Single Coil Two Transistor Boost Circuits

[Farmhand]

Mark, I didn't notice the "bootstrap" drawing you made at first but seen it the next day or maybe two later. With 1.4 volts max input voltage it should be ok. I'll test the lowest voltage the circuit will run to with the mosfet for the switch and no bootstrapping first, then I'll try it with the bootstrapping. I notice with the bjt when the input voltage goes below 0.7 volts the collector voltage begins to rise before tun off. And the output becomes very small. Maybe the mosfet will stay on better.

I think with the bootstrapping the circuit would run to produce a constant output from a single galvanic cell, enough to boost voltage into a 50 Farad cap or small rechargeable battery, which can then be used to dive several LED's at good brightness for some time with a non bootstrapped circuit. Many ways to utilize over 1.5 volts.

The regular CD4000 series chip will work from three volts ( a bit less ).

Thanks for the help.

[Farmhand]

The dirty trick that you might wish to try is to use a pair of Schottky diodes and a 0.1uF capacitor to form a bootstrap supply for your logic gate.  The Schottky cathodes would be common to your logic gate Vcc.  One anode goes to your battery, the other your output.  The capacitor goes from the cathodes / IC Vcc to ground.  You will have to be careful about exceeding the IC maximum voltage.  An LM4040 has a pretty small minimum current and would work to regulate the output.

Would it be possible to use one of those LM404's to turn off the oscillator when the desired voltage is reached ? Connect the cathode to the output and the anode to a logic gate input ? Maybe with an additional resistor or two.

..

[MarkE]

Would it be possible to use one of those LM404's to turn off the oscillator when the desired voltage is reached ? Connect the cathode to the output and the anode to a logic gate input ? Maybe with an additional resistor or two.

..

Yes, you can use it to kill the oscillations instead of just loading the source.

[Farmhand]

OK, I think this PCB design will work for a basic fixed frequency and fixed pulse width unit, I got confused a bit because the SMD
parts I'm using on the solderless board for prototyping are upside down and when I use both SMD and through hole components
I'll need to either solder the through hole parts up off the board, or put the SMD parts on the copper side and the through hole parts
on the other side. Should be ok, I think I'll do that to keep all the parts right down on the board. I could solder some through hole
parts like SMD parts such as the resistors, but I'll see how it looks.

The board should measure 45 mm x 22 mm. The connectors won't need to jut out, and I might modify the PCB design to include
2 x 5 mm LED's permanently between the diode output and the positive rail so there is always some load. Then when the output
between the circuit ground and diode output is loaded the LED's will go out if the load takes enough power to drop the output to
below the LED conduction voltage. I know loading the ground to diode output down to 4 volts puts out the LED's.

Just a basic design to test efficiency. I still need to nail down the exact timing capacitance and resistors, looks like 200 pF will be
enough timing capacitance so I'll work it out from there. On the solderless board there is loads of parasitic capacitance so the PCB
device will operate a bit different.

Anyone see a problem with the traces ? Logic gates at bottom, timing resistors above, timing cap to the right of those, inductor to
the right of that, and mosfet just below with the output diode near the output capacitor.

The board needs one jumper from the timing capacitor to the timing resistors and logic gates junction point.

..

P.S. I can add the loading LED's by simply soldering them to the spare bit of copper near the output diode and the positive rail,
I can drill holes if I want them solid mounted on the PCB or I can use wire and mount them in the case shining out.

And the circuit is also attached. With the optional LED's shown between the output and + rail.  I think it's all drawn correctly.
The circuit is working fine but I cannot rule out drawing mistakes. so I ask for checkers, please. Ileft out the timing resistor values
because they will depend on the inductor and the users wants.  The timing cap can be whatever is applicable to a persons needs
as well. I'll use a 1 mH inductor so I'll want about 50 to 80 kHz with about 30 to 40% duty, or probably about 5 uS "on" and about
7 or 8 uS "off". about.

..

[TinselKoala]

Looks pretty good to me so far.


Meanwhile I'm fooling around with Arduino. Here is a basic "barebones" pulser that gets to 50 kHz with a reasonable rectangular output pulse.



/*-------------------------------------------------------/     
   TK's Barebones pulser

  Enter # of microseconds for pulse to be HI in "dwell"
  Enter # of microseconds for pulse to be LO in "gap"
  +5 V pulse comes out Digital Pin 6 (or whichever you like)
  Check frequency and duty cycle on oscilloscope
 
/--------------------------------------------------------*/

#define OutPin 6

int dwell = 5, gap = 5; // Gives about 50 kHz, 50 percent HI

void setup() {
  pinMode(OutPin, OUTPUT);
}

void loop() {
  digitalWrite(OutPin,HIGH);
  delayMicroseconds(dwell);
  digitalWrite(OutPin,LOW);
  delayMicroseconds(gap);
}