Time to sleep. Finally. After a week of abusing my poor little body by keeping it artificially awake with a combination of coffee and energy drinks, the departure of Ong to Japan marks the beginning of a new era of leisure.
This second project (codename "Narcissus") was made possible thanks to the discovery of a nice-looking bottle from an energy drink called "Hang" which comes in three colors: orange, black and magenta. Orange tastes like shit and magenta (my favourite) well tastes pretty much the same as it looks like. But the black one is the first one I tried here and it has this strange taste of mint, like some kind of potion. I quickly became hung on Hang, specially because I like the look of the bottle and I had this idea of a very big sculpture involving many Hang bottles hanging from the trees. But given the difficulties I had with the First Absence project (codename "condom") I decided to work on something more simple. Anyway, I also drank a lot of Redbull, of course, and Lipovitan, and another one called "M-150" (no, it's not a lubricant brand). And, believe it or not, I even tried two bottles of "Beauti Shot", one of which contains collagen (but mainly consists of sugar, colorant and a lot of imagination).
But, coming back to the actual look of the bottle, the black Hang bottle won my heart. I thus passed the last week-end putting the circuit together. Very simple, it consists of a garden light solar panel, three 1.2V NiMh batteries, an ATmega168, a light sensor and an LED, together with a few other electronic components (see the schematics).
I used the solar panel voltage as a trigger for sleep mode during daytime. This avoided using the photoresistor for both analog light detection and digital sleep mode control (which would have forced to choose a unique voltage division factor for these two different tasks). The photoresistor being close to the LED, using it as the night/day detector would have meant getting biased results. The only issue I have with this configuration is that the trigger voltage changes according to Vcc, but this is a negligible issue. The operating voltage under batteries is 1.8V (the brown-out level of the ATmega168) to 4.0V (the voltage when the batteries are more than full). Therefore, the point of trigger is between 0.9V and 2.0V, which is way ok since the solar panel generates around 4-5V when placed under the sun.
I was a bit worried about the length of the wires, about 3 meters between the solar panel and the batteries and another 3 meter between batteries and circuit. But I tried the setup and there was no noticeable drop in voltage.
Finally, concerning the photoresistor setup, my experiments have revealed that it's better to use a small resistor between the GND and the analog input. In this circuit, I used a 10k resistor together with a 10k-50k photoresistor in the voltage divisor, which should bring an input of about Vcc/2 under the light and around Vcc/5 in the dark. When put in the dark and started, the blinking LED brings the average input voltage up but the small resistor prevents it to get too close to Vcc, ensuring a better resolution, as could be seen on the oscilloscope screenshots.
Enough for today. Tomorrow, the second part of the story, involving a boat, an accident and a night in the jungle.Time to sleep. Finally. After a week of abusing my poor little body by keeping it artificially awake with a combination of coffee and energy drinks, the departure of Ong to Japan marks the beginning of a new era of leisure.