Connect with us

Help with a fan controller project

Discussion in 'General Electronics Discussion' started by Kin3TiX, Mar 29, 2011.

Scroll to continue with content
  1. Kin3TiX

    Kin3TiX

    3
    0
    Mar 29, 2011
    Hello all, I am a new member and am looking for help on a project I am working on. I am an engineering student and this is an independent project I am taking on for a class, as well as something I am personally interested in pursuing.

    Essentially I want to make a fan controller unit I can mount into a drive bay on my PC which will provide a front panel with three knobs, a switch and an lcd temperature readout. The unit will be powered through a 4 pin molex adapter that I intend to pull off the PSU or the PC. I will have three PWM fans hooked up to the unit and the three knobs will control the speed of each fan. If the switch is thrown the unit goes into "auto" mode which attempts to regulate the case temp by automatically modulating the speed of the PWM fans through the temperature reading being compared to a desired value. The temperature will be measured through a thermistor.

    I have an Atmega 644PA microcontroller that we use in class and that I need to use for the project. The chip can provide 3 PWM signals and has ADC functionality for the manual knob speed controls and the temperature readings, as well as being able to drive the LCD screen.

    My question is more about the actual power circuitry involved. I intend to use the 5V line on the molex adapter to power the microprocessor, but each fan requires 12V power and I do not know enough about power electronics to know how exactly I can power the fans off the 1, 12V line on the molex adapter. Is this possible? Also, does the "sense" pin on the PWM fans (aka tach line) need to be utilized or can it be floating?

    Any advice would be helpful, I am trying to figure out where to start.

    Thanks
     
  2. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    The sense pin provides speed feedback. If you don't actually need to know the speed the fan is operating at, then I see no reason to use it. Note that there are 2, 3, and 4 wire fans. The 4 wire fans have one wire for sense, and another for control.

    Determining if you can use the internal power supply is as easy (and as hard) as determining if the extra power required by your fans and circuit will take the power supply into overload. It's very unlikely it would, given that fans are generally low power devices. So unless you have *huge* high powered fans, or an almost overloaded power supply, yo should be fine.

    I would take the power from a disk drive connector. This is straight from the PSU and you'll get 5V and 12V.

    Controlling 3 fans to achieve 1 temperature probably means that you would be setting each to the same relative speed (e.g. all to 40%, or all to 80%) which makes things somewhat easier.

    Controlling something like temperature by varying fan speed is not as easy as it sounds. You may wish to do some research. You may not want the fans going from 0% to 100% and back again as the temperature crosses the temperature you require.
     
  3. Kin3TiX

    Kin3TiX

    3
    0
    Mar 29, 2011
    Thanks for the response, and yes I plan on using a disk drive connector for the two lines. The PSU will be able to handle it cuz im just replacing the fans that are currently running in my base with PWM fans of the same type.

    As for the difficulty in the fan speed modulation to achieve a desired speed, I do have my doubts as well as to whether I will be able to accomplish this goal, but my teacher wants me to try. My original idea was just to have the fans run at a speed proportional to the detected temperature, so if I cannot achieve accurate or stable fan responses doing it the harder way, I will probably just go with a proportional type fan response.

    As far as the power goes I am forseeing problems powering three 12V fans (which is standard for PWM fans) with the one 12V line I can pull off the disk drive power. If I hook the three in parallel with the 12V line and the GND line will I be able to supply the necessary voltage?

    Also, do you think It would be possible to utilize the sense (tach) line on the fans to possiblly create some sort of closed loop control where I could monitor current fan speed to better reach a desired temperature?
     
  4. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    Placing them in parallel is what you do to ensure that they all get 12V.

    There is no real problem in turning the fans on when the temperature exceeds a certain temperature, and turning them off when it falls below that temperature other than the fact that fans suddenly running at full speed or suddenly turning off will be far more noticeable that fans running at a constant speed.

    You might decide to set the RPM based on temperature, say 0% at temperatures less than 40C, and linearly increasing to 100% at 50C.

    O you could investigate creating a PID controller.

    I would call the latter approach the "correct one", but it is by no means trivial and is not necessarily going to be better for your computer -- it's just the most elegant.
     
  5. Kin3TiX

    Kin3TiX

    3
    0
    Mar 29, 2011
    So would you suggest implementing software control? I am studying systems and control design atm and I think I would like to explore this possibility. I am thinking I could use the tach on the fans and the temperature reading to do some comparisons to a desired level and then design some sort of a software implemented feedback system to determine the error and ramp the fan speed accordingly.

    Thanks for all the advice!
     
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

-