Page 1 of 2 12 LastLast
Results 1 to 25 of 50

Thread: Fan Filter Airflow Compared

  1. #1
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338

    Fan Filter Airflow Compared

    Alright, have been meaning to test this since Round 1 of my fan testing, but kept forgetting/didn't want to do it when I remembered.

    Well, been getting started on Round 3 of my fan testing and was reminded of this.

    So I bought all the fan filters available from Petras and put them to the testbed. I have no explanation of any of the data yet, I usually put a day (or a couple months in the case of Round 2 fans) between testing and publishing to ponder/rest/graph/etc. This data is less than an hour old at the time of writing. And I expected almost none of it. This is my first time with a few of the fans too...so surprises all around!

    All filters were tested in pull (since who puts fan filters on their exhaust?).

    Fan selection was based on a desire to have:
    1) A wide range of RPM overall
    2) A variety of fans at similar RPM

    I have no intention of doing a dust test...just no good and controlled way to do it Also, no ghetto stuff yet, maybe that'll come with time or when I figure out a dust test I like.

    The fans:
    AeroCool Turbine (950RPM)
    Noctua P12 (1300RPM)
    Yate Loon D12SL-12 (1350RPM)
    Slip Stream M (1200RPM)
    S-Flex G (1900RPM)
    Panaflo U1C (2750RPM)

    The Fan Filters:
    "Wire Grille" - Click
    "Metal Mesh" - Click
    "Plastic Mesh" - Click
    "Blue Thing" - Click
    "MCR120" - Click

    Pretty graphs:
    All the airflow data:
    see below

    Airflow data of the ~50CFM fans (so it's easier to see)
    see below

    Airflow loss as a percentage, compared to open-air
    see below


    Words from me:
    I have little explanation for why the wire grille is restricting airflow so much. In fact, I just included it as a control of sorts. I'm also very surprised that some of the filters are as restrictive or more restrictive than the radiator (depending on fan). The variance from fan to fan is also a bit of a puzzle.

    I hate to hit it and quit it, but I have to run now. More from me later (tomorrow likely).

    EDIT: redid all testing on a different testbed (the one I plan to use for Round 3 of my fan roundup) and no longer had any sort of airleak issues and now have much cleaner (and better, IMO) data. See below for new graphs.
    Last edited by Vapor; 05-25-2008 at 02:11 PM. Reason: edit

  2. #2
    Registered User
    Join Date
    May 2006
    Location
    Quebec,Canada
    Posts
    27
    Nice job Vapor. Good information, thanks a lot man .

    CPU: Intel Conroe E6600@3.6GHz<>1.50vcore <> Scythe Infinity
    RAM: 2GB GSkill HZ 6400
    MoBo: Asus P5W DH Deluxe
    GFX: 3850_512<> 2007WFP
    PSU: PC&P 750W
    SB X-Fi Platinium
    Lian Li PC7-B

  3. #3
    Royal Administrator
    Join Date
    Jul 2005
    Location
    New York City
    Posts
    3,434
    Many thanks for this, Eric!

  4. #4
    Xtreme Enthusiast
    Join Date
    Apr 2008
    Location
    Canada, by the mountains
    Posts
    931
    Thanks for your efforts, Vapor. That's great info. What an eye-opener on the restirction by mere wire grills. You've removed any doubt I had about chopping out the meshed fan holes in my case.

    edit: really gotta wonder about the flow from this TT1200W psu. Big fan, but the air has to go through slots cut into the psu and then the fan hole mesh.
    Last edited by TedShred; 05-24-2008 at 09:32 AM.
    Time is an illusion. Lunchtime, doubly so. Douglas Adams
    QX9650 @ 3825mhz CoolIT FZone
    2x HD3870X2 @ 845/946mhz air
    Asus Maximus Formula 0907
    2x2048mb OCZ Reaper 8500
    2x Raptor 150G Raid0; 'Cuda 1000G
    1200W Thermaltake
    xFi Extreme Gamer/ Z-5500
    Vista Ult. 64
    Cosmos 1000
    30" NEC

