I have tried to put together a circuit based off Grumpy_Mike's PWM circuit but i am pretty sure i have stuffed something up.
Will the circuit diagram below work? With the right MOSFET would i be able to power the Delta fans?
I think you may indeed have stuffed some things up. It also looks overly complicated. I would start with something simpler:
At 4A, I would take care to keep the digital (board) ground quite separate from the high-current ground coming from the Q1 MOSFET. Do connect the 12V and 5V supply grounds together, but connect them together at one point, the power supply, then run one ground lead to the board and a separate ground lead to the MOSFET (and 47k resistor). The second ground lead will have quite a few bumps and things due to the high current switching, and you don't want that getting to the digital board.
Hi RuggedCircuits,
The point about the schematic is that you don't want to use PWM to apply to the fan, as the fan manufacturers (Sunon) recommend that the best way to control the fan speed (for reliability) is to use a DC level. Therefore, the schematic was based on one that smoothed the PWM to a DC level before applying it to a fan. In order to reduce the capacitive loading this was buffered by a transistor to allow smaller caps to be used. The P channel FET also allowed ground referenced current monitoring (required in the original application)
Lxmyers :- You could use that opto coupling output transistor as the FET drive transistor and simplify the circuit. You have drawn an N channel FET but you need a P channel one here.
The other circuit would work as well but this would apply PWM signals straight to the fan, and you will probably not notice that much speed control as a result. This is because the fan's inertia will keep it going at nearly the same speed as it is just kicked by the PWM signal.
Do you think i should add an octocoupler for MCU isolation or is that a bit of overkill?
My guess is it probably won't be necessary if you pay attention to my comments on grounding. And it's "optocoupler", not "octocoupler"
The point about the schematic is that you don't want to use PWM to apply to the fan, as the fan manufacturers (Sunon) recommend that the best way to control the fan speed (for reliability) is to use a DC level. Therefore, the schematic was based on one that smoothed the PWM to a DC level before applying it to a fan.
Oooo...didn't know that about the PWM. But doesn't this mean the MOSFET will be operating in linear mode and dropping excess voltage, thus likely to get quite hot?
I think a design based on a switching regulator would be much more efficient, though admittedly more complex.
But doesn't this mean the MOSFET will be operating in linear mode and dropping excess voltage
Yes it will it will need to be generously supplied with heat sink and have a good rating.
I think a design based on a switching regulator would be much more efficient, though admittedly more complex.
I think you are right, do the switching in the digital mode then smooth before applying it to the motor. So basically you need a computer controllable PSU.
Computer Controlled PSU? Wow, this is not a easy as i thought.
So PWM is the easiest way to control a fan but it is not particularly good for the fans lifespan.
Mike, you suggest that the voltage should be smoothed before applying it to the the fan but as a result of the smoothing, the MOSFET will operate in linear mode, getting quite hot.
Is PWM the best way to control the speed of a fan? Is the heat produced by the MOSFET normal and not something to worry about if a heatsink is used?
I changed the circuit a bit, am i getting closer to something that will work?
Is the heat produced by the MOSFET normal and not something to worry about if a heatsink is used?
Yes it is normal in the sense that it is expected. However, the heat sink will have to be big enough and the power rating of the FET will have to be big enough. As the power of something gets higher this pushes the power needed to be burnt off by the FET.
It is possible that a 45W fan controlled in the liner mode might end up having to burn off 25Watts or more. This is a lot of heat which why a controlled switch mode supply might be the best bet.
Hmm...your circuit is not likely to work in my opinion. I don't see any way for the voltage across the capacitor to decrease -- once it reaches 12V it's going to stay there. Maybe you want that 4k9 resistor to ground, not 12V.
At 4A I don't think a linear-mode MOSFET is going to be the way to go. I'd bite the bullet and get into switching regulators, e.g., the Linear LTC3605. It's easy to adjust the output voltage by varying one of the feedback resistors, e.g., by using a digital potentiometer.
