That's the biggest heap of chips I've ever seen to blink an LED and it's not even efficient! It's worse than the old LM3909.
If you want to go nuts with chips, at least use a flyback booster. You can even bootstrap the boosted LED voltage to drive the PWM circuit, allowing operation down to a few mV after a starting threshold is reached.
Hmm. Bit bulky. Currently pondering my design for a flashlight/compass/key-finder. Take a bit of sapphire a centimeter and a half in diameter, and 5cm or so long. Facet the outside a bit to knock the corners off and add a little sparkle. Now bore a hole and take a rod of sapphire of the appropriate size, and carve it to accept a few dice of varying colours, which are run at an average current of a few microamps, and pulsed every 30 seconds for a few milliseconds. The battery is inside a decorative 'totem pole', and all covered by a stainless/silver cap.
The flashlight is triggered (I think, design work continues) by a sharp knock.
Recharging happens due to a tiny bit of solar cell incorporated into the cap, which is charged by an optical charger. (or sunlight)
This was triggered after seeing a nice bit of sapphire on ebay. (pale blue)
Battery a 1/3 NiCd AAAA probably.
http://inquisitor.i.am/ | mailto:email@example.com | Ian Stirling.
I've never used it but the datasheet shows that at some point guaranteed to be no greater than 2.7V, the supply current takes a steep nosedive to off. Another catch to this type of comparator in applications with extremely slow input rates is the steep increase of supply current as input threshold is approached. This could jump to 100uA's or more, mA's in the case of logic Schmitts. Maxim has a good low voltage comparator specifically designed to suppress this effect. I notice the corresponding graph is absent from National's datasheet:-(
In article , firstname.lastname@example.org mentioned...
I have a question along this same line. On Dave Johnson's
he has an hour meter that uses a CMOS inverter as an amplifier. He shows the output tied to input with a 22 Meg resistor, to make it linear. He claims the whole circuit uses only 2 microamps, which seems low to me. I think that the CMOS chip would draw more than that because it's biasing itself right in the middle, halfway between + supply and ground. The supply voltage is 3VDC. The inverter is a 4069B. I don't have one laying around to try and see if it's really true.
###Got a Question about ELECTRONICS? Check HERE First:###
A 4069B is old-style high-voltage CMOS. At 3V rail you are probably at or below 2*VT, thus low consumption.
Since my simulation uses a device-level model (*not* behavioral) I just re-checked the results and zoomed in to the LMC6762 current... At V+ = 2V, I(V+) is nominally 10.25uA, pops up to 1.7mA at the start of the capacitor dump but has fallen back to 10.25uA before the cap is half dumped. Average power consumption of the whole contraption is
150uA at 2V, 250uA at 3V.
At 2V operation I'm smacking the LED with 6.7mA, decaying in 225ms. I've haven't played enough with LED flashers to know how long the duration needs to be to be noticeable.
From my LMC6762 data sheet, the fall-off of current seems to occur below 2.2V unless you're fond of -40°C ;-)
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |