Results 1 to 16 of 16

Thread: Push vs Pull vs Push+Pull Fan Noise Test Videos (Noctua P12s , GT AP15s)

  1. #1
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693

    Push vs Pull vs Push+Pull Fan Noise Test Videos (Noctua P12s , GT AP15s)

    Wanted to try some push/pull testing.

    Special thanks to Masika for sponsoring the Noctuas, and Hondacity of the GT. Thanks guys..

    I struggled trying to think of the proper scenario here, so I just went with some random ideas and figured I'd see how they turn out. In the end, I'm not completely satisfied... I think I need to build some sort of flow spreader between the radiator and anemometer vane for the CFM values to be more accurate. Oh well, the tests should provide some value and were a good draft run.

    Noctua Push vs Pull vs Push Pull Test 1 (ALL PUSHING INTO FLOW METER)
    http://www.youtube.com/watch?v=VHhKyViTSqE

    First I ran the Noctua with both fans mounted and switched around which fan was turned on, push then pull then push/pull. Obviously push+pull will be stronger, but I'm not believing it's THAT much stronger. Both fans were run in parallel for power, except the second fan had the RPM readout on the second controller. It appears there is either a small amount of Vdroop, or you can simply expect the pull fan to run higher RPMs...because that's what happened. It's probably better to run each set on a different controller so you can intentionally match up the RPM levels. I didn't do that here, I purposely ran them parallel so I could see what would happen.

    Gentle Typhoon Push vs Pull Test 2 (PUSH INTO FLOW METER & PULL)
    http://www.youtube.com/watch?v=wXk0pfzC3Co

    Second was just an arbitrary GT ran in push, then flipped the fanaround and tried pull and covered the microphone with a little wind deflector. I don't think my anemometer is particularly mounted such that it can test equally in each direction. I've also got all sorts of wind noise in the pull scenario. I guess it does at least give you some idea...the pull condition seems to have less harmonic issues than push.

    CONCLUSION
    I'm not quite sure yet...need to review the videos a bit. I do like the push/pull power and sound qualities at the lower RPMs, but not sure about the slight pulsing you get at higher RPM...but I suppose some of that could be tuned out if you had a controller on each channel.

    Regarding push vs pull...not sure. I know in radiator thermal based testing that I ended up with some varied results (pull being better in low speed, push being better in high speed). I suspect it depends on each fan to some degree. Noise wise, I think the deeper you can bury the fan the better but either way works fine. Right now it's looking like push is slightly better for decreasing noise but it's also not performing quite as well.

    Feel free to post your own experiences...I'm on the fence. I may consider some shroud based testing next.

    Cheers!
    Martin


    All of these are ran with the radiator now filled with water..and include the spectrum recording in the corner for a view at frequency spikes...

    Oh and a quick setup picture of the noctua test 1. The GT was similar except I had to pull everything apart and reset so you should not compare the two videos...they are intended to be stand alone tests.

    The second picture is a concept of a new flow chamber I'm thinking about that's really needed to do this sort of testing. What I have now doesn't seem to work very well.
    Attached Images Attached Images
    Last edited by Martinm210; 10-02-2010 at 07:01 PM.

  2. #2
    Admin
    Join Date
    Feb 2005
    Location
    Ann Arbor, MI
    Posts
    13,107
    Wow, loving push+pull at low RPM. 9V push Noctua is slightly worse to my ear than 7V P+P Noctua, but moves half the airflow.

    11V P+P Noctua seems to be pretty even with the 12V push GT in airflow and with 'noise disturbance' to my ear.

    Great stuff as always

    EDIT: rewatching the Noctua video...something wrong with the push airflow readings?

  3. #3
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693
    Quote Originally Posted by Vapor View Post
    Wow, loving push+pull at low RPM. 9V push Noctua is slightly worse to my ear than 7V P+P Noctua, but moves half the airflow.

    11V P+P Noctua seems to be pretty even with the 12V push GT in airflow and with 'noise disturbance' to my ear.

    Great stuff as always

    EDIT: rewatching the Noctua video...something wrong with the push airflow readings?
    Yeah, I do like the push/pull at low RPMs...and yeah I think I'm having trouble with my anemometer vane with fans behind the radiator. Because the air is so disturbed...I think it's not wanting to read right. I need to build some sort of air flow spreader. I think the Xbit guys did just that by gluing hundreds of straws together...I may have to copy that idea.. Unfortunately it will mean more restriction against the fan..

