Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[joecool85]

Alright, I've come up with a fan controller that I'm happy with.  It can run up to 1amp of current for a 12v fan so long as it is heatsinked.  With the fans I've played with, 4.5v is when they turn on and start blowing any sort of noticeable amount and 12v is full tilt.

Circuit Approximations:

When R3 = 0K
4.5v happens at 40C/104F
12v happens at 75C/167F

When R3 = 5K
4.5 happens at 28C/82F
12v happens at 63C/146F

When R3 = 10K
4.5 happens at 18C/65F
12v happens at 55C/131F

Notice how as the resistance increases, the turn on/full tilt temperatures decrease.  To make this more natural in how it works, I would wire the pot backwards so that all the way clockwise would be the hottest temps and all the way counterclockwise it would turn on the soonest and run the coolest.
Enjoy!

[J M Fahey]

Cool, Joe .   

[Roly]

Good one Joe; I love it when a chip (like a regulator) gets used for something the maker never imagined.   

[joecool85]

Good one Joe; I love it when a chip (like a regulator) gets used for something the maker never imagined.   

Thanks.  Your LM3914 circuit does that to an even stronger extent though!  (Especially with the use of a diode or transistor-as-diode for the temp sensor).

This circuit is well suited to more load than the LM3914 though, as well as keeping things very simple.

[joecool85]

I was worried about using a linear regulator like the LM317 for a fan controller since they lower voltage by dissipating it as heat, however, since brushless fans drop their current draw significantly as voltage goes down, it is still a net power savings running your fan at a lower voltage via LM317 rather than run it full bore at 12v all the time.  In fact, it uses almost exactly half the wattage at 6v than 12v.  Granted, if you used a PWM controller and ran it at the same rpm as 6v would net, you would use about 1/4 the wattage as running full tilt at 12v.

Here is some good info on the subject, from http://www.maxim-ic.com/app-notes/index.mvp/id/1784:

Linear Regulation
As the term implies, "linear regulation" adjusts the DC voltage across the fan by using a linear regulator. When using this method, it is important to make sure the fan is specified to operate over a wide range of voltages. One major advantage linear regulation has over PWM is that it allows the use of speed and alarm sensors. Unfortunately, linear regulation also has its drawbacks: mainly power dissipation in the pass element, as well as startup and stalling issues.

Linear regulators control the DC voltage across the fan. They do this by dissipating power in the form of heat. It probably seems silly to generate heat in order to cool something down. But it is not as ridiculous as you might think. During maximum and minimum cooling, power dissipation will ideally be zero. During maximum cooling, the pass element is fully on, so the voltage across it is nearly zero. Zero volts means zero power dissipation. During minimum cooling, the pass element is off (zero current flows), so again power dissipation is zero. As previously discussed, the current draw of the fan can be approximated as a linear function of the voltage applied, making it look resistive. With this in mind, worst-case power dissipation occurs roughly when the voltage across the fan is one-half its maximum operating voltage. See Figure 5. This means worst-case power dissipation in the pass element can be estimated by the following equation: P = 1/4(VMAX × IMAX), where IMAX and VMAX are the rated voltages and currents of the fan, respectively. For example, a 1.2W fan (12V at 98mA) will have worst-case power dissipation across the pass element of only 300mW when running at 6V with a 12V supply. It is comforting to note that maximum heat dissipation in the fan circuit occurs during minimal cooling requirements. Also, even though a power-dissipating device is being used, there is still an overall power savings when fan speed is reduced. See Figure 6.

[Roly]

It's not like the series-pass element is getting hottest when the fan is stopped, so there is a degree of self-regulation here if the series-pass element is also exposed to the fan (and why would you not?).

PWM is all very sexy and all, but not around low level audio.  With any sort of controller there are potentially large voltage swings, possible control "chatter", and the need to keep anything like this right out of the audio path.  This more than justifies the use of a linear controller.

[joecool85]

It's not like the series-pass element is getting hottest when the fan is stopped, so there is a degree of self-regulation here if the series-pass element is also exposed to the fan (and why would you not?).

No, but it is hottest at half speed (not full speed).  Brushless fans require less amperage as voltage drops which helps keep things under control since the lower voltage requires the regulator to burn more voltage as heat.  But at 6v a 200ma fan still needs 100ma.  Assuming a 14v power supply (so that you get 12v after LM317 regulation at full tilt), that means that 8v needs to get shaved off.  Multiplied out that is 0.8w of heat.  Not a ton, but some.  If you were to run 5 fans that would be 4 watts of heat - enough to start mattering.  With a PWM you wouldn't have that, but as you said, you would have added complexity as well as a good chance of adding a lot of noise to the circuit - no good for an amp.

For what it's worth, I like linear regulators for their simplicity, reliability and electronic quietness.  I'm also a nut for efficiency though, and so it bothers me that anything is wasted as heat    But, it is still the best solution in my opinion for amp fan regulation.