This is my "competition" solar roller. The long post is what differentiates my competition rollers vs my "sunlight rollers." It's to move the panel as close to the lights over the track as possible but still stay inside the size limitations. In sunlight, this is currently my second fastest roller with "Zip" (web page coming soon) doing the meter in about 4.5 seconds.
My model airplane hobby has supplied me with a variety of hi-tech building materials. For this, I used some carbon fibre tubes I bought from http://www.homefly.com for use in ultra-light indoor R/C aircraft. Ready to run it comes to 13.48 grams. This roller will do the meter in under 10 seconds under an incadescent bulb I use to simulate sunlight until the day it actually appears over Calgary. It's a little over the top I admit but I had a lot of fun building it and that's what it's all about.
Circuitry-wise it uses the good 'ol Miller engine. No sense
arguing with success.
Besides the low weight this roller has two aspects optimized to make it go fast. The first is the storage capacitor. I use one of the awesome aerogel .1 F 5.0 V capacitors. Grant from Solarbotics told me about them. Their low ESR (Effective Series Resistance) was a huge improvement over the pair of .047 F Tokin capacitors I used before. When I switched the time to traverse the meter once the bot triggered dropped from 1.8 seconds to around 1.2 seconds. I chose the .1 F capacitor because I wanted the bot to do the distance in one burst. I use my automated solar roller race track to test and time my rollers.
The next big trick on this bot is the solar panel used to power it.
I found these cells a few years ago at All Electronics Corp I bought a couple of hundred but if I knew how awesome they were I would have bought even more. I estimate their efficiency at about 14%. For the same surface area and voltage as a Panasonic Sunceram I found that a panel made with these cells produced about 80 to 100% percent more current. For example, one of the 20mm x 20mm cells give the normal .5 voltes (open circuit) but 120 mA (short circuit)! Their only fault is their voltage drops a lot when the light dims.
The cell on the lower right is one of the cells I have trimmed down. It is a for a 5-cell roller using a modified Miller engine based on an N-channel low threshold MOSFET.
The cells I bought only came in two sizes: 10 x 20mm and 20 x 20mm. I worked quite hard to develop a solar engine that will charge, trigger and reset on only two volts (4 of the 10x20 cells). The jury is still out on that so, in the meantime, I cut the cells down so I could make a 6 cell panel. Here is an example of a printed circuit board I made for an 8 cell panel.
To make the panel I made up a simple circuit board with square pads and little tabs to solder to. The bottom of the cell is one terminal and the other is the little strip along one side. I found that it was nearly impossible to solder to the bottom terminal as the conductor would delaminate very easily. I found the solution in conductive epoxy. I found it worked very well to use the epoxy to glue the cell to the pads on the PCB. The other terminal got a little 30 gauge wire soldered to it and a terminal going to the bottom pad of the next cell.
So far, the only problem I have been getting are cracks in the corners of the cells. The lower left cell on this panel has a crack. It slowed my roller down signficantly and the entire panel needs to be replaced. There's a tiny gap between the cell and the PCB substrate. Any pressure could cause the cell to crack. The next panel I make will have little strips of tape around the PCB pads to support the cell. They are incredibly delicate and break very easily.
Cutting down the cells required a lot of experimentation. The cells fracture instantly at the slightest pressure and all of my previous attempts to trim the cells resulted in a shower of tiny cell fragments. I found the best method was to gently clamp the cell in a soft plastic clamp, score the cell with the edge of large carbide drill bit, carefully snap the cell off and finish the trimming by running the cell against a wetstone. Sandpaper and files don't work. The files fracture the cells and sandpaper erodes instantly requiring constant changing. A wetstone works perfectly.
Update Oct 1, 2004
Um, yeah, the cracked solar cell
slowed it down. I made a new one last night and compared short
circuit current against the cracked one. Old one generated 5 mA while
the new one 25 mA under a lightbulb that I use to try and simulate
sunlight (when we get sun here). I hadn't expected it would result in
such a large difference.
Last modified: Oct 1, 2004. ATTACK OF THE FLUFFY MONSTER!! IEEEE!!!! No, really SHE JUST BIT ONE OF MY AIRPLANES AND PUT A LITTLE HOLE IN THE FIN! GGRRAARRRGGGHGHHHH