When the battery pool in your ship can feed app. 300 mA permanently into the flash, you should do exactly that. That way people with a poor FLARM (there are by far too many of those) or people with no FLARM at all (there are too many of those also) get warned too.
However, most glider pilots do not suffer from such excess battery power. Unused battery capacity of 5 - 10 Ah is scarce.
In my DG I do not have such an opulence of energy, a constant loss of 300 mA throughout the flight cannot be tolerated, even as I buffer my batteries with solar panels.
So what's to be done ?
One possibility would be to reduce the duty cycle of the flash (i.e. down to 4 %). With the How2Soar flash, that is an easy feat. It can be programmed to (almost) any duty cycle. Between duty cycles 20% and 4% there is factor of 5, by which the average power consumption would decrease. 300 / 5 = 60 mA in average would be tolerable.
I just don't want to do that. I am aiming for maximum perceptibility and luminosity and I stay with a duty cycle of 20 %- --> Goethe's last words : "more light".
It should be evident to any of my kind readers that a permanently active flash most of the time enlightens the sky and the clouds, but rarely ever another pilot coming my way.
To switch the LED flash manually, when a plane comes at me, and switching the flash off again, when that plane disappears, does not seem to be a too brillant idea.
That is why I build the "CalGnome", a controller for the LED flash, which digests the FLARM data stream (NMEA protocol) and activates the flash when there is reason to do so.
Now, that sounds trivial:
- Whenever the FLARM detects another plane, the flash gets activated.
- Where I fly (Germany and Alps) that would be equivalent to have the flash running constantly, to have your battery constantly depleted by 300 mA.
- You monitor the data stream and whenever a FLARM alarm appears, you activate the flash.
- That is also an off-target !!
The FLARM screams also, when an opponent closes up from behind.
Why should the flash be activated then ?
- And ...
FLARM alarms are notoriously late, too late so that starting to flash would be a reasonable means to warn your "opponent".
The CalGnome takes a different approach. From the information derived from the FLARM datastream it generates something I dare to call `situational awareness`:
- Is there another plane the pilot of which can see my flash without swivelling his head by dangerous angles ?
- Is that pilot in an altitude range where he could become a potential threat ?
Only if both conditions are met, there is reason to activate the flash.
If you compare the number of situations which the CalGnome deems appropriate to activate the flash, the FLARM alarm situations are rare (ratio 5:1 - 20:1). On top of that there are those situations where the FLARM screams but the CalGnome does not activate the flash because the flash cannot be seen by the other pilot.
I have performed an analysis of NMEA-Logs stemming from 28 flights adding up to 121 flight hours, inside Germany and in the French Alps. The result shows that in average between 1% and 11% of the flight time my LED flashes would have been visible to another pilot (meeting above conditions). FLARM alarms sounded only during less than 1 % of the flight time.
Overall result: The permanent load of 300 mA is reduced to 30 mA (or much less). The CalGnome itself needs about 15 mA. The remaining 45 mA are well within my tolerance of power usage. And it is even less than the power consumption for the 4 % duty cycle flash.
A permanently active flash is always advisable if you can spare the extra energy.
If you have to economize on energy, there is no way around some sort of device which activates the flash based on FLARM data.
If you stick to simple solutions (see above), then you will activate the flash too often or not often enough. And you might warn your opponent too late.
It is better to base the decision to activate the flash on a device with 'situational awareness' like the CalGnome, which produces 5 - 20 times more flashes, and more early flashes.
I am a bit shy to provide my own experiences as they certainly are biased and subjective. Nevertheless they might be instructive:
Since 2014 - my first year with LED flash and CalGnome - I have experienced far less situations of threat by head on encounters, compared to prior years. It became obvious to me, that pilots coming my way did see me earlier, made wider evasive moves, avoided scraping past very close.
I am certain, the analysis of NMEA data would have yielded a totaly different result, had I flown without flash and CalGnome. I am certain there would have been many more head-on encounters with a screetching FLARM.
The increase of reception area and the availabilty of ADS-B data make a Power-FLARM worthwhile.