Single Coil Two Transistor Boost Circuits

[Farmhand]

Yeah, again can you give any description of what we're seeing there ? I'm sure if I was using a transistor with a rating of 100 volts I could also light up an NE-2.  My intentions are to harvest low voltage energy sources that are unusable for much else and boost the voltage so that over time the energy can be stored and used from batteries or used intermittently like a flasher light.

The above circuit with less than ideal parts can run the battery down to 0.8 volts and it's still pulling 3.5 mA and boosting the voltage to 2.9 volts in a 25 Farad capacitor. When the cap is at about 3 volts ( can go to 5 volts ) it will then run the CMOS LED driver circuit for some time using the previously stored energy and the 4 x 5 mm LED's will be bright enough for a night light. I actually need a night light now because my goats are kidding and I need to get up during the night to check they're ok.

The circuit I posted I will be transforming, I hope by using an inverter gate in place of the pnp transistor and using MPSA18 NPN transistor for the main switch. Then by applying a high signal to the inverter gate input it's output should stay low and turn off the first stage according to a sensed voltage or such arrangement. CMOS is good because it has low power consumption. Gates can be paralleled to increase output power and they can drive transistors directly (with some help for some).

The CMOS LED driver circuit could be more efficient as well, and run from about 3 to 5 volts. It could also boost the 3 to 5 volts from the first circuit to up to 15 to 20 volts to dump into a 12 volt lead acid battery to rejuvenate/desulfate/charge a small one or similar.

I've ordered some 2N7000 mosfets and some Schottkys. As well I might look into a sub 3.3 volt micro controller, to control the initial and secondary circuits.

..

Some shots as the supply voltage drops.

..

[MarkE]

Is D1 helpful in that position?  With it in the circuit as is, Q2 and Q1 will only turn off very slowly.  I also recommend changing D2 to a 1N5817.

[Farmhand]

That works well Mark, Thanks very much and thanks for taking the time and effort to modify the drawing, much appreciated.
The circuit seems to be now taking about 200 nS to turn off if what I'm looking at is correct. Now is there a better way to adjust the conduction time of the transistor ? It's ok how it is, but there must be a better way. With your modification it now can go to a minimum off 20 uS and uses about 38 mA from 1.3 volts at 38 kHz, but it can be adjusted down to about 9.7 kHz and then it uses 111 mA with a 80 uS on time, Maybe 80 us is a bit long for that coil, not sure. Anyway it now adjusts a bit different I'll run it till the battery get down again and see how it goes.

Vortex did mention putting the diode there as well.

I might need to PM Stephan and get him to edit the first post for me edit time is over. I can't change the bad drawing.

Mark do you think that one of those 74AUC logic gates could be used in the place of the PNP 2N2907 ? And that way keep a neat and low part count "feedback oscillator".

Below is the wave forms now, the first is the two base wave forms NPN in yellow and PNP in blue.

Second is the collector of the NPN and the base of the PNP.

[MarkE]

You're welcome.  The 74AUC1G04 should do much better than the PN2907, particularly in being able to turn the MPSA18 off quickly.  The whole thing is a combination relaxation and blocking oscillator of sorts and depends a lot on the transistor betas and the saturation characteristics of the coil.  It is going to turn off at the earlier of C2 depleting, or L1 current building up to the point that Q1 lets go.  If that happens as a result of L1 saturation, then you've got a blocking oscillator.  If it happens because of C2/Q2/R1/Q1 then it's a relaxation oscillator.  Changing the value of C2 and R1 will change the timing in the former case.  I think that the POT R2 is kind of a dammit all.

If I were inclined to build a circuit like this to run from 1.2V, I would just pick up an NCP1400 with the output voltage I want.  Those oscillate at 200kHz, operate the inductor discontinuously, and once started with 0.9V or so will work down to about 0.3V.  It's much more of a challenge to make something work well using discrete parts, which I suppose is what's fun about these endeavors.

If you go with the logic gate, you might really want to try one of the low threshold voltage FETs that I have recommended, such as the DMG1012.  The gate charge is only about 1nC.  So if you switch around 50kHz that's only 50uA going into the drive.

As I have mentioned before, you can make further improvements with the choice of Schottky diode.  For leaded parts the 1N5817 is about as good as you will do for this kind of low power circuit.  In surface mount there are much better parts.

[Farmhand]

I want it to start from about 1.1 volts but work down to or at 0.6 to 0.7 volts or so, to drain old 1.5 volt alkaline cells. I could run an oscillator like in your drawing above from a plant pot cell even. The good thing about using a single winding coil is it can be a better coil.

Thanks.