I posted about an open loop, non-synchronous, high side N-MOSFET diy Arduino based buck converter design earlier. But after researching, I came to know that it will not work. I need to rewrite that blog post. But before doing that I wanted to write a new one and record what was wrong in that buck converter circuit.
Here is the circuit design of the buck converter that was failing.
There are several design errors in this circuit. The main problem with this buck converter circuit design is that I used N-channel MOSFET IRFZ44N on the high side (between the power supply and the load). The reason this will fail or will generate heat is that, when the output PWM signal coming from Arduino is high, the transistor BD139 will turn ON, and the Gate of the N-channel MOSFET is pulled low to ground turning off the N-channel MOSFET. Now as soon as the n-channel MOSFET turns on, the source terminal of the MOSFET rises to 12V. But to keep the MOSFET ON the gate voltage requires gate to source voltage of 10V(for IRFZ44N the $V_{GS}$ >10V) which means the gate voltage must be 15V$V_{GS}$) if we want 5V supplied to the load since $V_{GS} = V_{Gate} - V_{Source}$, that is,$V_{Gate} = V_{Source} + V_{GS}$=$V_{Gate} = 5\text{V} + 10\text{V} = \mathbf{15\text{V}}$. But the circuit only has 12V available from the battery. When the Gate reaches 12V, and the Source tries to reach 5V, the $V_{GS}$ is only 7V. This is enough to turn the MOSFET partially on, but not fully.Because the MOSFET isn't fully "ON," it acts like a resistor. It will "eat" the missing voltage and turn it into heat. Similarly, and this is the second reasons why this circuit will fail is that when the PWM signal is low(0V) then the transistor BD139 is turned off. In this case, the 10kohm resistor pulls up the gate voltage to the +12V power supply and $V_{GS}$ becomes 0V because the source voltage is same as gate voltage. This implies that the N-channel MOSFET is turned off cutting off the path between the 12V power supply and the rest of the circuit. Though the power supply is completely cut off, current does not stop immediately. The current stored in the inductor L1 tries to flow in the same direction and to do this it reverses its polarity. So, the junction where the diode, inductor and the MOSFET meet becomes now negative. This in turn forward biases the Schottky diode 1N4148. Now the current flows in a loop, from inductor to capacitor/bulb to ground, to diode and back to the inductor. In this case the diode is stressed. The 1N4148 diode is now doing 100% of the work to keep the bulb lit using the energy stored in the inductor. And here is another mistake in the circuit. The 1N4148 diode is 200mA rated and so if the bulb wants, say 1A, the diode cannot keep up and will most likely burn out during this situation.
So, the fixes to this wrongly designed Arduino based buck converter design is not to use N-Channel MOSFET on the high side but use P-channel MOSFET like an IRF9540 MOSFET. Using p-channel MOSFET at the high side the current driver BD139 Q2 will correctly pull the gate to ground to turn it on.
Another fix is about the diode. The 1N4148 is small signal switching diode rated only 200mA. The buck converter inductor "kicks back" significant current through this diode every time the MOSFET switches off and therefore burn out. The solution to this is to use schotty diode that is at least $3\text{A}$ to $5\text{A}$, such as a 1N5822 or SR540. Schottky diodes are necessary because they switch fast enough for PWM.
Another modification in the circuit that should be done is about the high-frequency noise on the 12V input line. Switching circuit like buck converter circuit produces massive high-frequency noise. The remedy for this is to use a large electrolytic capacitor (e.g., $470\mu\text{F}$) in parallel with the battery, as close to the MOSFET as possible, to stabilize the input voltage.
The corrected circuit schematic along with result is in the next article, Arduino Buck Converter Design - Open-Loop, Non-Synchronous, High-Side P-MOSFET.
Related
