how "positive" must airflow be for a tower case to keep dust out?

Discussion in 'Asus' started by Winey, Jul 26, 2004.

  1. Winey

    Winey Guest

    I'm upgrading a case (and the system inside it)? Actually two such
    systems. I want to have positive air pressure in the case sufficient
    to keep out dust. If you know this design, it has vents on the sides
    of the case near the top, plus a row of vents in the back above the
    adapter cards:

    Both cases will have an ASUS A7M-w266 D,TBred 2000, no overclocking.

    Power Supply is an Antec TruePower 550. (assume 0 cfm out at lowest
    fan speed, 40 cfm at highest fan speed)

    Cases are an Addtronics 7896A and a 6896A.

    Assume roughly 140 cfm in. (2 120 mm Panaflow "L" model)

    So, my question is:

    How positive do I need to keep the case air pressure, to keep dust
    out? That is, how much more should the cfm in be compared to the cfrm
    out?
     
    Winey, Jul 26, 2004
    #1
    1. Advertisements

  2. Winey

    Paul Guest

    I just tried "computer dust positive pressure" on altavista.com
    and one of the first hits was this.

    http://www.dustfreepcinfo.com/pages/184710/index.htm

    They use filters before their fans that pressurize the cabinet.
    Thus, clean air is used, and blowing that clean air under positive
    pressure, through any opening in the cabinet, prevents dust
    from entering.

    If your case blows dirty air in, I don't think it really matters
    what differential there is between "in" and "out". As long as there
    are dead spots in the case, where the air velocity is lower than it
    was originally, the particulate will settle out. (My thinking here,
    is that dust works on the same principle as silt in a river - as
    soon as the river widens and the water velocity drops, the silt
    falls out and collects on the bottom of the river.)

    To implement a solution like those dustfreepc people, you will
    need micron filters, a plenum, and your large 120mm fans. A plenum is
    simply a length of plumbing leading from the filter to the fan, and
    keeps the filter from getting too close to the fan blades. About
    3" of plenum between the filter and the fan should be good. You
    can try mounting the filter right on top of the fan, but you'll
    probably hear more noise coming from the fan that way.

    As soon as you use filters on a PC case, then you've got a maintenance
    item. The filters have to be cleaned or changed every x months. You
    have to remember to do it. If you forget, the computer might overheat.
    The most endangered component is probably the disk drive, and
    the computer will continue to run quite happily under conditions that
    are trashing the disk.


    Some other threads from my search:

    http://forums.pcper.com/showthread.php?s=&postid=2029294

    http://forums.overclockersclub.com/?showtopic=8412&st=10&

    The interesting tidbit on the following site is

    CFM = 3.16 x Watts / Delta_T_degrees_F

    http://www.chassis-plans.com/cooling_and_noise.html

    Using the AMD recommendation of a max case temp rise of 7C, which
    is 12.6F, and assuming 200W for a high end PC, we get 50 CFM.

    This one is a home brew solution, using an air cleaner filter
    replacement as the filter for the air. Positive case pressure
    is created by two identical fans, with one running 200RPM higher
    than the other:

    http://www.silentpcreview.com/Sections+index-req-printpage-artid-144.html

    So, if the air is filtered, then a positive air flow seems to be
    the solution offered by commercial designs. For unfiltered
    situations, a balance between in and out, or a slight negative
    pressure seem to be favored.

    I can tell you that one computer I use, which has only exhaust
    fans, is filthy inside. Yet, I've never had any trouble with
    the floppy or the CD on that computer.

    This is one of those questions, where if you ask twelve
    people, you'll get a dozen different answers :)

    HTH,
    Paul
     
    Paul, Jul 26, 2004
    #2
    1. Advertisements

  3. Winey

    Wizard Prang Guest

    Wizard Prang, Jul 26, 2004
    #3
  4. Winey

    BigBadger Guest

    Unless you have filters on the air inlets the dust is gonna get into the
    case anyway.... and if you do have filters then the performance of the fans
    will be much reduced.
    If you have filters and want to ensure positive pressure then the only sure
    fire way is don't have any exhaust fans.
     
    BigBadger, Jul 26, 2004
    #4
  5. Winey

    Michael S. Guest

    How would that keep out the dust?? It would seem if any fans are blowing in,
    they would bring in dust with the inbound air and any positive air pressure
    inside the case would limit the amount of incoming air for ventillation.
    Don't get me wrong, I would like to see a solution beside airflow limiting
    filters since I have cleaned out dustbunnies of record size and number from
    a few hundred enterprise computers over the years and I would like very much
    to keep the dust out of my new build.

    MkeSp
     
    Michael S., Jul 26, 2004
    #5
  6. Winey

    borolad Guest

    I do remember reading an excellent site years ago that gave
    scientifically measured answers to your question - but I don't have
    the URL anymore, if I come across it I'll post here.