  5. #5
    Xtreme Enthusiast
    Join Date
    May 2007
    Location
    Richmond (GVA), BC, Canada
    Posts
    541
    Excellent review with professionally done comparason charts.
    The metal mesh will be my best pick for the best dust filteration performance to least amount of air flow reduction ratio.
    Xtreme Air-Cooled OC System:
    Mobo: Asus Rampage Formula (X38 Edition) Rev 1.03G (BIOS: 0803)
    CPU: C2Q Q6600 "G0" L723A765, VID 1.2625, 3720mhz, FSB 465*8, Vcore: 1.464, Idle/load temp: 31c/64c
    CPU Heatsink: Thermalright Ultra120 Xtreme lapped (2* Scythe S-Flex SFF21F S-FDB 120mm Push-Pull configuration)
    RAM: Kingston HyperX T1 2x2GB PC2-8500 (DDR2-1116)(5-5-5-15-3-52-6-3-8-3-5-4-6-4-6-14-5-1-5-5)(2.264 Vdimm)(Rated @ 2.3v default clock)
    HDD: Western Digital Caviar Black 640GB WD6401AALS-00L3B (AHCI)
    Video Card: XFX 8800GTS 320mb Xtreme, Core 612 Shader 1420 Mem 900 (Stock cooling)
    Sound Card: Auzentech Meridian 7.1 (8788 chipset)
    DVD: Pioneer 212D SATA DVD-RW
    PSU: Corsair HX620W Modular
    Xtreme Air Cooled Case: Antec 900 case | 3*120mm intake | 1*120mm & 1*200mm exhuast
    OS: Vista Ultimate x64 SP1 Build 6001

  6. #6
    Xtreme Enthusiast
    Join Date
    May 2007
    Location
    Singapore
    Posts
    735
    Totally confusing since my silverstone tj-09 casing metal mesh doesn't kill airflow that much & fan grills, now thats interesting.
    Everyday is a NiceDay

  7. #7
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    I'm redoing this on a different testbed...found out I was having problems with airleaks on at least the "Wire Grille" and the "Blue Thing". The results were repeatable (I ran each test twice since the numbers were looking funny and some even non-sensical), but looking back, that only really meant that the airleaks were repeatable and had repeatable impacts on measurement.

    Testbed will also be the one for Roundup 3 (the box CyberDruid built), rather than the one from Roundups 1/2 (which I used two nights ago). Open air and radiator numbers (and metal mesh and plastic mesh even) shouldn't change outside of the margin of error.

  8. #8
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    Oh, and my theory as to why the leaks effected the low pressure fans the least is because they simply didn't have the power to pull any significant amount of air through the leak. Still collecting more data on the new testbed, don't have much of a picture yet, so maybe the data is repeating itself, haven't done open-air numbers yet

  9. #9
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    Yay for triple posting....

    (don't worry guys, I'll go lower my post count )

    Anyway, for the "Metal Mesh" filter...do people mount that on the outside or inside of the case?

    That is:
    1) FILTER | CASE PANEL | FAN
    or
    2) CASE PANEL | FILTER | FAN

    I know how the rest get mounted, but this one seems to be able to go both ways, not sure which people do more.

  10. #10
    Xtreme Enthusiast
    Join Date
    Aug 2007
    Location
    South Wales
    Posts
    651
    i've been saying for a while how much grills can restrict airflow. you need only place your hand 6 inches infront of your face and blow, introducing the grill in between every now and then. its so noticeable

    glad you got this up vapour. loads will benefit from it

    "Never under estimate the predictability of stupidity"


    =======
    Q6600 @ 3.6 / TRUE (both lapped)
    P5Q Deluxe
    4Gb Corsair DominatRIX 1066Mhz
    1Gb HD 4870 + HR03-GT
    2 x 320gb WD3200AAKS RAID-0
    Thermaltake Toughpower 850
    X-Fi Fatality / SR-71 Amp > Sennheiser HD650's & Logitech Z-5500
    HP W2207 TFT

    and me??? Powered by absolut, kebabs, tramadol & a new liver

  11. #11
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    Okay, new data...doesn't say much different in the grand scheme of things, but I feel a lot better about it now.






