fan speed control with temperature

Discussion in 'The Projects Forum' started by 805.bluebell, Mar 21, 2014.

  1. 805.bluebell

    Thread Starter New Member

    Feb 23, 2014
    10
    0
    OK , directly expressing the problem scenario .
    I have a temperature sensor [IC LM 35], comparator [IV 741 opamp] , a fan [18V D.C fan], thyristor [SCR TYN612M].

    Now, i want to use them over to form a circuit capable of controlling my fan speed in proportional to temperature.
    Is there a way out to form a simple but effective circuit out of this?
    HELP !
    :confused:


    i know what to do but don't know how to do (eg, i don't know what values of L-C pair should i employ for simple commutation of SCR)
     
  2. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,969
    744
    Yes use an op amp to amplify and buffer the dc voltage from the LM35, and feed this into an emitter follower transistor, so as the temperature increases, the speed increases also.


    similar to this circuit, add the Npn emitter follower

    forget the thyristor...
     
    Last edited: Mar 21, 2014
    805.bluebell likes this.
  3. joeyd999

    AAC Fanatic!

    Jun 6, 2011
    2,674
    2,717
    FYI, a simple proportional controller will not maintain a constant temperature. The temp will always be a little higher than the set point. You could try to compensate for this by lowering your set point, but there will then be variations based upon ambient temperature.

    Also, for small differences between the set point and temperature, the fan will not get enough power to start, and fan may lock up if the temperature increase occurs slowly.

    Consider a PID (or at least a PI) controller. This can be done in analog. One of my favorite articles by the late and missed Bob Pease is What's All This P-I-D Stuff, Anyhow?. Unfortunately, the online page doesn't seem to have the circuit diagrams. Bummer. Maybe someone else can fill in the blanks.

    Edit: The images are available if you Google for "Bob Pease PID" and search for images.
     
    Last edited: Mar 21, 2014
    805.bluebell likes this.
  4. wayneh

    Expert

    Sep 9, 2010
    12,090
    3,027
    As DD says, forget the thyristor.

    An improvement would be to use the LM35 voltage, probably amplified by the op-amp, to control the duty cycle of a PWM signal.

    See Bill's blog on how to use a comparator and a 555 timer IC for PWM control. I haven't worked out the details, but I believe you might apply your amplified LM35 voltage to the control pin (5) of the 555 to change the duty cycle. Or maybe you just use your amplified LM35 voltage as one input to the comparator.
     
    805.bluebell likes this.
  5. Nykolas

    Member

    Aug 27, 2013
    87
    31
    As this is a DC fan the SCR can not be used. The OP did not state the temperature accuracy required.
    I did a 3-speed + off fan using a LM35 (attached to the heatsink of an audio power amp) connected to a LM3914 in dot mode. The outputs of the LM3914 drove, via PNP transistors and low-ohm resistors, the fan. E
     
    805.bluebell likes this.
  6. wayneh

    Expert

    Sep 9, 2010
    12,090
    3,027
    Lots of ways to skin the cat. That's a good one.
     
  7. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    Or, replace the 8 pin opamp with an 8 pin microcontroller.

    Then just program it to control fan speed to give the right temperature. :)
     
    805.bluebell likes this.
  8. 805.bluebell

    Thread Starter New Member

    Feb 23, 2014
    10
    0
    can i have url of Bill's blog please :)
     
  9. wayneh

    Expert

    Sep 9, 2010
    12,090
    3,027
    Go here, and scroll down to Figure 10.4 in Chapter 10. It's on page 2 for me.
     
  10. 805.bluebell

    Thread Starter New Member

    Feb 23, 2014
    10
    0
    As SCR needs to be ruled out of scene i am thinking of using BJT transistor with enough rating to start working with LM35 output being fed to its base.
    Will it work?

    [​IMG][​IMG][​IMG]

    If yes, then what rating of transistor should i choose and are there any other pheriferals(resistors,capacitors etc) required
     
    Last edited: Mar 22, 2014
  11. bertus

    Administrator

    Apr 5, 2008
    15,645
    2,344
    Hello,

    Your image does not show, so I have downloaded and cut the essential part of it to post.

