1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Cooling Fan Testing and Comparisons

Discussion in 'Hardware' started by LennyRhys, 22 Apr 2020.

  1. LennyRhys

    LennyRhys Oink!

    Joined:
    16 May 2011
    Posts:
    5,961
    Likes Received:
    453
    Yeah, the one with the "wind tunnel" and anemometer with a hot wire. It's pretty cool and I guess it's suitable for ballpark figures, which is what most PC users want. If you wanted to get an accurate airflow measurement from an air velocity reading, you'd need a much longer tunnel with proper flow rectifiers (the mesh won't do much of anything at all), and the air velocity readings should be taken at various points within the cross-section and averaged out.
     
    Bloody_Pete likes this.
  2. cobalt6700

    cobalt6700 Active Member

    Joined:
    17 Oct 2005
    Posts:
    610
    Likes Received:
    100
    Spot on there @LennyRhys. Hot wires are great for measuring small apertures / slots. When used in a duct, as you say multiple readings need to be taken and averaged out (on a straight duct, air usually travels at a higher velocity in the centre, due to the friction from the walls of the duct).

    @MLyons If you are looking at building something like this, I would suggest whilst Gamers Nexus rig is a starting point, it could be much better.

    The mesh fitted to their 'wind tunnel' wont really be helping much at all - in fact it will be creating eddy currents which won't help matters.

    For accurate measurement of a flowrate, you should be looking at 10 x duct diameters upstream. Flow straightening is a thing that helps, but it does introduce a small resistance. Something like plain drinking straws cut down to 30mm lengths would probably do it. This will help a lot with the 'swirling' motion introduced into the airstream by the fan rotation.

    For the ~ 150mm tube used in their design you would probably only need to take 5 measurements (a centre and 4 quadrants) to get a good flow reading with a hotwire.

    You could also look to measure with a vane anemometer for accuracy - again you will encounter the issue of resistance. Most, if not all vane anemometers are 100mm Dia.
     
    LennyRhys and Big Elf like this.
  3. MLyons

    MLyons Half dev, Half doge. Staff Administrator Super Moderator Moderator

    Joined:
    3 Mar 2017
    Posts:
    3,376
    Likes Received:
    1,477
    You absolute beauty. The rest of the information you provided is great but this, this is the amazing bit. I've been looking for the name of this for bloody ages. "circle anemometer" is what i previously used and the results were crap. Anyone know of a vane anemometers that outputs live to USB and can be read by a PC running python? I want to use a program that controls the PWM signal and can automate the whole testing process. The goal is to install the fan and after pressing a button I get CFM and noise readings graph from 1 - 100% PWM
     
  4. cobalt6700

    cobalt6700 Active Member

    Joined:
    17 Oct 2005
    Posts:
    610
    Likes Received:
    100
    No problems at all :)

    No idea if you can integrate with python, but DPM make some really nice manometers with live USB logging. We have a couple of these around for pressure/flow measurement at work:

    ST 650M + DPM ANE £685 https://www.ttseries.com/product/dpm-st-series-mini-test-set/

    You do also need the software which is £95. https://www.ttseries.com/product/st6-micromanometer-download-software/

    All in - it's not cheap.

    I'm unsure of another 100mm vane anemometer with live USB data output. I'm aware of the Pacer DA400 series - but these are usually 70mm vanes, and they come in at £1100+
     
  5. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    11,166
    Likes Received:
    1,673
    Hey man sorry for the delay. I've been hella busy.

    PM me your addy. I boxed the fans up earlier, mum said she will post them sometime next week. I won't be here, so I need to get it labelled up before Monday.
     
  6. LennyRhys

    LennyRhys Oink!

    Joined:
    16 May 2011
    Posts:
    5,961
    Likes Received:
    453
    Awesome - thanks Vault, much appreciated. PM Inbound!
     
  7. LennyRhys

    LennyRhys Oink!

    Joined:
    16 May 2011
    Posts:
    5,961
    Likes Received:
    453
    Hey folks, sorry it's been so long since I've posted in this thread. I had a really long wait for the fan that I got in the post this morning (4 months to be exact), and there was a bit of a fiasco with the seller in terms of cancelling the order and getting a refund, only to be told it's actually in the UK now even though we deemed it lost in transit.

    Aaaaanyway, I reckon it was worth the wait, and in the end it cost me just under £20. It's a [very] rare fan, and a bit special: it's a counter-rotating fan with an 11-blade impeller at the front and a 7-blade at the rear, with 9 vanes sandwiched in the middle. The max power draw on the ticket is 8.2A, and each "fan" has its own set of four wires for power/tach/PWM, so there are eight wires in total coming from the frame. I've tested the fans each by themselves and also together, and I can honestly say it's the only fan I'm legit scared to ramp up to 100%. Even by themselves the fans are seriously strong.

    On paper this thing produces 293CFM, so it should (quite literally) blow all my other fans away. Sanyo Denki make a fan much like this which has similar performance, but it costs £65 which is why I opted for this one instead.

    I'll try to get a video together soon. The fan is cleary very well used but there's no nasty noise coming from the bearings and it works flawlessly.

    [​IMG]

    [​IMG]
     
    cobalt6700, enbydee and Big Elf like this.
  8. Big Elf

    Big Elf Oh no! Not another f----ing elf!

    Joined:
    23 Apr 2009
    Posts:
    3,557
    Likes Received:
    270
    Wow, just wow!
     
    LennyRhys likes this.
  9. Mr_Mistoffelees

    Mr_Mistoffelees The Cat Lies Down on Broadway

    Joined:
    26 Aug 2014
    Posts:
    2,674
    Likes Received:
    836
    Earplugs at full speed methinks.
     
    LennyRhys likes this.
  10. MLyons

    MLyons Half dev, Half doge. Staff Administrator Super Moderator Moderator

    Joined:
    3 Mar 2017
    Posts:
    3,376
    Likes Received:
    1,477
    [​IMG]
     
    LennyRhys likes this.
  11. Goatee

    Goatee Well-Known Member

    Joined:
    19 Apr 2015
    Posts:
    1,590
    Likes Received:
    556
    I would love to stick that on a 120 mm rad and see what I could cool in terms of components. Is it really pulling 100W?
     
    LennyRhys likes this.
  12. enbydee

    enbydee Member

    Joined:
    10 Jul 2014
    Posts:
    96
    Likes Received:
    23
    Would you need to watercool the fan?
     
  13. Goatee

    Goatee Well-Known Member

    Joined:
    19 Apr 2015
    Posts:
    1,590
    Likes Received:
    556
    That's a good point, might need another fan on there to cool the fan.....
     
    LennyRhys and enbydee like this.
  14. Mr_Mistoffelees

    Mr_Mistoffelees The Cat Lies Down on Broadway

    Joined:
    26 Aug 2014
    Posts:
    2,674
    Likes Received:
    836
  15. liratheal

    liratheal Sharing is Caring

    Joined:
    20 Nov 2005
    Posts:
    11,667
    Likes Received:
    1,246
    Sweet jesus.

    That's one aggressive fan.
     
  16. LennyRhys

    LennyRhys Oink!

    Joined:
    16 May 2011
    Posts:
    5,961
    Likes Received:
    453
    That's max inrush current at power on, assuming it's set to 100% duty cycle. I haven't measured it yet (I plan to) but I'd guess around 5A / 60W running at full tilt, possibly more. The intake fan is a 5,500rpm beast so I'd be surprised if it pulls less than 3A.

    Edit: not far off! Intake pulls 3.1A at full tilt (haven't measured rpm yet) and exhaust pulls 2.4A, so that's 66W. Whoa!
     
    Last edited: 26 Jul 2020
  17. Goatee

    Goatee Well-Known Member

    Joined:
    19 Apr 2015
    Posts:
    1,590
    Likes Received:
    556
    Have you thought about adding an AIO test to your fan test suit? Take something like a bog standard AIO, get a copper plate and drill a couple of holes and stick some 3d printer extruder heaters in it (2 x 40W heaters would give an approximate representation of a CPU).

    Stick a temp sensor in/on the block and look at the delta of room vs heater plate with each fan? Might give a good indication of the cooling effectiveness of the fans in a practical setting, see if there is some sort of diminishing returns.
     
  18. Bloody_Pete

    Bloody_Pete Technophile

    Joined:
    11 Aug 2008
    Posts:
    7,303
    Likes Received:
    525
    Would probably be better with a high end tower cooler, as heatpipes have a far far higher thermal conductivity!
     
  19. LennyRhys

    LennyRhys Oink!

    Joined:
    16 May 2011
    Posts:
    5,961
    Likes Received:
    453
    I considered doing some AIO testing on my "test rig" (read: work PC, which is currently at home...) but I reckon it'd end up being more about the cooler than the fans. I now have a 360 rad which I'm planning to use for sound level testing at some point. I've always found that the ol' faithful Thermalright Ultra 120 Extreme is consistently the best cooler for fan testing as it scales extremely well with more powerful fans.

    This evening I found a little time to take a peek at the two fans that make up the GFC1212DW. I planned to remove the impeller but that monstrous circlip had other ideas, and I don't have circlip pliers to hand. Here are some test results from the box, and a couple of photos:

    Front fan (by itself) achieves 23mmH2O at 4,900rpm. Rear fan achieves 17mmH2O. Together they manage 33mmH2O, which is bang on spec. It's obviously not a fan designed to develop high static pressure, but it can sure as hell move air.

    For scale, the smaller ball bearing is from a "standard" 120x25mm server fan.
    [​IMG]

    [​IMG]
     
  20. Goatee

    Goatee Well-Known Member

    Joined:
    19 Apr 2015
    Posts:
    1,590
    Likes Received:
    556
    The use of an AIO attached to a fixed heat source is more so you can measure the coolant delta after equilibrium is reached. The repeatability is much better than using PC and you can just create a testing rig and use it for testing fans.

    Low water delta at equilibrium = better cooling from the fan.

    Those pictures are awesome!
     
    LennyRhys likes this.

Share This Page