With a couple of transistors, we can turn a MOSFET into a diode with almost no voltage drop, making it an ideal-diode.
This is a proof of concept for using a MOSFET instead of a diode in a buck converter. This is also called an ideal diode.
Table of Contents
- Device for testing
- Voltage bias
- Capturing the right moment
- Inverting the signal
- Safe zone
- Driving the bottom MOSFET
Device for testing
I made a very basic buck converter, so I have something to experiment on.
What you see here is the switching part.
The TO-247 device with the heat sink is the high-side MOSFET.
The other one was a diode and after cutting it off by its feet, I soldered a MOSFET onto them.
And I added a Dupont wire to the gate of it.
I made the controller on a separate board, which makes development a lot easier.
There are two linear voltage regulators and a high side MOSFET driver.
The device works, switching 3 watts of power to the light bulb.
That is well enough for this experiment.
Time for some probing.
I always have the yellow signal connected to the PWM output.
The blue one is the voltage over the diode.
This diode is the MOSFET’s body diode.
If I zoom in on the signal, we can see the voltage of the source actually go below ground.
That label there shows the ground voltage.
This is the moment we want to switch the MOSFET to ON, so the diode is bypassed. That would make it an ideal diode.
But for now, that is the job of the MOSFET’s body diode.
Capturing the right moment
I have made a circuit that is normally HIGH, but gets pulled down when the source is well below ground.
For some reason that also happens when the top MOSFET is on, but that is of no concern, as I will fix that later.
Let’s look at the circuit.
5 volts come in from the controller, through the resistor and connected to the collector of an NPN transistor.
That is also where the probe is at.
This construction is called a pull-up resistor.
The ground comes in through a limiting resistor and a diode.
And if the ground voltage is higher than the source voltage, the transistor conducts through the second diode and pulls the signal to ground.
Inverting the signal
With an additional piece of circuit, I now have the previous signal inverted.
If I overlap the two signals, we can see that we now have a high signal right after the high side MOSFET turns off.
Let’s take a look at the additional circuitry here.
5 volts from a digital IO pin goes to another pull-up resistor.
This goes right into the second NPN transistor’s collector.
When the previous signal goes high and comes in through a limiting resistor, the transistor conducts and pulls the signal low, making it an inverted signal.
Please note the additional circuitry that is going to drive the bottom MOSFET with this signal later on.
Now we need to know when the bottom MOSFET is allowed to switch on.
We start with the original PWM signal.
Create a second one with a slightly bigger duty-cycle.
Make sure they are in phase correct mode.
And invert the second signal.
Now we have a signal that is HIGH, only when the top MOSFET is off.
Driving the bottom MOSFET
At the bottom we have the new inverted PWM signal.
If we zoom in, we can see the delay between the two signals.
Don’t mind the noise in the signal.
With all the hoops and wires, this circuit picks up its own noise.
The 5 volts that went to the second pull-up resistor is now replaced with the second PWM signal.
It can now only be up when it is allowed to.
Outside that space, the pull-up resistor becomes a pull-down resistor.
I feed this final signal to a TC427CPA DIP8 IC, which is a low side MOSFET gate driver, very simple.
It turns the 5 volts into a 12 volts signal for the MOSFET, that’s all.
So, this is the signal at the gate of the bottom MOSFET.
When I increase the input voltage to the converter, a second signal appears at the end of the cycle.
I am happy to see the bottom MOSFET now doing the work of the diode.
Here is the circuit for the ideal diode we just created.