Two weeks ago I flew pretty heavy IFR long x-c (Leesburg, VA to
St Loius, MO) the plane sat on the ramp in the rain overnight one
night. Next morning the "Alternator Failed- the battery is the only
power source" red warning came on.
OK, I'll state the blindingly obvious: You've got a loose connection
I think the rain is a clue; a marginal connection got
wet, maybe got corroded a little, and is continuing to cause you
Also, Odin is really the wrong Norse god for this sort of thing. You
need to talk to Thor.
I continually watch the system info to make sure the vlots are 28 +/-
when the engine is at cruise, so I know the alternator is working.
Well, you _know_ that the G1000 says "28" on it, so you _think_ that the
alternator is working. I can think of a couple of reasons that the
G1000 might be lying to you.
First, the display update rate may be much slower than the sampling rate.
This is done on purpose; it's distracting to have the last couple of
digits of a digital readout jumping around when you're trying to read
it. The alarm, however, may be set to trigger on any low voltage reading,
including one that doesn't last long enough to show up on the display. A
very short glitch might be enough to trigger the alarm but not enough to
show up on the display. Hand-held digital voltmeters usually have a
faster update rate, but a glitch might still sneak by. A digital
storage oscilloscope will almost surely find it, but not every shop has
one of those.
Second, it is possible for an alternator to partially fail but still
sort of work. If one of the diodes (out of at least usually three) in
it fails, it may still put out enough volts to charge the battery, but
it won't be able to put out full current, and the voltage will have a
lot of ripple on it. The G1000 could be looking at the ripple on the
input voltage to detect partial alternator failure, or it could be that
the troughs of the ripple are enough to set off the alarm.
It looks like the G1000 can tell you if your battery is charging or
discharging, which means that there is probably a shunt in the electrical
system somewhere that all of the battery charging current flows through,
and that would be another place to look. (A shunt is a calibrated,
low-value, high-power resistor - it can even be a known length of wire
in the harness.) There may also be a shunt for the alternator. Or,
instead of a shunt, it could be using a Hall-effect current sensor -
this sometimes looks like a little ring that goes around one of the big,
high-current wires in the electrical system. If the connections to
either a shunt or current sensor are intermittent, that might trigger
Another possibility is that the G1000 is seeing an _over_voltage that
it doesn't like. Imagine that you're driving up a hill in a manual
transmission car, and the clutch fairy suddenly steps on the clutch
pedal without you knowing it. The engine would speed up quite a bit
until you figured out what was going on and backed off of the throttle.
The alternator/voltage regulator can do the same thing if an electrical
load is suddenly disconnected, as might happen with a bad connection.
The warning light now predictably comes on during most run-ups, and
stays on for climb out and back on at decent.
This is actually good - if you (or the shop) can make it happen on
demand, it's a lot easier to find. It sounds like it's correlated with
increased vibration, at least for run-up and climb out, which also
suggests a loose connection. Another thing to think about - did anything
else "interesting" happen to the electrical system lately? New battery,
existing battery got so dead you had to put a start-cart on it, new neon
lights underneath, new LED windshield wiper nozzles?
One place to look first might be at the high-current connections in the
electrical system: battery, master relay, alternator output, main bus,
and shunts (if so equipped). If these connections aren't happy, nobody
is going to be happy.
Another place to look first would be in the G1000 operating or service
manual to see what conditions trigger the alternator alarm, and then
chase those conditions backwards from the panel. The manual should say
whether it is making a simple voltage measurement, or also looking at
voltage ripple, current, current ripple, etc.
If the alternator is really easy to get off of the engine, and if the
test gear is available, it can be eliminated in a hurry by pulling it
off and bench-testing it. If it's a pain in the butt to remove, more
tests should be made with the alternator installed before deciding to
remove it. Same for the battery - if it's easy to get at, pull it out
and test it.
This makes me think a sensor is overheating or something.
I'm not sure how likely this is. It's probable that at least one of
the voltage "sensors" is built into the G1000 itself - it'd be a real
simple measurement to make on the main power line coming in. If it
makes any current measurements, those sensors are probably external.
Disclaimer: This is based on experience with electrical systems and
electronic equipment in ground vehicles and in fixed installations.
I don't have an A&P; I don't even have a TG&Y. Your mileage may vary.