  12. #12
    Xtreme Addict
    Join Date
    Dec 2005
    Posts
    1,035
    I always wondered about how filters impact airflow...Thank you !

  13. #13
    Xtreme Member
    Join Date
    Jul 2007
    Location
    Florida
    Posts
    213
    Yes - a big thank you! Thanks for taking the time to be thorough and explaining your test methods! It is greatly appreciated.

  14. #14
    Xtreme Addict
    Join Date
    Jul 2007
    Location
    Canada
    Posts
    1,397
    Very interesting. Surprising that some of those metal meshs were less restrictive than the radiator, but go figure.

    Also surprising about the respective fan performances of the 50cfm group. I assume those are Petra YL's? Based on other reviews I've seen (as well as personal experience) I would have expected the P-12's to edge ahead of the YL's and the Slipstream. And if I'm not mistaken, in one of your fan reviews, you mentioned that you wouldn't expect particularly good performance of the Slipstreams under restriction. Any ideas?
    i7 2600K | ASUS Maximus IV GENE-Z | GTX Titan | Corsair DDR3-2133

  15. #15
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    Honestly, I was surprised to see any of the filters being more restrictive than the radiator. I've always just thought a radiator is the most restrictive thing you could throw at a fan

    Yeah, that D12SL-12 is a PTS Yate. I expected the P12 to be better than it was as well--and the Slip Stream to be worse (well, a couple weeks ago I expected it to be worse--when I first received my fans, I put the Slip Stream SH up against the S-Flex G and realized the Slip Streams actually weren't so bad on a radiator for just raw airflow). The Slip Stream is easily the loudest of the three, however. Frankly, I have no explanation for the P12's performance. I'm shocked it performs poorly on the radiator--maybe it's just a bad pull fan? Will try push shortly I guess.

    Then again, Petra's Yates are excellent radiator performers. They're almost competitive with 38mm thick Panaflos in terms of airflow loss--no easy feat. I didn't expect either the P12 or the Slip Stream to match it, but I certainly didn't expect the P12 to lag behind the Slip Stream. Oh well.

  16. #16
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    Okay, retesting the three ~50CFM fans on the radiator now...Yate Loon and Slip Stream haven't changed at all, but the P12 is now doing 36.6CFM pull (26.7&#37; loss) and 40.8CFM in push.

    Moves it ahead of the Slip Stream @ 34.65CFM, but still behind the D12SL-12 @ 38.78CFM.

    No clue why the P12 was bad (@ 33CFM) before--it doesn't seem to be a finicky fan and nothing has changed in my environment, not even temperature. No clue what happened

    Graphs updated.

  17. #17
    Xtreme Member
    Join Date
    Feb 2008
    Posts
    285
    Vapor, it would be a lot better to group the graphs by the fans, not the filters, IMHO. We could see the results a lot better.
    Like this:
    Fan1(open)/Fan1(wire grill)/Fan1(some filter)------Fan2(open)/Fan2(wire grill)/Fan2(some filter)

    Also I was really surprised by how much restriction a wire grill has. I was expecting almost the same performance as open gap
    PC: none at the moment
    Lappy: Dell Latitude D520
    CPU: Core Duo T2300E 1.66GHz
    RAM: 1 Gb generic
    VGA: Intel 950 gma
    HDD: 60 GB Hitachi 5400 rpm + 250 GB Phillips external
    Mouse: Razer Lachesis
    Mousepad: Razer Mantis
    Headphones: AKG K701 + LittleDot MK II + Behringer U Control UCA202
    OSes: Fedora 9 - work || XP-games (only DOTA)
    Additional cooling: seated on cigarette packs with custom made USB powered fan

  18. #18
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    Ask and ye shall receive:

    All fans:


    50CFM fans:


    Percentage loss:

  19. #19
    Xtreme Member
    Join Date
    Feb 2008
    Posts
    285
    Quote Originally Posted by Vapor View Post
    Ask and ye shall receive
    Holy , you are fast
    PC: none at the moment
    Lappy: Dell Latitude D520
    CPU: Core Duo T2300E 1.66GHz
    RAM: 1 Gb generic
    VGA: Intel 950 gma
    HDD: 60 GB Hitachi 5400 rpm + 250 GB Phillips external
    Mouse: Razer Lachesis
    Mousepad: Razer Mantis
    Headphones: AKG K701 + LittleDot MK II + Behringer U Control UCA202
    OSes: Fedora 9 - work || XP-games (only DOTA)
    Additional cooling: seated on cigarette packs with custom made USB powered fan

  20. #20
    Xtreme Cruncher
    Join Date
    Sep 2007
    Location
    PA, USA
    Posts
    1,504
    Quote Originally Posted by Vapor View Post
    Yay for triple posting....

    (don't worry guys, I'll go lower my post count )

    Anyway, for the "Metal Mesh" filter...do people mount that on the outside or inside of the case?

    That is:
    1) FILTER | CASE PANEL | FAN
    or
    2) CASE PANEL | FILTER | FAN

    I know how the rest get mounted, but this one seems to be able to go both ways, not sure which people do more.
    i use #1

    thank you for you time and effort, these results are very good to know. i'd love to know how restrictive the mesh on my tj07 is or on a rad grill
    XS WCG Rules: #1: don't pull fart_plume's finger #2: Dave aka Movieman, don't give him your phone number if you like your hearing
    XS WCG Note: There are 2 sets of points, WCG and Boinc. WCG = 7x Boinc

    Project: Dark Matter (<- link) - Asus Maximus II Formula, Intel X3330 3.4ghz @1.32v under load, corsair ddr2 1066 8gigs, evga gtx260 core 216, pc p&c 750W, EK Supreme HF Nickel, iandh 175 res, Swiftech MCP355, Black Ice GTX G2 240, Lian Li v1200b

    silverstone tj07 build log


  21. #21
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    12,338
    I figured most did #1 so that's what I tested with. They're probably both really similar...maybe even zero

  22. #22
    Xtreme Enthusiast
    Join Date
    May 2007
    Location
    Singapore
    Posts
    735
    slip stream doesn't like the plastic mesh but works great with metal and blue thingie... thats interesting. Always knew the Yate loons were awesome fans but this is mind blowing.

    makes the p12 fan seem horrible for a push pull config on a cooler doesn't it ?
    Everyday is a NiceDay

  23. #23
    Xtreme Cruncher
    Join Date
    Feb 2008
    Location
    C:\WINDOWS\system32\
    Posts
    1,451
    Awesome work Vapor, thanks for the tests.

  24. #24
    Registered User
    Join Date
    Feb 2006
    Location
    Portugal
    Posts
    81
    Great Job! I was always curious how my "metal mash" would affect the airflow.

    (And answering to your question: i use it on the outside of the case)
    Amd PII x2 550 @ x4 // HDT-s1283
    MA790XT-UD4P
    Gskill Ripjaws PC3-1600 7-8-7-24 2x2Gb
    XFX 9600gt // AC S2 + NB XL2
    Samsung F3 500Gb
    Corsair VX-550w

  25. #25
    Registered User
    Join Date
    Apr 2007
    Posts
    36
    I thought fan restriction operated on the inverse square law, i.e. twice the distance - 1/4 the restriction.

    Thanks for the test data. I've been using the 3-piece plastic on the basis that they would do better over time in a dusty environment, IOW they wouldn't clog up enough to affect performance as quickly as the other designs and would be more effective (how long until a noticeable buildup inside the machine - another aspect of filters available for testing ;-) . Also be aware that the plastic, 3-piece jobs can come with two different porosities of filter material (see the Mouser catalog).
    . Perhaps when you get some time you could investigate the clogging rate aspect of filters, Vapor.

    Here is a random datum for you: the metal "mesh" on the front of the fan holder in the Silverstone Kublai cases drops the fan RPM about 100. My "mesh" grows an intolerable skin of dust in a week or so. I don't mind that when I replace the "mesh" front with one of the plastic filters as it is supposed to collect dust...

    .bh.

Page 1 of 2 12 LastLast

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •