This is a DC-DC converter with 5 volts output.
Input voltage range is 0.95 volts to 4.8 volts.
With 3.6 volts in, output current is over 150 milliamps.
Efficiency at 3.6 volts in and 100 milliamps out is 84 percent.
No-load input current is 1 milliamp with 3.6 volts in.
It will start into a 10 milliamp load with 0.95 volts in.
The inductor is always an important part. For best efficiency choose something with low resistance and rated for 300 milliamps or more.
Almost any transistors will work, assuming the polarity is correct. Substituting 2N3906 for the PNPs and 2N4401 for the NPN might even improve it.
SwitcherCAD from www.linear.com actually wouldn't simulate accurately unless I used a 2N4401.
Just use your local generic transistors and it should work, or play around with R2 until it does.
But please, use some perf board, don't just build it in the air:
Q1 and Q2 form a Less-Than-Perfect™ oscillator with feedback provided by R1 and C1. Bias is provided by R2. On-time is fixed by R1 and C1.
Q1 charges inductor L1. Flyback is routed to the output through Schottky rectifier D1.
Voltage regulation is provided by multiplying Q3's junction with R3 and R5. This actually gives better performance than a zener diode, since low voltage zeners are terrible especially at low currents. Q3 throttles the circuit by extending the off time.
When Q1 switches off, Q3 is reverse biased and this disturbs the voltage reference as it recovers and can cause unwanted bursting - C4 neutralizes this effect.
Here's the schematic in SwitcherCAD format if you want to run the simulation.
This web page, created March 20, 2005, is temporary and will be rewritten and moved to a new location. Please do not link to it.
This page and contents copyright © 2005 Bob Blick