    Bottom line, measuring air flow sucks (very finicky about being disturbed)...I like water much better(takes a beating and straightens itself back out)..lol!
    Last edited by Martinm210; 10-02-2010 at 06:36 PM.

  4. #4
    Xtreme Addict
    Join Date
    Mar 2009
    Posts
    1,233
    Quote Originally Posted by Martinm210 View Post
    Yeah, I do like the push/pull at low RPMs...and yeah I think I'm having trouble with my anemometer vane with fans behind the radiator. Because the air is so disturbed...I think it's not wanting to read right. I need to build some sort of air flow spreader. I think the Xbit guys did just that by gluing hundreds of straws together...I may have to copy that idea.. Unfortunately it will mean more restriction against the fan..

    Bottom line, measuring air flow sucks...I like water much better..lol!
    Cant you just use a big shroud ? Like a 30cm tunel and mesure teh flow at the end of the tunel ? There will be less turbulance than next to the fan.


    Nice review !!!
    BTTB - Gigabyte Z87X-OC - WCed I7 4770k - 2x8gb Ballistix 1600mhz - Zotac GTX 780
    Asus Xonar Pheobus - OS -> Toshiba Q 256gb - Games -> 2x Agility 4 256gb Raid0
    Corsair HX850 - Tecnofront HWD BenchTable - Asus VE278Q 5760x1080

    Serveur - Asus Z77m PRO - 2500K - NH-C12P - 4x4gb G.Skill Ares 1600mhz
    Agility 4 128gb - Corsair CX430M - 1TB Black - 2TB green - 2TB Red

    KatPat - Gigabyte 990FXA-UD3 - FX-8320 - PH-TC14CS - 2x4gb Viper 1600mhz - GTS 450
    Samsung Evo 120gb - Corsair HX750 - Bitfenix Survivor White - Asus VE247H

  5. #5
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693
    Quote Originally Posted by Boulard83 View Post
    Cant you just use a big shroud ? Like a 30cm tunel and mesure teh flow at the end of the tunel ? There will be less turbulance than next to the fan.


    Nice review !!!
    Probably if it's long and big enough, I'm just not sure how long is enough. I recall reading Bill Adams air flow chamber was something like 6' long. Don't think the wife would appreciate that sort of monstrosity in the walk in..

    That's why I'm thinking about the flow spreader idea. Just need to go down to the dollar store and see if I can buy a load of straws..

    I think it is time to rethink this flow chamber though. It's been a good learning experience, but it has some issues to resolve especially when it comes to push vs pull type testing.

  6. #6
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693
    Quote Originally Posted by Vapor View Post

    EDIT: rewatching the Noctua video...something wrong with the push airflow readings?
    I see what you're asking. I think that may just be that fan in pull just sitting there. Too much restriction for the fan and also pushing the air to the side around the anemometer.

  7. #7
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693
    This is what I'm thinking for a flow chamber rebuild:


    Flow spreader would be similar to what the xbit guys did here:




    Here is what a more industrial flow chamber looks like...so this is a similar idea in needing some way to straighten the flow path.


    Just don't expect me to be building one of these anytime soon...


    I'd just like to make something a bit more accurate...

  8. #8
    Xtreme Addict
    Join Date
    Mar 2009
    Posts
    1,233
    Quote Originally Posted by Martinm210 View Post


    Like this !
    BTTB - Gigabyte Z87X-OC - WCed I7 4770k - 2x8gb Ballistix 1600mhz - Zotac GTX 780
    Asus Xonar Pheobus - OS -> Toshiba Q 256gb - Games -> 2x Agility 4 256gb Raid0
    Corsair HX850 - Tecnofront HWD BenchTable - Asus VE278Q 5760x1080

    Serveur - Asus Z77m PRO - 2500K - NH-C12P - 4x4gb G.Skill Ares 1600mhz
    Agility 4 128gb - Corsair CX430M - 1TB Black - 2TB green - 2TB Red

    KatPat - Gigabyte 990FXA-UD3 - FX-8320 - PH-TC14CS - 2x4gb Viper 1600mhz - GTS 450
    Samsung Evo 120gb - Corsair HX750 - Bitfenix Survivor White - Asus VE247H

  9. #9
    I am Xtreme
    Join Date
    Aug 2008
    Posts
    5,584
    wish the vane anemometers have a 120mm diameter..it would make our lives easier.....the channelization contraption (funnel) i made sorta skews the cfm......

    this testing posts are soo informative


  10. #10
    Registered User
    Join Date
    Oct 2009
    Posts
    53
    has anyone tried using a pitot static probe / 5 hole probe to traverse downstream of the fan/shroud/rad assembly and then integrating the result to give airflow data? - could be interesting to show which parts of the radiator are receveing most airflow / how even a given shroud makes the flow.

