LDO Voltage Regulator, 3 MOSFETs and a Zener

I love this circuit topology and I have done a lot of experimentation with it. It is an active voltage controller. I have incorporated it into many of my circuits. In fact, one is incorporated into a coil driver in the background of one of the photos for this Instructable. In a previous Instructable using an NPN and PNP pair, I use it to provide 6.25 V to an SN74HCT74 Flip-Flop running on an otherwise 12V circuit board.

In my work with high voltage coils and power coming off of those coils, I needed voltage regulators that could handle higher voltages than the NPN and PNP pair could handle. So, I designed a circuit that could use HV MOSFETs in the control circuit. I attempted to use an IGBT for Q1 as I thought it would be a good higher voltage substitute for a BJT. But since IGBTs are current driven and not necessarily voltage driven, the circuit worked in a less than satisfactory way. It did work, but the regulated voltage and the Zener diodes were far apart in their values.

For this topology, one P function is needed in the logic of how the circuit operates. In the origin circuit, it was in the power transistor, a PNP power BJT. In my circuit linked above I moved that function into Q2, the 2nd of the two control BJTs. That way, you can use the cheaper, faster, and higher current carrying N-Channel MOSFETs as the power transistor, or transistors, because you can run a bunch of them in parallel. In the circuit of this Instructable the P function is in Q2, a P-Channel MOSFET.

Analysis of the circuit shows that when the voltage is below the breakdown voltage of the Zener diode, the voltage at the gate of Q1 is low and the transistor is OFF. The gate of P-Channel MOSFET Q2 is at the same voltage as the incoming unregulated supply voltage and is also OFF. When Q2 is OFF, the gate of the power MOSFET Q3 is high and Q3 is ON allowing voltage to go to the source side of Q3, the regulated voltage side of the circuit. When the regulated side of the circuit reaches the breakdown voltage of the Zener diode, the diode allows current and voltage through, raising the voltage on the gate of Q1. Q1 is switched ON, dropping the voltage on the bottom side of the 100 K, 2 W resistor. This in turn drops the voltage on the gate of Q2 switching it ON. When Q2 is ON, it drops the voltage on the Q3 power MOSFET switching it OFF, thus not allowing any more voltage or current into the regulated side of the circuit. The capacitor, or capacitors, smooth out the ripple of the oscillations of the circuit and maintain a smooth and steady voltage on the regulated side of the circuit.

In this circuit, you should use MOSFETs that can handle greater-than-or-equal to the unregulated voltage coming into the circuit. Also, the capacitor(s) you use, should also be able to handle the unregulated voltage, even though it or they are on the voltage-regulated side of the circuit.

1. Smaller N-Channel MOSFET, TO-220 package. I used a P60NF06 because it was the the only smaller N-Channel MOSFET that I had on hand. Vds = 60V. It is not ideal, but, it worked suitably for this project. In the future I will use an IRFB38N20D (Vds = 200V), when they arrive from eBay.
2. Smaller P-Channel MOSFET, TO-220 package. I used an IRF9540 (Vds = -100 V). I have a bunch of these.
3. 100 K, 2W resistor
4. 100 K, 1/2 W resistor
5. 10 K, 1/2 W resistor
6. Zener diodes, 1 W work satisfactorily.
7. 47 to 100 uF electrolytic capacitor, 100 V. You can use multiple capacitors in series or parallel to meet these requirements. If you are not sure how this works, please do a little reading about how to do this.
8. Breadboard
9. Higher voltage circuit board to hold your power N-Channel MOSFET or parallel MOSFETs
10. Standard DMM
11. Oscilloscope (optional, but fun)
12. Some kind of voltage supply. I have 12 V, 60 V, and 100 V supplies on hand.

Assemble the circuit as diagrammed. A photo of mine is shown here. In the background is another one on a more complex circuit board where you can see the voltage regulator right above the "Freq" in the label. There is an NPN right next to a PNP BJT (the PNP has a strip of masking tape on top of it).

Because the components on the control board, the breadboard, consume little power, they only need a small ground wire connected to the negative rail. The power coming into the control board powers everything so no separate power connection is necessary.

Note: What might look like a single + rail in the circuit diagram is not one rail. The left side is the unregulated voltage. The right side is the regulated voltage and these two are not connected together. The left side is the Drain of the power MOSFET, and the right side is the Source of the power MOSFET.

You will need to try different Zener diodes. Mine are not that high of precision (apparently). They are in the 1N47xxA series. I have put in 5 W Zener diodes and the circuit works exactly the same. Zener diodes with the same voltage rating will produce slightly different output results. If you don't get exactly the output you would like, try another diode with the same, or different, voltage rating. I have included a pdf of a spreadsheet of my results with the same and different diodes with 2 different input voltages.

I connected my oscilloscope to the gate of the power MOSFET. At low frequency it looks like a bunch of noise. When I set the scale to higher frequency, I can see blobs of voltage that are separated by about 150 kHz. When the scale is set to even higher frequency, I can see individual cycles that are at about 10 MHz. I don't know if this is simply ringing of the MOSFETs, or the entire circuit is oscillating with this waveform. The gates of all 3 MOSFETs have the same signal, so I am leaning toward the entire circuit oscillating at 10 MHz.

I included a photo of the DMM reading which remains steady at 12.06 to 12.07 V for hours. I have a 560 Ohm resistor between the + and - rails as a load (It is only there for testing. Remove for standard operations). Also included is a 10 second video of the waveform, which I find very interesting. It doesn't change with changing the values of the capacitors. I recommend giving it a view. If nothing else, it's entertaining.