    I do however want to make the following observations :

    - at 1+1/2 to one going through filters and case perforations you will
    need probably to go for 2+ 1/2 to one to achieve the required CFM.
    Those filters need to be external, internal ones hardly ever get
    changed often enough, people just can't be bothered opening the case.

    I'm sure you won't want 5.5k deltas so bigger sizes the better running
    at 5 or 7 volts will keep the noise down and the flow up. I've been
    fitting 120's for about 7 years now but almost without exception this
    requires case mods.

    - all cases need a gentle laminar airflow input, a simple ' hot air
    rises ' is often the cause of 'dead spots', all cases have thousands
    of holes and air will take the path of least resistance. Baffles can
    be used to eliminate re-circulation.

    - chassis 12 volt cooling fans draw very little power; less than two
    watts for a 60mm fan, less than four watts for a 120 mm fan. Even if
    you have six 120 mm fans and four 60 mm ..fans, the total power draw
    from the 12 volt line will be less than three amperes.

    - removing the chassis perforations from both input & exhaust fans
    will boost the throughput substantially, this would particularly be
    the case for the exhaust / PSU fans. It's a contradiction in terms
    because to maximize airflow, any resistance should be minimized but
    filtered positive pressure does exactly the opposite.

    input fans should be sealed against the panel, if they are not air
    leaks at a rate of knots, reducing input by a considerable margin.
    Pairs of fans should both suck - or - blow otherwise the tend to
    cancel each other out.

    - you don't give the VID, but modern AGP cards generate lot's of heat,
    if you can, leave the slot next to the AGP 'open' to allow airflow',
    you should do this anyway to avoid sharing an IRQ address with AGP
    slot.

    - if you can the rule is mount your MOBO low and your Drives high, all
    drives produce a lot of heat particularly 7-10k RPM, I find the two
    platter 3 year warrant ' spinpoint ' line from samsung are the coolest
    drives I've ever used and certainly they are very very quiet.

    Just audible is 10 dBA
    Soft whisper at 15 feet is 30 dBA
    A quiet room is about 28dBA
    A noisy drive measures about 36-38 dBA
    A quiet office is about 40 dBA
    Air conditioner, normal speech, 60 dBA
    Noisy restaurant, freeway traffic, noisy office, 70 dBA
    Hearing protection recommended at 80 dBA
    Lawn mower on grass is 85 dBA
    Heavy truck in traffic measures 90 dBA
    Rock concert is 110 dBA
    Auto horn at 3 ft, maximal vocal effort results in 120 dBA
    Thunderclap is 130 dBA
    Jet air ops on a US Navy carrier deck is 140 dBA

    - look a series of numbers keeps going through my head, I've no idea
    how factual / reliable they are but I seem to remember reading
    somewhere that ....... " 35 CFM of airflow is required for a system
    that dissipates 200W "

    All the above ramble is experience, guesstimates and my personal
    twopenny worth.

    BoroLad
     
    borolad, Jul 26, 2004
    #6
  7. Winey

    Winey Guest

    As the OP, agree 100%. I guess I should have stated my question a bit
    differently:

    How can I keep * a lot off the dust* from settling in my case.

    If there is positive air pressure, some of the dust might get
    exhausted, and some will simply have to be cleaned out every X months.

    Both my Addtronics cases have a cleanable mesh filter on the
    removeable front of the case. It does trap some dust, and the total
    filter area is big enough so that a small amount of dust obstruction
    probably doesn't affect fan performances.

    These cases aren't 100% dust-free in use, but they are a lot cleaner
    than systems without any dust filters. And that is probably good
    enough.
     
    Winey, Jul 26, 2004
    #7
  8. Winey

    Amir Facade Guest

    I think the exhaust hose on a medium size shop vac would do the trick
    nicely.

    Amir
     
    Amir Facade, Jul 27, 2004
    #8
  9. Winey

    Ben Pope Guest


    You included over 100 lines of previous post, what ARE you talking about?

    Ben
     
    Ben Pope, Jul 27, 2004
    #9
  10. Winey

    Amir Facade Guest

    I'm upgrading a case (and the system inside it)? Actually two such
     
    Amir Facade, Jul 27, 2004
    #10
  11. Winey

    Ben Pope Guest

    Bloody hell. learn how to post.

    The exhaust system from a vac would be warm air and not ideal. Might be a
    bit noisy too.

    Ben
     
    Ben Pope, Jul 27, 2004
    #11
  12. Winey

    NEM Guest

    Hi Michael S.,

    By making the room the computer sits in have less dust to begin with.
    I'm NOT talking about a million dollar Clean Room, but there are things
    that can be done to minimize the amount of dust that could get inside.

    Carpet, as one example, is bad. As the carpet gets older, it breaks off
    and becomes airborne and gets sucked into a computer. Pets that think
    the computer case is a companion, since you spend more time with the
    computer then with them. <G> Hey, they get feed, what more do they want.
    Heh. Consider what makes up dust, human byproduct. Stop shedding. <G>

    The CPU fan in this computer was the main culprit. The fan just plasters
    dust into the heat sink and no longer allows the air to flow through the
    fins.