  11. #11
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693
    Quote Originally Posted by Hondacity View Post
    wish the vane anemometers have a 120mm diameter..it would make our lives easier.....the channelization contraption (funnel) i made sorta skews the cfm......

    this testing posts are soo informative
    Yeah, I wish there was a higher quality larger vane type out there. I've had mine setup in a duct, just keeping it fixed in the center, but it relys on clean airflow that way. Seems to work ok on a single fan test with the rad in between(which acts as a flow spreader), but even with that I find it hard to repeat tests within a few CFM. I just don't think vane probes are all that accurate...very convenient, but not all that accurate.

    Quote Originally Posted by SuperSilent View Post
    has anyone tried using a pitot static probe / 5 hole probe to traverse downstream of the fan/shroud/rad assembly and then integrating the result to give airflow data? - could be interesting to show which parts of the radiator are receveing most airflow / how even a given shroud makes the flow.
    I thought about it, but haven't tried it. I think it would be more accurate, just hard to time consuming to record and convert. It would eliminate any noise from the vane probe and also eliminate warm up error by the vane. I was also thinking about building an array of pitot tubes to average the pressure across a round cross section. Unfortunately, you need a high sensitivity manometer to measure pressures that low and they are not common enough on ebay to find one reasonably priced. I bought an analog digihelic on ebay for cheap, but it sticks and is a bit of a pain to use. I'd want to go digital before exploring that again.

    You could buy one of these, they even have pressure tips you can buy, but it all costs some $.

    http://www.dwyer-inst.com/Products/P...eName=Ordering
    Last edited by Martinm210; 10-03-2010 at 09:11 AM.

  12. #12
    Never go full retard
    Join Date
    Feb 2008
    Posts
    3,980
    I really am deaf as a post... had to turn the volume up quite a bit to hear the differences.

    I do like the spectrum in the upper right of the video, nice touch. Excellent work as always!

  13. #13
    Registered User
    Join Date
    Oct 2009
    Posts
    53
    hmmm, im kind of temped to give the pitot measurement a go, ive got access to automated datalogging and 2 dimensional traverse equipment at work (university aerodynamics department ) just need to get hold of a fan and rad i guess, and find some spare time when no one else is using the wind tunnel . I presume the GT1850 would be what people would want results for.

  14. #14
    Xtreme Member
    Join Date
    Dec 2009
    Location
    Gomel, BY
    Posts
    235
    from our country "standarts" (called GOST) to measure airflow You should use tube of fan diameter and 6 fan diameters long.

    in middle section of tube "straighter" 2 fan diameters long.

    full article here in russian.

  15. #15
    Xtreme X.I.P.
    Join Date
    Jun 2007
    Location
    Bend, Oregon
    Posts
    5,693
    Quote Originally Posted by SuperSilent View Post
    hmmm, im kind of temped to give the pitot measurement a go, ive got access to automated datalogging and 2 dimensional traverse equipment at work (university aerodynamics department ) just need to get hold of a fan and rad i guess, and find some spare time when no one else is using the wind tunnel . I presume the GT1850 would be what people would want results for.
    Anything we haven't explored before would be interesting.

    Quote Originally Posted by Vetalar View Post
    from our country "standarts" (called GOST) to measure airflow You should use tube of fan diameter and 6 fan diameters long.

    in middle section of tube "straighter" 2 fan diameters long.

    full article here in russian.
    I like their flow spreader and anemometer. I can't seem to find one with a vane that big.


    BTW..took the boy to the pipe yard this afternoon and found some remnant 8" PVC that'll suit chamber V2 just fine. Of coarse since he found the piece, he wanted his picture taken..lol!

    This will literally be about 10x bigger than the old chamber so I can build a nice flow spreader inside, and I'm planning to make it capable of testing both 140 and 120mm fans.
    Attached Images Attached Images
    Last edited by Martinm210; 10-04-2010 at 06:32 PM.

  16. #16
    Xtreme Enthusiast
    Join Date
    Mar 2009
    Location
    Toronto ON
    Posts
    907
    Getting him started early....good for you teach him to sleeve and put him to work :-)

    Sent from my X10a using Tapatalk
    MagisD
    ____________________________
    Lurker, Dreamer, Planner, Noob



    6 Quad rads 1 case Maybe I went a little overboard....Overkill Cube

Posting Permissions

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