Model Aeroplane Navigation Strobe / Battery Level Indicator
The following article is a copy of a submission I made to Silicon Chip (http://www.siliconchip.com.au) which printed an abridged version in their January 2001 issue. I hope I am not breaking any copyright law by giving the full version here.
If you are interested in model aeroplanes and have the capital, a radio controlled electric park flyer is a must have. An RC aeroplane under 400 grams floats around like the rubber models of my youth and contact with the ground doesn't seems to involve that sickening crunch that comes with larger models. Finding a deserted park in which to fly one (that propellor can still be dangerous) and waiting for a windless day isn't always that easy but it sure beats driving an hour to a club field. Park flyers are definitely stress beaters and are also just plain fun.
Two things I've discovered about calm days in Canberra. One is they tend to be a little damp, so make your park flyer water resistant. The other thing is that the calmest time is in the twilight. In these low light conditions a navigation strobe would be both useful and practical. LED's could be used but have a narrow viewing angle, so I chose miniture bulbs instead. Since the propellor sucks a fair amount of juice anyway, the 50mA @ <10% duty has little effect on the battery endurance.
Another aspect of park flyers is that they generally need some power to fly at all and (depending on the model) have an appaulling glide angle (slow flight often = lots of drag). You can tell the battery is about to go when the model looses it's umpth but this can leave little time to set up a good landing approach. I decided that a strobe light that beat with increasing urgancy as battery voltage fell might be useful, hence the following circuit, designed largely by trial and error.
This circuit should work with 6, 7 or 8 cells as zener D1 and trim pot VR1 make the duty cycle of the 555 timer chip adaptable and the transistor Q1 limits the lamp current to 50mA. R1 and D1 provide a near stable voltage so that C1 (I used tantalum 35wV) is charged through R2 and R3 at a rate irrespective of the supply voltage.
I didn't want to waste too much current biasing D1, so the actual voltage across D1 is closer to 3.4V. Small zeners in the range of 3.3V up to 5.1V will probably work. A reference diode such as the LM336 might be better but watch that bias current.
The circuit is setup with a fully charged battery using VR1 to trim the control voltage so that the 555 triggers when the voltage over C1 is about 95% of the zener voltage. This gives a strobe about once a second. Capacitor C1 then discharges through R3 with sufficient duration for the lamp filaments to reach their operating temperature. For the sake of component count, C2 could be left out but it may prevent problems in the field. VR1 can be replaced with a fixed resistor to ground but the internal voltage divider impedance in the 555 varies with the manufacturer and you'll be lucky if a single resistor will do the job.
I've labelled the lamps as F for fuse as I fear this is what they will do in time. My father once told me that the reason light bulbs tend to give up when they are first turned on is that the cold filament resistance of the bulb is much lower that it's operating resistance. As such, when the voltage first appears over the bulb, any weak spots in the filament will be given a dose of a low resistance current and therefore heat up much quicker than the rest of the filament. 1000 cycles later the crack at the weak spot is complete. This shouldn't be a problem for the average park flyer as flights tend to last only for 5 to 10 minutes (about a 1000 stobe cycles!). However, my supplier has lithium cells on his flyer and gets bored doing touch and goes for 45 minutes!
In an attempt to lengthen bulb life I have included a current limiting circuit. I tried using two transistors in a current mirror but this brought much frustration and eventually I found that the voltage across LED D2 and the Vbe of Q1 for the chosen value of R4 gave a pleasingly constant 1.5V across F1. The circuit will work with 1, 2 or 3 lamps in place of F2 & F3. F2, being green, is in the starboard wing, F3 (which should be red) in the port wing. As such there is a fair length of wire, so to reduce weight I used very fine varnish insulated wire from an old solenoid. The current limiting circuit should prevent an inflight fire occuring but keep the wires apart just the same. For a 1 meter wing, this wire weighed less than I could measure (< 0.5 grams) and gave a resistance of about 3 ohms. The white lamp is in the belly of the model (between the undercarriage - yes, park flighers will R.O.G.) and so the leads it came with will do. The LED included in the circuit is visible in my model in the cockpit and will flash even if all the bulbs have gone (I later found this to be not entirely true as F2 & F3 must bias Q1 to prevent Q1 effectively shorting the LED). Note that I used Orange (cataloged as Amber) instead of Red for L3 as for some reason the red lamps are so thickly tinted that they are much dimmer; also I have only lost one lamp so far and that was a red one (who said I can keep a grudge).
Construction uses the 555 timer as a base on which all compenents are soldered top and bottom. The result was 3 grams ( 2 of which were solder; so I think you can do much better). This may not seem much but when the RC microservos are 5.4 grams and the microprocessor controlled speed controller is 1.3 grams, my circuit seems a bit clumsy (don't you just love/hate the swiss).
Copyright Robert Parker 2000