprise, but the kit fans all tested relatively the same (most within 3dbA or less differences which fall within the “barely perceptible” level).  The Helix fan did for some reason have a bit higher than expected harmonics on the radiator bench which didn’t seem to be as noticeable when actually testing in a case, but it is something I heard a little when trying push only.  In push+pull I noticed most of that helix harmonic disappeared.

I would consider the kit results to be relatively similar, they are like most fans and all perform roughly the same.  The Gentle Typhoon however does seem to retain that unique ability on a radiator and tested upwards to 8-9dBA lower in noise level at 12V than other fans producing the same flow.  The H100i fans and their 2700RPM capability did produce the highest maximum air flow, but it comes at the prices of having a fairly gritty noise quality.  Noise quality isn’t captured well in the graph and really only something you can listen for in the videos.

The other aspect I’m now noticing that is missing from this single fan test bench is harmonics between the two same fans.  In the thermal testing using the kits and earlier noise testing, I had significant RPM harmonics issues with the H100i fans, but a single fans test scenario completely misses that.  This is something I seriously want to consider in fan flow bench future upgrades.  I think it is important to capture the “paired fan” harmonics effects as it can be fairly significant.  The helix H220 fans did really well paired together in the kit testing, but you just can’t see that in a single fan test.

Also as noted some of the pressure harmonics issues can also be mitigated for by going push + pull.  The helix fans don’t show real well in this single push test, but I found when testing four fans in push/pull on a radiator the fans worked very well together.  They are not up to Gentle Typhoon silence or build quality standards, but in use I would say they fair better than what the above chart or single fan test result demonstrates.

I also think the Larkooler fan subjectively sounds quieter than the produced dBA.  I’m not sure how to describe it, but the sound type is more lower in frequency and seem to contain less motor noise and gritty noise that is more prevalent in the other fans.  It has a noise quality that reminds me of the noise blocker series which I’ve always liked.  Noise level doesn’t measure anything special, but I think this fan does have pretty good noise quality particularly at slower speeds.  This is another one where my own ear and the meters don’t really agree all that well..:)

This at least gives you one more perspective on the sound.  I would suggest listening to the fans at like air flow levels and make a decision not based on noise level, but what you perceive as being less irritating.  That is likely a combination of frequency, noise quality, and noise level.  Don’t put too much weight on the noise level, it is important, but it’s not the entire picture and each person and each setup will be slightly different.

So there is another round and the Gentle Typhoon retains it’s low noise/rad air flow ratio crown.  Nothing comes close…

This is just a quick graphing demonstration of why fan box specs using static pressure and max air flow are generally poor tools in comparing fan performance.  Very much like pumps, fans have a pressure vs flow rate relationship that is not at all linear.  Some are flat some are curved, some have big bumps and dips and it’s the area between these two points that ultimately matters, not what is at the ends.

For example, let’s compare these two fans using box specs:

Cougar Vortex PWM

http://www.cougar-world.com/fileadmin/downloads/COUGAR%20VORTEX%20PWM.pdf

Max Airflow = 70.5CFM

Static Pressure = 2.2 mm H20

Gentle Typhoon AP-15

http://www.nidecamerica.com/fanpdfs/d1225c.pdf

Max Airflow = 58CFM

Static Pressure = .08inwg = 2.03mm H20

Using these specs alone you would assume the Cougar would produce more air flow on a radiator.  Both max air flow and max pressure are greater, you can’t go wrong with that can you?

Well, luckily both of these fans do provide P-Q curves(Most do NOT), so I spent some time cropping out their P-Q curves, converting units and overlaying them.  First here are their native P-Q curves in different units of coarse making a comparison difficult.

CougarVsGT15part2

So I brought both into Excel and plotted them out the best I could, here is how that looks:

CougarVsGT15part4

In addition noise box specs are “Open Air” with no restriction and no adverse “Real World” affects when mounted to a radiator.  Since most fans out there don’t bother publishing P-Q curves, you essentially don’t know what there real world performance will be and the noise specs don’t mean much on a radiator either.

So, just goes to show you…don’t rely too much on fan specs, they are not very good measures of real world performance.  Just like pumps, it’s the pressure somewhere between these two points that matters in the real world and that isn’t static pressure and it isn’t max air flow, it can only be represented by a full P-Q curve.

And my intent here is not to pick on any one fan. I would like to try some cougar fans out myself considering they are supposedly producing this curve at only 1500RPM, but the curves were available so I wanted to see for myself how the PQ curves compared.  Sure enough it produces a very contradicting conclusion vs. looking at only static pressure and max air flow.