    [​IMG]

    You will need some circuitry between the LM and the BJT to compare the temperature and drive the BJT.

    Bertus
     
  12. joeyd999

    AAC Fanatic!

    Jun 6, 2011
    2,674
    2,717
    Yup. Those boobs will work just fine... ;)
     
  13. donpetru

    Active Member

    Nov 14, 2008
    186
    25
    A good solution requires knowledge of characteristics:
    1). input current fan vs. speed fan. Here it would be useful to know that: air flow values is required at certain temperatures. According to this aspect, then you can find some points on the current drawn by the fan. These points, actually some fan input current values, ​​will be used in the next step;

    2). choose a bipolar transistor, not a thyristor. Depending on the maximum current and voltage drawn by the fan running, you can choose a transistor. In datasheet transistor will be looking to exit characteristic, more precisely: base current vs collector current. I recommend to look: base-emitter voltage vs. the current transistor collector;

    3) I saw that you want to use LM35 which gives the output a voltage proportional with measured temperature. For example, if the output LM35 is 250mV, this mean 25 gr. Celsius. If you want to apply the post no.11 schematic (there should be installed a resistor between the transistor base and the output of LM35), then you should think that the transistor will go into saturation only after it exceeds ~0.4V between the transistor base and ground. This means that the transistor will begin to enter the saturation above 40 degrees Celsius. In your application is that enough?

    And another thing, that the transistor begins to enter the saturation that does not coincide with the fan operating point, because in that threshold for starting the transistor conduction, current starts to flow through the transistor and the fan, and the current may be too low to start the fan. In this case, the fan should be running at full speed when the temperature barely exceed 50 degrees Celsius, not 40, and, return the question: In your application is that enough? Think well.
     
  14. wayneh

    Expert

    Sep 9, 2010
    12,090
    3,027
    +1
    Right, the LM35 is a thermometer (continuous output of voltage proportional to temperature), not a thermostat (on/off). As drawn you will get no fan until temperature exceeds ~60°C, putting >0.6V onto the base of the transistor.

    Also, the LM35 likely does not put out enough current to drive the bjt. The fan will never approach 100%. Depends on the details of the fan and the transistor, but not a good design without the circuitry in between as noted by bertus.
     
  15. snav

    Active Member

    Aug 1, 2011
    100
    10
    Why not use a power darlington Like a TIP120 and Bias the input with the junction voltage of a diode or Potentiometer depending on budget?
     
  16. bertus

    Administrator

    Apr 5, 2008
    15,645
    2,344
  17. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,516
    1,246
    Many DC fans will not run below 50% of their rated voltage, and most won't start in that condition. So you have to adjust your requirements. A typical fan controller runs the fan at 100% for a short time to get it moving, then backs off the speed depending on the local temperature. But they never back the speed down below 50%, so the fan is always running even it if is cold. So, what is the temperature at which you want the fan speed to start to increase above 50%, and what is the temperature at which you want the fan to peak at 100%? You can't start without knowing these two things.

    Also, while your sketch conveys what you want to do, those parts will not do it. Try this. Based on the two answers from above, use the LM35 data sheet to calculate the output voltage at those two temperatures. Then calculate the base voltages needed to get the fan to run at 50% and 100%. Since those two sets of voltages will not agree, you need a circuit in the middle to translate things. That's where the opamp comes in.

    ak
     
    Last edited: Mar 29, 2014
  18. snav

    Active Member

    Aug 1, 2011
    100
    10
    that's the reason I'm using this to drive a fan, but it ranges from 'low' at about 5.6v to full 'on' at 9.4v. Put a gain circuit for the thermister that takes place of 10k pot and you're done.
     
  19. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,516
    1,246
    Not quite done. Most DC fans *hate* pwm drive.

    ak
     
  20. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,127
    266
    Is your fan a brushless DC fan by any chance?

    These fans do not like PWM, they have internal motor controller chips that commutate the coils, they go haywire when fed PWM. You can control the speed by changing the applied voltage, but only over a narrow range. Harder than it looks at first glance.
     
Loading...