Connect with us

Most efficient fan control method?

Discussion in 'Electronic Design' started by Michael, Jul 11, 2003.

Scroll to continue with content
  1. Michael

    Michael Guest

    Hi - I'd like to find the most efficient possible method for manually
    controlling two fans (seperately) in a computer system. The two fans are
    these: http://vantecusa.com/sf12025l.PDF. They operate at 12V with a
    current of 0.08amps. My first impulse was to use a simple linear voltage
    retulator like teh LM317 for each, with a potentiometer setting the
    voltage, but then I realized that that would produce heat in the computer -
    so the slower I ran the fans, not only the less heat they would remove from
    the case, but also the more heat the voltage regulators would create in the
    case! And I just don't like that idea too much. So I got to thinking about
    PWM fan control - and I even posted here about it not too many weeks ago.
    But then yesterday I thought of one other thing: Why not try to smooth out
    the signal coming from the PWM circuit? That way - you wouldn't have excess
    heat production, but you would also not suffer the problems that motors
    have running at low speeds under PWM control. I was thinking I'd use a
    mic502 for this. So - does this sound like a reasonable plan? Or am I
    missing some benefit of actually having a PWM signal controlling fan?
    Thanks for your input!
     
  2. Michael

    Michael Guest

    Damn! and I thought I was going to be a millionaire with that idea! (no
    not really =( ) Anyways... Is there any benefit to not smoothing out the
    PWM output? And can anyone point me towards the proper way to choose the
    right size of inductor and capacitor? Thanks!
     
  3. cpemma

    cpemma Guest

    One benefit of *not* smoothing is that the motor gets a series of 12v kicks,
    and normally will start & run at a lower average than from linear DC. I've
    run fans down to 3v cf 5v on linear. Downside is more noise (clicking or
    growling) at low speed and possibly bearing wear, and any speed sensor won't
    work.

    Most of the motherboard pwm fan speed control systems use a very high
    frequency (>2kHz) and smooth it back to a rough DC, IIRC with just a small
    electrolytic (32uF?) cap across the fan. This means the fan electronics get
    a continuous supply, so the speed sensor still works. Search out the Winbond
    83782d datasheet for an example circuit.

    A better system is to integrate the usual low-frequency PWM signal with an
    op-amp, again just to allow speed sensor to function. Look in the MIC502
    Application Note #34
     
  4. Active8

    Active8 Guest

    i'm sorry to have to post this without providing links, but i would
    suggest subscribing to free publications like EEM, Electronic Product
    News, etc. one of the big things these days is SMPS chips. there's a big
    demand for efficient regulators, so there's always a new chip or more.

    i remember seeing a dual CPU fan controller. i'm sure you can find one
    of these if you search hard enough.

    mike c
     
  5. CFoley1064

    CFoley1064 Guest

    Hi - I'd like to find the most efficient possible method for manually
    If you're looking for theoretical minimum power, you will have to consider the
    time your FET or transistor is spending in the linear or non-saturated, non-off
    region. For any PWM controller, that is a consideration, and should be added
    to the "waste heat" involved.

    If you're interested in maximum efficiency, why don't you try a "bang-bang"
    on-off controller with hysteresis, such as the Maxim MAX6665. It has
    factory-programmable trip points, pin-selectable hysteresis, an open-drain
    output that can switch 24V at .25A, and only consumes .065 mA supply current.
    Good luck, and I hope you find the Grail. I think this might be it, at least
    as far as commercially available one-chip solutions.
     
  6. Spajky

    Spajky Guest

    you can get 2 (1 via 10 poz) rotational micro switch & 20 cheap diodes
    1N4007. connect them between contacts in serial conduct way with a
    fan. On any diode there is a drop of 0,8V ... but IMHO would be a 5-7
    poz switch enough & combining No of diodes betwwen contacts... no
    extra heating on diodes!

    -- Regards, SPAJKY
    - http://freeweb.siol.net/jerman55/HP/Spajky.htm
    Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
    E-mail AntiSpam: remove ##
     
  7. Fred Bloggs

    Fred Bloggs Guest

    Well- whatever the hell you're saying, at least you're enthusiastic
    about it...
     
  8. Ian Stirling

    Ian Stirling Guest

    Look up
    "switched mode" "boost converter" design
    on http;//www.google.com/ to get some idea on how these work.

    You probably don't want to do this.
     
  9. Spajky

    Spajky Guest

    these ones are 1A Si diodes ! no worry about that much IMHO ...


    -- Regards, SPAJKY
    - http://freeweb.siol.net/jerman55/HP/Spajky.htm
    Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
    E-mail AntiSpam: remove ##
     
  10. cpemma

    cpemma Guest

    There's no "extra heating" with any controller. If the motor voltage is
    reduced, the current is reduced, so the overall watts & heat falls.

    And diodes still each lose (Vdrop x I) watts, same as a 317T, etc. You just
    don't need to heat-sink a 1N4001.
     
  11. cpemma

    cpemma Guest

    Been very practical for me for the last few years. Switched Diodes - The
    Diobus ©: http://www.cpemma.fsnet.co.uk/sdiodes.html ;-)

    1N4001 drop is 0.75v @200mA, which is typical PC fan ball-park.
     
  12. Spajky

    Spajky Guest

    just an approx. calc. :

    fan @ 11V/200mA regulator drop 1V = heat power 0,2W
    7V/150mA 5V 0,75W
    5V/100ma 7V 0,7W

    for diodes instead of regulator (same heat power in TL but partially
    much less per diode, so they do not need any HS like regulators on
    them - (6 for 5Vdrop, 9 for 7V drop) price same TL like for regulator
    IMHO

    -- Regards, SPAJKY
    - http://freeweb.siol.net/jerman55/HP/Spajky.htm
    Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
    E-mail AntiSpam: remove ##
     
  13. Well it *ain't* typical diode ball-park! Spajky stated that *any*
    diode has a 0.8V drop, which is palpable nonsense.
     

  14. Erk - you guys can make a simple thing so damm complicated. First off,
    why do you need to speed control a fan anyway - if it needs two fans,
    then run them. If it needs one fan, run that. Drill a few more holes
    in the case while you are at it - does wonders for heat removal. Make
    sure the PC is not burried in pizza boxes or piles of paper etc.

    The most efficient switching method is called a "switch" - stick it
    on the front panel, use fingers to move up/down as required. Has
    inbuilt memory - will always stay in its set state. If you want to be
    really hi-tech, get a bi-metallic thermostat and stick it on top of
    whatever gets hottest - use that to switch them "on" or "off".

    73 de VK3BFA Andrew.
     
  15. Spajky

    Spajky Guest

    :) that I was reffering too ...

    -- Regards, SPAJKY
    - http://freeweb.siol.net/jerman55/HP/Spajky.htm
    Celly-III OC-ed,"Tualatin on BX-Slot1-MoBo!"
    E-mail AntiSpam: remove ##
     
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.
Electronics Point Logo
Continue to site
Quote of the day

-