I'm getting a bit ahead of myself - this test was without a MOSFET...

So far my idea for a mutually reliant (symbiotic - yet not living...) charger and torch system relies on boosting the DC 5V supply from the USB wall charger to my TP4056. I could have paid a small amount of money for a boost converter module from eBay but as I'm trying to learn more analogue electronics, I had a look what could be made from discrete components and found the wonderfully named Dickson Doubler. Here's my LT Spice schematic of this circuit:

SFM = Shrinkify for Manufacture

47uF electrolytic capacitors are readily available but are relatively large for the PCB layout I'm aiming for (small) so I wondered if I could shrink the size of the capacitor by reducing its capacity into the range where SMD ceramic capacitors would be affordable. The LT Spice simulation showed that a 1uF capacitor would actually produce a broadly similar, if not slightly higher boost voltage. I think the trade-off would be in the pulse smoothing but we'll get onto that.

Homemade Dickson Doubler

I bought the four passive components required (I get a free clock source from my ATMEGA328P) and popped them onto the breadboard:

Relative to the schematic, D1 and C1 are on the left, D2 and C2 are on the right. The blue rail on the left is actually the clock line, not ground (which is out of shot to the right). After connecting the oscilloscope as shown and providing a 500Hz 50% duty clock pulse, here's the surprising plot:

Note that we're on 5V per vertical division here - 10V! This would be perfect for the MOSFET Gate drive. What about the pulse frequency? I got pretty much the same result at 50Hz.

Add Amps?

What about if the Gate takes some current (which it shouldn't do - we're switching it on an leaving it on to charge the battery)? I stuck a white LED and a handy 6.7k Ohm resistor across the legs of C2 and this is what I got for a 50Hz clock pulse:Oh dear - we're back at 2V per division and we're getting 7V but it's rapidly dropping to 5V. What about back at 500Hz?

That's more like it! Back up near 10V.

This latter exercise is all pretty pointless because we're not switching the charging MOSFET on and off very often and Gate current is a function of it's capacitance and the switching rate (see this application note).

Summary

It's nice to see what the limits of this circuit are as a voltage doubler. I'm quite impressed as the LED was drawing 0.75mA and we were only using a pair of capacitors, a pair of diodes and a square wave.

Next job will be to try creating the two halves of the mutually reliant charging circuit and sticking my tongue test equipment across the terminals...