    Or, stop using fans and use water cooling. Or pantyhose. I read an
    article that suggested that the material used to make pantyhose is
    perfect for allowing air flow and keeping a fair amount of dust out of
    the case. A simular material is used in recording studios in front of
    microphones.

    One way or another, it is a good idea to go into the case and clean it
    out. There is no reason to spend less time in there cleaning, then you
    would your own living space two or three times a year. <G>
     
    NEM, Jul 28, 2004
    #12
  13. ever thought of vacuming out the guts once in a while?
     
    notritenoteri, Jul 28, 2004
    #13
  14. Winey

    Ben Pope Guest

    Yeah, but the static build up worries me.

    Of course compressed air (or LPG or..) would be more suitable

    Ben
     
    Ben Pope, Jul 28, 2004
    #14
  15. Winey

    Ed Light Guest

    Air cleaner with filter!
    Ion generator to stick dust to the walls!


    --
    Ed Light

    Smiley :-/
    MS Smiley :-\

    Send spam to the FTC at

    Thanks, robots.
     
    Ed Light, Jul 29, 2004
    #15
  16. Brush the dust out with a soft artists brush. Have a vacuum cleaner
    sucking beside the case, to stop you inhaling the dust, and stop it from
    just settling back where it'll go through the computer again.
    Dust was the reason the original IBM design, had the rear fans blowing
    into the case. The idea was that the intake here was higher, and further
    from the carpet/user, and the outflow of air at the front would keep the
    floppy disk from filling with fluff. It was one of those ideas that was
    'forgotten' with the passage of time.

    Best Wishes
     
    Roger Hamlett, Jul 29, 2004
    #16
  17. Well, they do. But the other half of the story that's untold there is that,
    after making the design, they also learned that people hate having their
    computers 'blow' on them.
     
    David Maynard, Jul 30, 2004
    #17
  18. Winey

    Ben Pope Guest

    BTX design should improve a whole load of things, apparently.

    Ben
     
    Ben Pope, Jul 30, 2004
    #18
  19. Winey

    Bx.Cornwell Guest

    you'd think that computer case engineers (designers?) would at least be
    keeping an eye on things like that...
     
    Bx.Cornwell, Jul 31, 2004
    #19
  20. [snip]

    Good idea.
    [snip]

    Sounds typical. But the specific details generally aren't really relevant to
    your core approach -- *all* cases have sufficient vents and gaps of various
    sorts to allow all manner of air ingress, if a (partial) vacuum exists inside
    the case.
    [snip]

    Offhand, that seems like serious overkill, unless there's something else
    you're not telling us. But that's another issue.
    [snip]

    Both assumptions are probably too extreme. AFAIK, the PSU fan(s) will never
    shut off completely (at least under the PSU's own control), so no "0 CFM"; and
    if you believe any manufacturer's "maximum CFM" ratings, I've got a toll
    bridge to sell you in New York.
    [snip]

    Not familiar with that particular model; but *if* both of those fans are
    drawing through proper filters, that should be more than adequate unless
    you're doing something silly like "serious" overclocking.
    It doesn't work quite like that.

    First, "flow" and "pressure" are two different things. Further, they tend to
    oppose each other, in any given situation.

    Second, as implied above, manufacturer "CFM" ratings are mostly a crock, in
    part because *any* "raw" CFM rating presented without all the context details
    (i.e., what sort of test jig; how much pressure differential is the fan
    maintaining, etc.) is effectively near-meaningless. And even then, the
    typical test jig doesn't come close to representing the real-world conditions
    under which an *installed* fan operates. (For that matter, I strongly suspect
    that at least most of the fans sold for hobbyist computer applications are
    *never* actually tested on a proper flow bench.) At best, these ratings can
    be a *very* rough guide to "this fan is more powerful than that fan", within a
    single manufacturer's line.

    Besides, the degree of pressure differential is not really important, as long
    as it is even slightly positive (in fact, as long as it is even non-negative).
    Even a zero pressure differential would still be sufficient to keep from
    drawing dirty air into the case via all those assorted vents, gaps, etc.; and
    that's all you're really trying to do WRT to managing the internal case
    pressure. Beyond that, it mostly becomes a question of trading off total air
    flow (i.e., cooling) vs. noise. All other things being more-or-less equal,
    more fan speed (and/or bigger fans) == more flow == better cooling (presuming
    decent overall system layout, so that you don't get serious internal
    "stagnation" areas).

    --

    Jay T. Blocksom
    --------------------------------
    Appropriate Technology, Inc.
    usenet01[at]appropriate-tech.net

    "They that can give up essential liberty to obtain a little temporary
    safety deserve neither liberty nor safety."
    -- Benjamin Franklin, Historical Review of Pennsylvania, 1759.
    - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    Unsolicited advertising sent to this domain is expressly prohibited under
    47 USC S227 and State Law. Violators are subject to prosecution.
     
    Jay T. Blocksom, Aug 1, 2004
    #20
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.