I added a 1 Farad capacitor to the Xee2 flasher circuit I blogged. I left the 10000 microfarad capacitor connected, but this is only 1 percent of the 1 Farad capacitor so it really doesn’t make any difference. I used my LED task light to illuminate the 4 LEDs and charge the capacitors. I connected it up in the morning, and left it running for several hours. I came back a few hours later and the LED was flashing, and the flash rate was about 1.5 per second. I noticed that as the input voltage increased, the flash rate also increased.
The voltage across the capacitor had risen to about 1.28 volts, and seemed to have leveled off, as if the amount of current drawn by the flasher was equal to the current from the charging LEDs. I thought the flash rate was a little bit too fast, so I added another resistor to the 680k total, bringing the total up to 1 Megohm. The flash rate slowed down a bit, but the charging current was now more than the flasher current so the voltage across the 1 Farad capacitor started to rise again.
Later in the evening I checked the voltage and it had risen to 1.58 volts, which is almost as high as the maximum voltage that I measured on the 10000 microfarad cap. This indicates to me that the flasher current is slightly less than the charging current, and the circuit should fully charge in bright sunlight. I then turned off the task light, and left the circuit flashing overnight. I woke up this morning after about 8 hours and the voltage had dropped to 1.225 volts, and the LED was still blinking, like it should be. This means that the 1 Farad capacitor is large enough to sustain the flasher for most of the night.
But one thing I haven’t tried is to charge the capacitor with direct sunlight. I have been using the task light mainly because it’s cool and overcast outside and because I can charge it when it’s dark. But I have thought about a few possible problems. The LEDs have a lens that should focus the sunlight on the LED chip when it’s pointed at the sun, but what happens when the LED is held fixed, and the sun moves across the sky? Will it get charged for only a short time when the sun is focused directly on the chip? Even so, will this and the remaining charge time be enough to fully charge the capacitor? I won’t know until I try it out outside.
One possible solution would be to arrange the LEDs so they each point to the sun at a different time of day. They would point in an arc from east to west. As the sun moves across the sky, each would get maximum sunlight at a different time of day. But will this be enough current to fully charge the capacitor?
In the past I have modified the LED lens by sanding it down flat and then polishing it with very fine sandpaper and finally toothpaste to get the lens to be clear. This spreads the light out, and would allow the sunlight to hit the chip from a much wider area. I can do that with these LEDs, but I don’t see this as being practical when I need to do four LEDs per flasher times two dozen flashers – that’s nearly a hundred LEDs.
Another solution would be to add more LEDs. Red LEDs are inexpensive, probably about ten cents apiece. But if I have to use more than 6 or 7 of them, it would probably be better to buy a 2 volts solar cell for $1.15 (USD) from Futurlec.com. One or two of these could probably charge several 1 F capacitors, and supply enough current to run several flasher circuits. But then I would have to run a pair of wires between the flashers, and this Xmas tree light wiring mess is what I was trying to avoid. I would rather have each flasher completely independent from the others.
There is another thing that I need to do. The current from the LEDs may be enough to allow the flasher to use more current. At this time I have a 1 Megohm resistance and a 1 uF capacitor, which gives a flash rate that looks okay. But I may want to reduce the resistance and increase the capacitance so that each flash uses more power and is brighter. It’s a matter of getting a balance between how much current the LEDs can supply in bright sunlight and how much current the flasher can use during darkness and still flash for about 6 to 8 hours.
Sunday Nov 24 – Since it was sunny, I decided to put the four LEDs outside. I mounted them high enough to get hit by the sun all day long, and ran a pair of wires from it, through the wall, and inside where I can monitor it without going outside. By the afternoon the 1 F capacitor had charged to only 1.025 volts. I think the problem is the LEDs get direct sunlight only a fraction of the day, and the rest of the day the sun doesn’t hit the chips directly, and they put out very little current.
Update Nov 28 – The day was sunny, so I went out and pointed the four LEDs directly toward the sun for a few minutes. I tried to get the shadow below the LEDs to be right below them, hoping this would align the lens and the chip. Later, at dusk, I measured the voltage across the 1 F capacitor and it read 1.238 volts. This was higher than past days, but I can’t say whether it was because of the time I aligned the LEDs or if it was because it was a sunny day.
In the past few days, I have watched the LED blinking late at night, and the voltage was then only about 1.025 volts, so I would say that the four LEDs are capable of doing an adequate job of charging the 1 F capacitor if the sun hits them long enough.
One idea I came up with was to use the plated disk from a bad hard disk drive. These have a mirror finish, and would make a good mirror to concentrate the light on the LEDs. The trick seems to be getting the disk or disks mounted so that it or they direct the light properly to the LEDs.