LR7843 has Very Low RDS(on) at 4.5V VGS
i already did that [CU PWM SIG] - [FET's gate ]
but FET get's too hot over 40 ℃
it's too hard for the MCU's I/O pin to onnect FET's gate directly
that is the reason why i'm doing this.
i think i have to fully open FET's gate.

What is the stall current of the fan motor?
While Rds,on is important, current capacity is related to the Rds,on. If you're driving something that takes 2-3W at 24V using a FET capable of 100A+ is extreme overkill - and wasting space and cost to manufacture - not to mention your time. I suspect your FET is way larger than it needs to be. A microcontroller pin can drive a small logic level FET for a computer fan. I know, I've done it, several times. However, it must be designed and not just thrown together by reverse engineering someone else's design.

Alternatively, use a gate driver if you want to drive a nail with a sledgehammer.

 

This is the method I use, using a MIC4427 gate driver
https://forum.allaboutcircuits.com/...nal-with-microcontroller.185813/#post-1719933

 

Alternatively, use a gate driver if you want to drive a nail with a sledgehammer.

Explain (the gate drivers come with the variety of peak sink/source capabilities = you can pick the suitable size for your sledgehammer)

 

What is the stall current of the fan motor?
While Rds,on is important, current capacity is related to the Rds,on. If you're driving something that takes 2-3W at 24V using a FET capable of 100A+ is extreme overkill - and wasting space and cost to manufacture - not to mention your time. I suspect your FET is way larger than it needs to be. A microcontroller pin can drive a small logic level FET for a computer fan. I know, I've done it, several times. However, it must be designed and not just thrown together by reverse engineering someone else's design.

Alternatively, use a gate driver if you want to drive a nail with a sledgehammer.

Suggesting to use an adequate driver for the device being used is certainly both suggesting a proper design choice and a reasonable design approach. And there are specific times when driving a nail with a sledgehammer is the very best choice.
as in most designs, selecting elements that work well together is more likely to lead to better results.

 

This is the method I use, using a MIC4427 gate driver
https://forum.allaboutcircuits.com/...nal-with-microcontroller.185813/#post-1719933

Referring to my old IR Power Mosfet Designers Guide from the 90's, They typically show The advantage of totem pole drive for a Mosfet gate using discrete transistors.
Now we have the advantage of the likes of the MIC drivers, Both inverting and Non.

 

MOSFET driver ICs are cheap - as cheap as logic gates. I don't like the idea of the load being coupled directly to the microprocessor by Cgd.

 

Also when using a Micro, eliminates the need for a logic gate Mosfet.

 

this is my circuit
(15Khz PWM)


R2 -> 330
MOSFET temp : 40 ℃




R1 -> 100
MOSFET temp : 31 ℃

R1 -> 100 is sutable to me!

i'll try MIC4427 but it will takes time, i have to buy that
thank you all!

 

With the two capacitors across the motor the diode is not needed, nor does it serve a useful purpose. In addition, a 0.1MFD capacitor scross the motor will be entirely adequate for spike and noise reduction.
The concept of PWM is not to smooth the pulses into a DC level, except in the D/A converter schemes, which driving a motor is not.

 

Charging a 220uF capacitor abruptly 20000 times a second isn’t going to do the MOSFET or the capacitor any good!

 

With the two capacitors across the motor the diode is not needed, nor does it serve a useful purpose. In addition, a 0.1MFD capacitor scross the motor will be entirely adequate for spike and noise reduction.
The concept of PWM is not to smooth the pulses into a DC level, except in the D/A converter schemes, which driving a motor is not.

Charging a 220uF capacitor abruptly 20000 times a second isn’t going to do the MOSFET or the capacitor any good!

yes, i removed all capacitors

thank you!

 

you should protect the mosfet from the inductive kick - not the motor . . . i doubt the SS54 can handle 2.5A for long // re-checking the validity what i just wrote . . .

. . . while the Schottky likely protects the MOS-FET - the SS54 without caps needs to be confirmed to be a proper choice

. . . the SS54 seems to be quite capable - except that you should check the instantaneous discharge current of the ("shorted" by the SS54) motor coils

although the forward RMS through the Schottky is likely less than \({\large \frac{2.1A}2}\) - it will be heated some extra with the alternating bias on it

 

With the two capacitors across the motor the diode is not needed, nor does it serve a useful purpose.

The opposite is true.
The capacitors are bad, since they generate large peak current through the MOSFET when it turns on.
The diode is needed to carry the motor inductive current and prevent spiking when the MOSFET is turned off.

 

it will be heated some extra with the alternating bias on it

How does that add extra heat?
The turn-off recovery current in a Schottky is negligible.
It's the forward current through the diode drop that causes heat.

 

How does that add extra heat?
The turn-off recovery current in a Schottky is negligible.
It's the forward current through the diode drop that causes heat.

yes i changed my diode 60V/6A




this is my new circuit




mosfet doesn't work at all (just turn on, can't off / can't controlling)

 

I think this has been said before: the resistances in the gate drive circuit are too high.
Your gate capacitance/gate resistor time constant is 20us, which is 40% of the period. It’s not going to work well unless it is <3%.
(Takes about 3 time constants to charge/discharge the capacitance, and the switching time should be <10% of the period for a respectable degree of efficiency)
You are also driving it commin-drain with a N-channel FET. You need to get the gate voltage at least 5V higher than the positive supply to get an adequately low voltage drop across the MOSFET. Why are you not driving it common-source?

 

this is what i found and i changed Resistors value







but this circuit is working...

 

but this circuit is working...

That’s a matter of opinion.
Judging by the waveform, I would disagree.

 

I think this has been said before: the resistances in the gate drive circuit are too high.
Your gate capacitance/gate resistor time constant is 20us, which is 40% of the period. It’s not going to work well unless it is <3%.
(Takes about 3 time constants to charge/discharge the capacitance, and the switching time should be <10% of the period for a respectable degree of efficiency)
You are also driving it commin-drain with a N-channel FET. You need to get the gate voltage at least 5V higher than the positive supply to get an adequately low voltage drop across the MOSFET. Why are you not driving it common-source?

i'm sorry but i cant' understand.

i changed R2, R3 (680 / 1K / 2.2K / 4.7K / 10K)
and moved R3.

but it didn't works

[Edit / Result]
I removed all BJT and directly connected (MCU's PWM pin - MOSFET's Gate (with 47R))
obviously it works good.

but i still wonde.If I keep using it like this, I wonder if it will affect the life of the MCU.

 

Try using the post #37 circuit, but with R1=1k, R2=10k and a 10 Ohm gate resistor.