....or we could continue were as well, please tell me which you prefer?

Let us continue this discussion in chat.

Hmmm, I added the VCC and GND connections to the LM2903, there is no change to the behaviour, I have to go out now, but will continue on my return.

Do you think the circuit is wired correctly now (prior to the mods)?

Right I will do that, in the interim, I tested the circuit on a breadboard again, and got the following result: Output voltage on the F connection follows the input voltage on the D connection, until the input voltage reaches 3.0v then the F voltage falls off to zero.

The final stage hacktastical (I hope), is a quick question about the LM2903. In the circuit diagram it shows +, _ and i/o pins, but doesn't mention the Vcc and Gnd pins. Are these pins meant to be left unconnected, or are they connected to the positive and negative rails?

You're right of course, I was ignoring the circle. I took a proper look and it was obvious.

So, correct me if I'm wrong, when looking at en.wikipedia.org/wiki/MOSFET under the heading "Circuit symbols", the Mosfets are drawn, as in the fourth type pictured, looking from left to right. There is no physical wire connecting 2 legs together.

This might seem like a silly question, but your circuit shows 2 of the pins on the Mosfets connected together. I suspect that this is just how they're drawn, rather than how they're meant to be connected, and is the probable reason why my circuit isn't working. Looking at the Mosfet from the front, the pins are Gate, Drain and Source, right? If that is true I have Gate and Drain shorted on one Mosfet, and Drain and Source shorted on the other.

I built your circuit using the Mosfets I linked to, on a breadboard. I tried to test it using a variable DC power supply. The way I tested it was by putting a voltage of 3v on the positive rail and gradually increased it to 12V. I thought that a DMM connected to the F connection would see maximun Volts when the supply voltage was 3V, decreasing to 0V when it got to around 5.4V. I measured the F connection between the point marked F on the circuit and the same ground as the power supply.

The other is a SQP60N06-15_GE3-ND its datasheet is here: vishay.com/docs/63554/sqp60n06-15.pdf Do you mind checking that they are suitable? Thanks Dave

One of the MosFets I am using hacktastical is an IRF9540N P-Channel Mosfet. It has the following specs: Drain Current, VGS @ -10V -23 ID @ TC = 100°C Continuous Drain Current, VGS @ -10V -16 A IDM Pulsed Drain Current -76 PD @TC = 25°C Power Dissipation 140 W Linear Derating Factor 0.91 W/°C VGS Gate-to-Source Voltage ± 20 V EAS Single Pulse Avalanche Energy 430 mJ IAR Avalanche Current -11 A EAR Repetitive Avalanche Energy 14 mJ dv/dt Peak Diode Recovery dv/dt -5.0 V/ns TJ Operating Junction and -55 to + 175

Hi hacktastical, I have started to build the circuit that you kindly provided, but I do have a question: Does the type of Mosfet matter? I am assuming that as long as it can handle the output current of the generator and that it is a P type, or an N type as shown in the circuit diagram, all will be ok? Is too high a current rating for the Mosfet an issue? I have an N-Channel Mopsfet rated at 60V 56A already, but I'll have to get a P-Channel one.

Just to let you know, I ordered the parts today from Digikey. I should get one built in a week or 10 days. I'll let you know how it goes.

Thank you very much for creating this drawing Hacktastical. If I build this, my problems should be solved. (as long as it works :))