Hope this muddy’s the waters for the box spec discussion out there…:)

Martin

Welcome to my second value ($100-$150) 2 x 120mm water cooling kit review, the XSPC Raystorm 750 RS240 DIY Water Cooling Kit w/ Free Dead-Water. These XSPC kits have been very popular among the forum users due to their extreme value and good Do-It-Yourself parts. This kit is very much DIY-custom regarding installation and does not come pre-filled like the AIO units such as the Corsair H100 or Swiftech H220. It does however come packed full of value and the performance level we expect from a DIY-custom water cooling parts.  Unlike the sealed AIO kits with their tiny little 1-2watt pumps intended for CPU only needs, this kit affords you flexibility and enough pumping power (6W) to expand later to include a GPU and extra radiator.  It also provides your usual DIY flexibility in allowing custom barbs/compression fittings and any flavor tubing you want to suit your custom build needs.  Last but not least, it comes in a mostly individual component package (Except for pump/reservoir combo) that makes future upgrading things like the CPU block less costly.  XSPC wraps all that up with some additional visual bonuses such as blue LED modules for both CPU block and reservoir that does add some nice visual flare to your custom water build.  The kit also comes at the ready with extra hardware to mount external radiators which is something those AIO kit’s just can’t do because you need to route tubes through case walls, etc.  In the end it provides you with a big box of water cooling goodness intended to start you in the journey of water cooling without breaking the bank and retaining as much flexibility in installation as possible.

This review is a “LIVING” review as I’m still in the process of testing and writing up the results.  If you have testing or review request, please post in the comments. 

A very special thanks to Mark from FCPU for sponsoring this XSPC Raystorm 750 kit, your one-stop-shop for all your PC modification supplies.

Logo-FrozenCPU

XSPCRaystorm750-01

A quick photo of the Kit’s Raystorm block in darkness action!

Product Description

The XSPC Raystorm RS240 Extreme Universal CPU Water Cooling Kit comes complete with everything you will need to cool your CPU. This kit is designed to handle your CPU and can be expanded to handle more blocks as well.

The kit uses the newest XSPC CPU block, the Raystorm as the core cooling component. This block has a pure copper base and is a top of the line in performance and looks. XSPC coupled this with their RS240 radiator which is a thin profile radiator giving more flexibility with compatibility.

The reservoir is a combination of the XSPC dual bay reservoir along with a XSPC X20 750 pump. This pump has been revised and is much more reliable than previous revisions. It is now injection molded for seamless joints and this new revision 4 model is a black color.

This kit comes compete with all fittings, clamps, fans, mounting hardware and hose.

Note: Kit does NOT come with fluid or additive.

Features

  • Designed for Multi-core CPUs
  • Complete Kit
  • Top End Performance
  • Full Copper Base Water Block

Specifications

CPU Block
Compatibility: AMD Sockets 939, 754, 940, AM2, AM3
Intel Sockets LGA2011, LGA1366, LGA1156, 603, 604
*Requires mounting holes
Block: XSPC Raystorm – AMD and Intel
Sockets: AMD Sockets AM2, AM2+, AM3, AM3+
Intel Sockets LGA1366, LGA1156, LGA1155
Base Dimensions: 56mmx56mmx3mm
LED Support: 4x 3mm LED Holes
Radiator
Radiator: XSPC RS240
Dimensions: 121x35x277mm (WxDxH)
Ports: G1/4
Screws: 6-32 UNC
Fans: 2 x 120mm (4x with push/pull)
Reservoir
Reservoir: XSPC X20 750 Dual 5.25″ Pump / Reservoir Combo
Fitting Ports: 2 x G1/4″
Dimensions: 149 x 100 x 85mm (WxDxH)
Tubing and Fittings
Tubing: 2 Meters – 7/16″ x 5/8″ Clear
Fittings: 1/2″ Barb (Hose Clamps Included)
Pump
Pump: XSPC X20 750 (Black Revision 4)
Pump Performance: 750 lph
Delivery head: 1.8m
Voltage: 12V (4pin)
Fan
Fan: 2 x XSPC 120mm x 25mm Radiator / Chassis Fan – 1650 RPM
Size: 120mm x 25mm
Airflow: 65.2CFM
Noise: ~29dBA
Static Pressure: 1.8mmAq
Operating Voltage: 5.5 – 13.8V.
Connector: 3-Pin
Wire Length: 45cm
Included: RayStorm CPU Waterblock
X2O 750 Pump/Reservoir
RS240 Dual Radiator
G1/4″ to 1/2″ Barb (Black Chrome) x6
Plastic Hose Clip x6
XSPC 1650rpm 120mm Fan x2
120mm Fan Grill (Black) x2
Intel and AMD RayStorm Brackets
Socket 1366 and 1155/1154 Backplates
Socket AM2 and AM3 mounting kit
80mm to 120mm Radiator brackets
3mm Twin Blue LED with 4Pin Molex
5mm Blue LED with 4Pin Molex
2 Meters of Clear 7/16″ Hose
24pin ATX Bridge Tool
K2 Thermal Paste

One particular to take extra note of is that this kit does not come with fluid. It does however come with a bottle of IandH Dead Water (Biocide), so you will need to buy a gallon of distilled water or a liter of water cooling fluid. In spirit of “Value” I chose the first because a gallon of distilled at Wal-mart is only 88 cents which fits this value theme nicely.  Also noteworthy is the 1.8m head pump and low restriction Raystorm CPU block that should leave you with some extra pumping capacity to handle adding more to the loop later.  There are also several other nice features not expected in a value kit such as the LED modules for pump and block, the 24pin ATX bridge tool to make bleeding easy, and the external radiator mount brackets so you can install the kit on back or on top of cases with limited space.  There is quite a lot in just one box..

