The 2nd week of feb brought me a good news and a bad news…
The Good: I had to make a small Solar Bot
The Bad: It had to be designed in just one week!
The Ugly: I didn’t have enough component samples…
So I went online and did what any self-respecting hobbyist would do, logon to solarbotics.net 
Since I really had a week to design, I didn’t count on spending time on optimizing my bot for a better performance. So I promised myself that I would improvise the bot later and went on to copy the circuits in the resources section. The circuits that you design for running the bot, are called ‘Solar Engines’.
The working principle of these engines is more or less the same: Charge a big capacitor through a solar panel and when you think the cap is charged enough, release all the energy through the motor in one go. Then again wait for the capacitor to be charged.
The Engines are broadly defined in to 3 types:
1> Time Controlled: The charge is boosted from the cap in periodic intervals. It doesn’t wait to check whether the cap is actually charged or not.
2> Charge Curve Controlled: The charge is boosted to the motor when the charge curve slows down. i.e we assume the when the curve becomes flat, the cap is charged and release the power.
3> The voltage controlled: when the voltage across the cap rises above a specified level, the power is given to the motor. These types of engines are the most common.
I, of course, chose to go with my fellow robotists and started working on the Voltage controlled Circuits.
The major components that you might need for this type of Solar Engines are:
A Voltage detecting IC:
a transistor:
a super capacitor (i.e. a capacitor with a big value)
and a solar panel:
Although I knew I didn’t have much time, I wanted to try all the voltage controlled circuits explained on different websites. Solarbotics.net seemed to satisfy me with most of my requirements.
So I tried the following circuits: (I wouldn’t want to steal credits deserved by solarbotics.net So instead of explaining them here, I’ll direct you to their link)
1>Zener diode based SE (Solar Engine): http://solarbotics.net/library/circuits/se_t1_zener.html
2>1381 based SE: http://solarbotics.net/library/circuits/se_t1_1381.html
3> Miller Engine SE: http://solarbotics.net/library/circuits/se_t1_mse.html
Whew!!… 3 days up building debugging and tweaking the circuits. Now I had to fix one engine and I chose the last one.. the Miller Engine. It seemed to bee a pretty efficient engine.
(BTW, I was building these circuits on a breadboard and testing them with a 200W bulb. The Miller Engine worked pretty well. Even after I turn the bulb off, the motor worked for around 20-30 seconds, which is pretty cool. But now I had to get it out on a field run…)
So then I built 1,2,3 PCBs as prototypes and the 3rd one worked just fine. The motor, now acting on load, did not run as much as it did when it was on a breadboard. But I was happy with it.
What I soon realized was that, by changing the contact area of the motor, I could make it run for longer,
i.e, if the motor is inclined such that more of its shaft is in contact with the ground, it runs faster, but since the load is more, it doesn’t run as long when the bulb source is removed. If I change this inclination such that lesser area is in contact with the ground, the motor runs for a longer time comparatively.
Have a look at the videos; the bot is kept in direct sunlight falling from the window of my room.
