Single Coil Two Transistor Boost Circuits

[Farmhand]

OK Gotcha, When the parts arrive I will try the basic RC oscillator (which might be the better option) and the feedback setup a few ways and see what happens.

When the supply voltage gets very low the circuit should go into "burst" mode, where the circuit working draws down the battery voltage till the oscillations stop, then the battery voltage rebuilds and the oscillations happen in bursts that give the same amplitude output voltage of the continuous oscillations (or very close to it) so the bursting oscillator will still charge a capacitor to a higher voltage it'll just take longer.

I have a Galvanic cell with a JT powered by it that lights 2 x 5 mm LED's in bright bursts at about 2 to 12 Hz of 60 kHz oscillations, the supply voltage is less than 0.5 volts RMS by my DMM. It lights the LED's bright but they "blink" so I called it "Blinky Bill"  .

I intend to use that JT to flash at some plants powered by the "pot cell" while at the same time putting current through the soil for electro-horticulture experiments. That setup has been running for weeks now still with the initial wetting of the soil (no more water added yet) still moist, blinking is getting faster it seems. Likely better contact for conduction (less electrode to soil resistance).

Low DC resistance to inductance in a coil is just as important as low resistance switches ect.

..

P.S. Here's a block diagram of what I envision. I already have a less than properly efficient stage 1 - 2 and 3,... 4 can be done by simply adjusting the output of the VLV boost circuit and 5 can be done the same way but an appropriate capacitor dump/current surging charging circuit might be better suited to irregular input/output.

[Farmhand]

I mainly wanted this thread as a discussion for alternatives to the two winding coil Joule Thief system. So any circuit that can oscillate a single winding coil "like a Joule Thief" or better then it's relevant. 

I set up a simple air core Bifilar coil and connected it by one of the windings bottom to the ground and the top of the same winding to a capacitance insulated from the ground, then I connected a variable 40 to 1400 pF capacitor across the other winding to tune it in and then I connected one side of the "tank" to the ground as well and the voltage became much more and the local radio signal looks quite good. See scope shot below. I tuned it roughly to the local 840 kHz - A.M. station.

I would bet that I can use a crystal radio type circuit to rectify that and maybe double it so that a cap can be charged to 0.8 volts or so. 

If not it looks good for a simple crystal setup for one frequency.

Also below is a picture of the coil and arrangement when I took the scope shot, that was BEFORE I connected the tank coil to the receiver coil and doubled the voltage and even before using a top capacitance (going by the photo), so I think one diode voltage drop could still almost double the 1 volt PP. Maybe.

I better set it back up again and check it to the drawing before I post how I connected it, my desktop computer is not fixed yet so I can't use the video camera because no fire wire on the laptop. (I'm lazy like that)

[Vortex1]

I suggested the diode on the base of the pnp to protect it if you are using a large value C where appreciable current could take out the base junction.

If you are shooting for high frequency operation, the diode capacitance could be a problem, it all depends on what you are trying to design.

As I said, what I posted is the elementary circuit. It was for a variable intensity led flashlight.
I'm used to designing for high volume, lowest cost of components, and best overall performance.

Imagine you had to design a variable flashlight circuit that would sell in the millions of pieces for $1 , and cost of components had to stay under 25 cents.

Where cost is no object there are numerous options.

[Farmhand]

Hi Vortex, I did mention that you already mentioned about the diode at the PNP base, In a single line statement so you would see it.
I remember. Thanks.

For curiosity sake I ran the circuit down to 0.75 volts last night, but was unable to get it to charge a capacitor to over 3.8 volts at under 0.9 volts input, But three volts no problem.

I'm interested in building prototypes for collecting energy from as many sources for low power devices or for charging their batteries.

One use I just thought of is a light that flashes and can run from the 0.5 volts of a regular pot galvanic cell. I've made one already it is just a joule thief flashing 2 x 5 mm LED's 24/7. I sealed it up in a plastic box to make it waterproof with just the LED ends poking out like eye's, then I glued a light tube over the LED's to keep the light "directed" at the plant. So rather than just shorting out the Pot battery to get current through the soil, I can get light flashing on the plant as well as current through the soil.

Finding the best flashing frequency or frequency of the oscillations in the bursts which light the LED's might require some research.

Anyway, to run from a Pot cell the oscillator needs to have a very small current draw or be able to work in burst mode.

One situation where as much current as I can get is best.

I get current through the soil output and I get light output and I think the plants will benefit from the oscillator near it. Time will tell.

There are a multitude of uses for these type of circuits and extreme efficiency is not always required, especially if the input is free scavenged energy or energy otherwise not usable. Any circuit that works will work. Logically.

..

P.S. As Pirate said and the electro-negativity scale shows that a carbon magnesium battery works best for what we can get a hold of easily. Part of an old Mag wheel for the more negative element and a carbon arc gouging electrode or other form of carbon
for the less negative element. If no Carbon or magnesium then zinc/galvanized steel and copper.

..

[Farmhand]

The circuit with the PNP and NPN Bipolar transistors shows the below wave form at 0.76 volts input while working into a 25 Farad supercap. With the MPSA18 transistor a better PNP or a single logic gate (which is a bunch of mosfets working as one switch) and Shottky's  , I bet it could go much lower for a small expense in parts.

..