I'm using the following circuit to regulate voltage running to a fan using PWM as the control method. The PNP is just a placeholder as Multisim doesn't have the transistor I'm actually using.

What I would like to do is prevent the output voltage to the fan (represented by the 6 Ohm resistor) from dropping below say 5v). So effictively I want something to check the output voltage and if it's below the stated minimum it just makes the output 5v, until such a time as the output voltage rises and it allows it to rise.

Is this possible without discrete simple components? I was thinking something like a Comparator to activate a secondary circuit to take over powering of the fan when the output voltage sags with low duty cycle but am unsure where tio begin with that. I'm not even sure what this is called otherwise I'd look into examples of it.

Any ideas? Or some clues to help me find a solution.

Thanks

 

Why would the voltage sag?

 

As I reduce the duty cycle the output voltage drops more or less in line with it. What I want to prevent is stalling of the fan at too low a voltage when the duty cycle drops to below 20 or 30% when the capacitor can no longer hold enough energy to supply the fan load.

It's not really a deal breaker as if set up correctly the circuit works by varying the 1k resistor between mosfet and PNP so that it won't drop below 5v but I'd like it to be more fail safe and idiot proof.

 

Is R4 pretending to be a PWM circuit?
Need more information to find a slick way to do this.
What's your maximum current at 12 V, at 5 V?

 

XFG1 is the 5v@25Khz PWM generator here. In many PWM circuits there's usually a resistor between the PWM output and the rest of the circuit so I put 1k there. I use a frequency generator as it's easier to adjust and construct than a 555 PWM circuit. Sorry if I wasn't clear.

As for maximum current at 12v I'm able to drive around 0.5A with the 1k at R3 rising to 1A with a 250 Ohm instead.

So I'd say maximum would be 1A at 12v but would drop to around 0.3A at 5v.

Thanks.

 

No, I read it wrong. My bad.
At least you revealed that you intend to use a 555 chip.
Mumbling to myself...
The 2SA1213 transistor has a limit of 2 amps and is designed for a gain of 20 for being a switch.
The BS170 will be about 1.2 ohms at 8 volts gate drive.
I'm thinking in terms like this...you provide the resistors.
http://www.mouser.com/ProductDetail...GAEpiMZZMstCHp3EWKGl7RSz/WB%2byfUiAYcVnTa5PE=

 

Oh no don't get me wrong. I do use a 555 PWM circuit to do testing of the circuit but ultimately this circuit is to run off the CPU PWM header on a PC. They are 22-28kHz and around 3-5v PWM. The function generator is usedin Multisim only

The circuit works just fine and I've got it running multiple fans off the PC PWM duty cycle, it also provides a tach signal to the motherboard.

 

You would need to use an RC filter to give the average fan voltage, and apply that to a comparator with a 5V reference. The comparator output would then go to a CPU input, telling it to maintain the PWM duty-cycle at that minimum level.

 

Thanks.

RC filter and comparator I can do but talking back to the motherboard is going to be too much for this little project.

Was just hoping there'd be a way to throw some other source of energy at the output when required.

 

Thanks.

RC filter and comparator I can do but talking back to the motherboard is going to be too much for this little project.

Was just hoping there'd be a way to throw some other source of energy at the output when required.

Well then you could use the comparator to switch out the PWM signal from the PC when it gets below 5V average and switch in a 5V source, such as from the PC. A couple of MOSFETs would work for the switches.

 

Could you put a little hall sensor or optical sensor on the fan to act as a pulse generator to shut the fan off or increase the pulse width in a stall situation.
It is well beyond my capabilities , but is there any possibility of extracting the back EMF from the motor and filtering it from the pwm pulses? It would be a really neat solution.

 

It is well beyond my capabilities , but is there any possibility of extracting the back EMF from the motor and filtering it from the pwm pulses? It would be a really neat solution.

Here it is. http://ww1.microchip.com/downloads/en/devicedoc/41233a.pdf You will find it under tip#7
I would love to try this sometime.

 

Thanks but I think that's way beyond my capabilities too and I wanted to avoid microcontrollers as I have no knowledge of them and no funds to buy all the stuff needed.

Was just hoping there'd be something simple enough for me to do myself using junk box components but even the comparator/mosfet switching concept above puzzles me.