AC fan control of CPU with temperature as feedback

Thread Starter

805.bluebell

Joined Feb 23, 2014
10
Hi,
I want to make a fan control system which modifies the speed of fan based on temperature changes and tries to maintain a specific speed.
I just don't know what grade/specifications of parts should i purchase and implement it since its my first ever project and have read about them in textbooks only.
I want to use

  • thermocouple as temperature sensor,
  • op-amp as comparator of feedback temperature

  • SCR receiving the feedback from op-amp such that its waveform is chopped and given as power supply to my fan.
please help me make this project a reality !
 

Thread Starter

805.bluebell

Joined Feb 23, 2014
10
i have referred these circuits.
i want to make a single (DC supply and motor) device as mentioned above.
i want to use an SCR which is fired by connection from the red light LED in second diagram. of course i will use a forced commutation circuit in parallel (of a capacitor and inductor of required value) with SCR.

The problem is what rating of SCR and commutation circit elements should i purchase.
u may refer these links to get technical details
http://circuiteasy.com/temperature-sensor/
http://circuiteasy.com/automatic-fan-controller/



 

#12

Joined Nov 30, 2010
18,224
Those drawings are absurd. The LM35 delivers .01 volt per Centigrade. You would have to be about 50 to 60 C to get any results from the first circuit and the LM741 op-amp is even worse. It will not sense within a few volts of ground, 200C to 300C. By then, the sensor will have melted.

Start over with what temperature range you need and we'll go from there.
 

Thread Starter

805.bluebell

Joined Feb 23, 2014
10
as for temperature range,
i will demonstrate it in the class with [cold glass of water] and [medium hot cup of tea] if you insist on higher temperature ranges.

Also, we will give normal earthing to the opamp which is available for households.
 

#12

Joined Nov 30, 2010
18,224
if you insist on higher temperature ranges.

Also, we will give normal earthing to the opamp which is available for households.
I do not insist on anything except that you not waste your time on circuits that absolutely will not work, but that is negotiable.

Earthing the opamp to the house ground will not fix anything and is not necessary.

The first drawing might barely work as a switch with very hot water. An LM34 would work even better.
 
Last edited:

AnalogKid

Joined Aug 1, 2013
10,971
In the original message you say you want to maintain a specific speed. If you realy meant a specific temperature, as in closed loop feedback control, then here is a basic approach.

Figure out how much heat you want to move. For example, if your fan is cooling a heatsink with some power transistors on it, what are the low power (for low fan speed) and high power (for max fan speed) numbers. This will tell you how much air you need to move, which translates into fan size and speed.

Figure out where the sensor goes. Temperature sensing is extremely localized around the sensor, so put it where it will tell you what you reallly want to know, the temp of only one small thing. You can put it in the middle of a large heatsink, but it will tell you the temperature of only the nearest 1/4 inch.

I prefer thermistors over over devices because even though they are non-linear, they are very accurate. Over a reasonable temperature range of 20 to 30 degC, a simple circuit can be accurate to 0.5degC without calibration.

After all of that you have a control voltage that varies as a function of temperature with everything scaled to you application. Next up is a triac power stage for the motor.

Keep in mind that you won't have much of a control range. Most fan motors (AC, brushless DC, whatever) do not like to run slower than 50% of rated speed. It is pretty easy for what you want to do to hold a sensed temperature fairly constant over a 10 or maybe 15 degC change at the sense point, but after that the fan will max out.

ak
 

Thread Starter

805.bluebell

Joined Feb 23, 2014
10
well AnalogKid,
if you are really sure that IC LM35 won't work can u please suggest a remedy for it.
I know u have suggested thermistors. But, of course the circuit elements around opamp will also change accordingly.
Please guide me regarding the thermistor i should use and the th modifications which would occur as i have no idea about the tangible parts.

In the original message you say you want to maintain a specific speed. If you realy meant a specific temperature, as in closed loop feedback control, then here is a basic approach.

Figure out how much heat you want to move. For example, if your fan is cooling a heatsink with some power transistors on it, what are the low power (for low fan speed) and high power (for max fan speed) numbers. This will tell you how much air you need to move, which translates into fan size and speed.

Figure out where the sensor goes. Temperature sensing is extremely localized around the sensor, so put it where it will tell you what you reallly want to know, the temp of only one small thing. You can put it in the middle of a large heatsink, but it will tell you the temperature of only the nearest 1/4 inch.

I prefer thermistors over over devices because even though they are non-linear, they are very accurate. Over a reasonable temperature range of 20 to 30 degC, a simple circuit can be accurate to 0.5degC without calibration.

After all of that you have a control voltage that varies as a function of temperature with everything scaled to you application. Next up is a triac power stage for the motor.

Keep in mind that you won't have much of a control range. Most fan motors (AC, brushless DC, whatever) do not like to run slower than 50% of rated speed. It is pretty easy for what you want to do to hold a sensed temperature fairly constant over a 10 or maybe 15 degC change at the sense point, but after that the fan will max out.

ak
 

Thread Starter

805.bluebell

Joined Feb 23, 2014
10
its the first time i have heard these IC LM series.
i need to make it clear here that temperature difference will be maintained by me no matter what the conditions are.
Only i need to demonstrate it in class that such closed loop systems actually work.
If you insist i would purchase LM34 if it works better.

I do not insist on anything except that you not waste your time on circuits that absolutely will not work, but that is negotiable.

Earthing the opamp to the house ground will not fix anything and is not necessary.

The first drawing might barely work as a switch with very hot water. An LM34 would work even better.
 
Top