The initial idea was to build this LED flashlight:

• http://www.popsci.com/flashlight
• http://nutsvolts.texterity.com/nutsvolts/200802/?pg=34
• http://nutsvolts.texterity.com/nutsvolts/200802/?pg=35

We were succesful at putting the nuts and volts kit together. The next step was to alter the design for my Analog Design course. We attempted to program the chip but were unsucessful. We then decided to imitate the Attiny 11L chip in the circuit ( we did the one with 5 LEDs), but couldn't not figure out what the chip was doing from its outputs. We now just want to see if the chip is what is makin the circuit more efficient so in other words is it a factor in the 88% efficiency.

The professor said we should try building a circuit that powers 5 LEDS and see if it is as efficient, as far as power, as circuit with the chip.

Any circuit ideas or suggestions?

 

EDN published a Design Idea a month or so ago, using an extremely similar circuit; except instead of an ATtiny, they used a CMOS 555 timer, and a larger inductor. The 1N5819 Schottky diode and 1uF cap were omitted.

The 555 timer circuit will provide reasonably comparable functionality for the purposes of your experiment without the complexity of having to program a uC and trying to figure out what the heck it's doing (or not doing).

The EDN Design Idea runs from a single 1.25v NiMH cell for up to 192 hours. The circuit could certainly be modified to run from 4.5v. R1 and R2 should be increased for the higher voltage. Running from 4.5v would enable you to use the 2N7000 MOSFET instead of a 2N4401 NPN transistor.

Don't try to use a standard bipolar 555 timer; you need a CMOS version.
ICL7555, TLC555, etc.

Here's a link to the article; it starts at the bottom of page 58
http://www.edn.com/contents/images/6598371.pdf

 

I thought a CMOS 555 was rated for 2.0VDC? They are showing 1.25V.