I found this 1.5V switched capacitor flasher circuit on talkingelectronics.com (I can’t get a direct link to it). It flashes a red LED; I haven’t tried any other color but the similar Bowden circuit will brightly flash a yellow LED. This is essentially the same circuit I blogged earlier, except this TE circuit uses only two transistors. My measurements show that it is oscillating at about 1.1 Hz, and draws less than 0.5 mA average current at 1.5V. The circuit is still (not brightly) flashing at 1V, and can be barely seen flashing at 0.8V. The red LED is what I’m measuring; other colors such as amber and yellow have a higher forward voltage and may flash brightly as long as the cell is fresh but may dim well before the cell is depleted. Several of the yellow ones I built have bright flashes, but I haven’t checked how they hold up as the cell is depleted.
This circuit is essentially the flasher part of the Bowden circuit, without the third driver transistor. I’m assuming (but I don’t really know how accurate that assumption is) that the flash is not as bright as my Bowden flashers, because the current draw is not as high. I’m assuming this because the C2 in Bowden’s circuit is more than twice the capacitance, and R3 and R4 are less than half the resistance. However the Bowden circuit has the additional transistor; that and the power dissipated in R2 may reduce the difference.
I made very few changes. The original circuit used a BC547 and BC557, which are not high current switching transistors. I used the ubiquitous PN2222A and the BC327-25 for the driver. TE recommended that R1 be from 100k to 470k. I have found that using a resistor lower than 1 Meg often causes this type of flasher circuit to stop oscillating when the supply voltage is at 1.5 or so volts. I usually recommend that this resistor be 1.3 or more Megohms. With that resistor, and a 1 uF capacitor it flashes about once per second.
This TE circuit puts out a bright flash and draws very low current, giving long battery life. It’s simpler and cheaper than the Bowden circuit, but the difference should be accurately quantified so that the builder can make a judgment on whether or not the Bowden circuit has advantages that outweigh the added cost. Still, the experimenter can build this circuit and expect good performance from it with red, orange, amber or yellow LEDs. I tested one of the yellow LEDs for current and forward voltage. At 2.4V, the current was almost 30 milliamps, and at about 20 mA the voltage was about 2.3 volts.
Update Feb 28 – Since the above, I’ve built a few more of this circuit. It’s a bit smaller than Bowden’s flasher, since it has one less transistor. I want to try experimenting with two flashers on the same board, driving a single LED. I would like to get the LED to look like it’s flickering, similar to a candle. This may require more than two flashers to get enough randomness to simulate the candle. I think when I have a few more single boards built, I may try to wire them together. More in the near future, I hope.
Update Mar 2 – I wired two of the boards together. I removed the LED from one board and wired the junction where the anode was to the same point on the second flasher. I really wasn’t surprised when I found that the two flashers synchronized so there was only a single flash. The two flashers are coupled too closely, so that most likely the faster one triggers the slower one to fire early when the faster one fires. My idea has failed to produce the results I wanted.
One thing, I haven’t tried experimenting with the 10 nF capacitor across the 1k resistor. Most circuits of this type that I’ve seen have no capacitor across the resistor. If there is one, it’s usually much smaller than 10 nF. But I have included it in every one I’ve built. I have a bag of a few hundred of them, so I figured that it was almost no cost to put it in the circuit.
Update early Nov, 2013 – I have drilled and populated seven more boards for decorations during the Xmas holidays.
Back to experimenting…
I’ve assembled this circuit and it worked very well. It flashes at 1.2hz very bright.
It was capable to lit a blue led using a single cell.
I put one on a 50 farad supercap and left it on the desk and it flashed for more than 30 hours, however it was getting dim toward the end. I left for work and it quit sometime during the day. What’s this mean? Well, the circuit uses only a half a milliamp average power. It will run a long time on a AA or even AAA cell. I’d estimate several months, but I haven’t timed one. Six months is just a guesstimate.
On the other hand, the circuit abruptly shuts off because the total voltage is not enough to forward bias the LED. So if you have it flashing on a blue or white LED, as the cell voltage drops the LED will quickly dim and go dark in a much quicker time than if a green, yellow or red LED is used.
More circuits based on the charge pump technique:
http://www.youtube.com/watch?v=JenPZMHREfg
http://www.instructables.com/id/Broken-nose-preventer/
That “brief” Youtube video would have been shorter than 7.6 minutes if he had skipped the verbiage about the album. I think this particular circuit is unnecessarily complicated. I haven’t done any analyzing to figure out what the extra transistor does, or if it’s really needed. From my own experience with these, I predict that with the flash rate of once every 5 or 10 seconds, that D cell will outlive the guy. My Blue Blinky ran for two years on a partly used AA cell, at 1 flash per second. I think the D cell will last for much more than ten years, maybe 30 years! He should have sped the blink rate up to 2 per second. Thanks for the links.