Logo-FrozenCPU

In the continued pursuit of the ultimate radiator fan, I bring you round #10.  This round was brought to you buy “cpachris” from overclock.net.  He sponsored 5 never tested before fans including the new Bitfenix Spectra Pro Black, Noctua NF-F12 PWM, Silenx Effizio, Noise Blocker M12-S2, and Phobya Naon-G12.  I put these up to the test against the previous undefeated Gentle Typhoon AP-15.  While many fans are good in case fan applications, few seem to rival the low noise output of the GTs once mounted to a radiator.  Unfortunately the GTs are not perfect either, they are a rather boring grey/black color scheme, they do have some resonance issues at specific RPMs, and they are often out of stock and suspect to overpricing due to their high demand and inadequate production.

With that, let’s bring it!

TEST CONDITION NOTES

For the test rig, I’ve built an exhaust collection chamber that’s basically a long piece of 8″ PVC with insulation, a bunch of flow spreader tubes, and an MCR120 radiator mounted to the face of it.  The anemometer hot wire sensor is then mounted to the back of this to measure air speed through a smaller port.  The chamber’s purpose is to collect the fan exhaust, straighten it out, and provide a consistent means to measure it in a fixed spot without creating additional noise.

I am using an Extech Hotwire anemometer that is set to zero at the start of the round left alone the duration of the test.  For control and check, I test the first fan again after all other testing to be sure the hotwire didn’t go out of calibration.  I have found in the past that it can happen, so this extra test is to ensure data is good relative to the fans tested in this round only.  It’s a good meter, but I still don’t trust it without doing the checks.  It is however a hotwire meter which more importantly doesn’t have a vane probe to warm up or create noise.  The only drawback is dust and temperature changes can cause calibration issues. For noise level I am using a basic noise meter measure A-weighted dbA.  The radiator is a Swiftech MCR120 as shown below:

The video is via Canon T2i and audio via Zoom H1.  Not shown in the photo, but I place the Zoom on a tripod in line with the fan at a 12″ distance and the zoom has a foam wind protection sock on it.  The lens used this round was the kit 18-55 as opposed to the Tokina shown below, but you get the idea good HD video and good HD stereo audio are at the heart of the tools.

I combine the video and audio in Sony Vegas 9, add the text notes, and export to an 8MB/sec wmv before uploading to youtube.  During this combining effort, I set the T2i audio to mute and turn on the Zoom H1 audio. I use a double finger snap to create a wavform mark so I can line the two up.  Then I trim the whole thing to start/end of the test done.

Alright, enough of the testing setup, let’s get to the results

Summary

Here is a summary of what the video meters produced.  Sorry about the lack of RPM data on the Silenx, but something was odd with the Hz readings, so I pulled the data off.  The Gentle Typhoon AP-15 results are still the strongest of the bunch in terms of CFM per dBA, however I would as usual encourage you to evaluate the noise characteristics in addition to noise level for a more complete picture.  I have received messages by some readers indicating that they can’t stand the noise type of the GTs and I would agree that they do produce a different type of noise that may or may not be what you are willing to put up with.  The GT’s also do have a habit of resonating as specific RPMs as can be seen by the bump in the noise chart below. The GTs just have a special fan blade that seems to bury nearly all air type noises into the radiator which is good, but it also tends to have a little bit of bearing whine and the motor noise is present perhaps a bit more than other fans.  I think the noise quality of some of the other fans is superior, but the GT-15 remains dominant in noise level.

The noiseblocker actually has good noise ber RPM, but isn’t pushing as much air per RPM.  The Noctua has a very good CFM/RPM similar to the GT15 leaving some of the others as much as 200RPM behind. Regarding looks, I liked the noiseblocker, Phobya, and Bitfenix better than either the Noctua or GT.  As far as build quality goes, the GT probably gets my vote as it has a larger hub with metal bearing casing and dual ball bearing construction. The Noctua probably provides the most accessories giving you several resistor options to reduce fan speeds if you didn’t have a fan controller. Finally, there is PWM control which is only provided by the Noctua in this bunch.  You could however control the others via software tools like a Sunbeam Rheosmart without much extra cost. So there you have it, another round of 120mm fans and the GT-15 is still king of noise level, however there are other factors to consider that may sway you toward other options.

A special thanks goes out to “cpachris” from OCN for sponsoring these fans and my many supporters for sponsoring the tools, equipment, and website.

Cheers!

Martin