Craig Johnson, who runs the website www.ledmuseum.org, where he posts some highly entertaining product reviews, had a few choice words to say about this Chinese-made flashlight advertised as "no battery needed."
He writes in the Discover Circuits forum: "I measured the open circuit voltage at 2.3V and measured the short circuit current at 37ma in sunlight.. I started tracing out the circuit used in this path light assembly. It appears that they used a 1N5819 schottlky diode between the solar panel and a single 1.2v NiMH rechargeable AA cell. The 1.2V from the single cell is too low to power any LED directly, so they used a transistorized (no ICs) DC to DC converter, to boost the voltage high enough to drive a single, most likely yellow, LED." See circuit here. Dave's further calculations revealed that while the light lights, it's only good for about six hours -- not nearly enough for the long Colorado winter nights. In other words, he says, "this design stinks." measured the open circuit voltage at 2.3v and measured the short circuit current at 37ma in direct sunlight. I started tracing out the circuit used in this path light assembly. It appears that they used a 1N5819 schottlky diode between the solar panel and a single 1.2v NiMH rechargeable AA cell. The 1.2 volts from the single cell is too low to power any LED directly, so they used a transistorized (no ICs) DC to DC converter, to boost the voltage high enough to drive a single, most likely yellow, LED. During sunlight charging, the four solar cells act as a constant current source, feeding a maximum 37ma of current to the single battery cell. Other solar powered path lights use a higher solar panel voltage, which can charge two AA cells. The advantage of two cells is that the LED driver circuit is nothing but a current limiting resistor connected to a simple circuit, which turns off the LED when there is sunlight. This is typically done by monitoring the voltage from the solar cells. The thing I picked up seemed to be a cheap knock-off. But, would it work? Let’s make a few assumptions. Let’s say that this thing was placed in a sunny area and got a full 6 hours of sunlight during the day. With a current of 37ma, the current pumped into the rechargeable battery would be 37 x 6 or 222 milliamp-hours into a 1.2v battery cell. That is 1.2 volts times 0.22 amps for 0.27 watt-hours of energy. Now, let’s assume that the yellow LED had a forward voltage drop of 2 volts at 20ma and was driven by this battery. We will not even factor in the losses of the DC to DC converter or the battery charging losses. Two volts times 0.02 amps is 0.04 watts. If we divide 0.27 watt-hours by 0.04 watts, we get 6.75 hours. That means that the energy during those 6 hours of sunlight would be enough to keep the path light lit for about 6 hours at night. At least here in Colorado, that would not be long enough to make it through the night. If the sun went down at 5:00 in the afternoon, the LED path light would turn on but would be off again by 11:00 PM. The battery would be dead. Maybe they drive the LED with less current to extend the operating time. But, overall, I think the design stinks. The manufacturer was trying to save money by using only one rechargeable cell, and a very small solar cell area. Sure, the light lights but not for very long." "Horse puckey! The LED Survival Light is a plastic flashlight that is supposed to be self-charging (by giving it a shaking or "spanking the monkey" motion), but this shake mechanism is just a phony - the flashlight actually operates from two CR2032 disposable lithium coin cells. The "magnet" inside is just a steel rod, and the two wire leads from the coil are shorted to one another." Apparently, though, they aren't all bad, as one of Craig's readers reports: "My light looks exactly like those pictured, right down to the number of bumps on the body, but the magnet actually IS a magnet! I can lift the unit by holding a jeweller's screwdriver above the magnet. The leads from the coil aren't shorted, and are soldered to different connections on the board. I removed the two coin cells and turned the light on, and it works, but the output is very dim. There is a small capacitor on the underside of the board that, I assume, is charged by the magnet/coil. The light was re-assembled and turned on for 42 minutes with no discernable dimming, using the CR2032s, no doubt." Morale of the story: Buyer Beware, when you're shopping at a dollar store!