Confusing title, let me explain. I am not concerned about the power source as I can work with that or build one . This is a 5volt project.using a 3.3 volt fan...What I do need is this: a PWM that works a fan and continues to work the fan up to the point 3 minutes after the power is shut off. This is for cooling a project. So you turn off the power to the unit the fan is cooling but the fan remains on for three minutes then shuts down and stays off until the power is once again reapplied by the unit you are cooling. So far I have not yet tried to build anything. The PWM I am fine with, so don't need any help there. You might say I am becoming lazy and don't want to spend the time thinking. Not that I don't enjoy it, as I really do, but it is becoming harder for me to focus and visualize a circuit in my head any more. I can still draw out circuits from memory but finding frustrating to say the least to have to even ask for help now. Medication you know. It's like trying to build something when you are all liquored up. I don't drink either, maybe that would help. If I sit down for a few minutes I suppose I can figure it out, but what are your ideas? maybe better than mine. One Idea I was going to use a 555 timer in one shot mode. Connect the PWM to the one shot so it continuously receives pulses. It times out each pulse that activates it, at 3 minutes, checks for power on condition and receives another pulse etc. I just thought of something, brain surge, maybe I am thinking too complicated. how about this: my PWM will drive the fan and supply pulses to the 555 timer which times out 3 minutes, 555 drives a transistor as a relay driver which keeps the relay on for 3 minutes each pulse that arrives at the input of the 555 that activates the 555. The relay switch or contacts will be in series with the fan power. The driver transistor and 555 will be connected to power source that will never shut off. So as soon as power is once again applied to the PWM it activates the relay, which allows power to be applied once again to the fan.
Any other ideas that might be simpler? My PWM will use a 555 driving a mosfet running at about 25 KHz.

 

Where does the PWM come from?

What means "power off"? Just turning the PWM off?

Your last 555 idea will work (power for the fan in series with a relay contact)

What's the fan's current consumption and nominal voltage?

 

One idea you might want to try is to decouple the two altogether. Why does the fan need to go on right when the unit is powered up? What is magical about 3 minutes? Maybe today two would be fine but tomorrow four would be better. Instead, use a thermistor (or some other temperature sensing element) to measure the temperature of the air after blowing over the unit. This works best if you can identify the part of the unit that heats the most and/or cools the slowest and can place the sensor close to it (or even mount it on the hottest part of the heat sink). Then have a circuit that simply turns on the fan whenever the temperature is above a preset threshold and shuts it off after it is back below it (or a somewhat lower turn-off threshold in order to build in some hysteresis).

 

nominal fan voltage 3.3. current at full speed 2.2 amps. this fan is currently running at 20%. nothing magical about 3 minutes, could be 5 minutes. 3 seems to be enough. I like the temp sensor idea, then it could shut down sooner or later depending on the threshold. That was a good idea. The Pwm is only for providing the pulse frequency for the fan speed. I am looking to cool parts down after unit shuts off so the unit is not sitting there baking. The 3 minute idea is a fixed option but I do like the temp sensor idea and will incorporate that. I have thermistors I can use. Negative coefficient as I recall. You know that is a good idea cause I can control the fan speed while it is on therefore most likely no need to run it after it shuts down. I suppose if I cant keep it cool at that point, something is definitely wrong and maybe would need to redesign the whole thing.it helps to hear ideas. sometimes mine make sense to me at the time, but as always there are always more ways to get to the end result. Thats why I like electronics so much. There are many many ways to get to the same end result. You cant say there is only one way to do that. there are many ways. Like WBahn said i could loose the 555 one shot and just use the PWM, a thermistor and a few control components. I will draw up the schematic and check it out.