Generators, redundancy, and old stories

Many people have asked me why I have generators on my airplane rather
than converting to alternators. My airplane is a PA-30. The only
legal alternator conversion is the InterAv STC. This is a
non-redundant conversion - the alternators used require power to the
field and the regulators use interruptor relays. Thus if the battery
or master contactor fails, both alternators will immediately go
offline. No gear, no flaps, no lights - all from a single point
failure. Unacceptable in a serious night/IFR machine.

Generators, of course, ground the field. Even if the battery or
master contactor fails, the generators keep going. The voltage will
fluctuate and some avionics may die, but some (especially the modern
uncertified ones with good power supplies) should keep working, and of
course flaps, gear, and lights will never know the difference.

An old-time pilot told me I was wrong - that the generator system was
also non-redundant due to the paralleling relay. But of course even
if that fails, you can still run on one generator, I replied. No, he
told me - it can fail and make the whole airplane dark. I never
understood how - until it happened.

For those not familiar with the paralelling relay, here's how it
works. It's actually two relays, each hooked to the armature line of
a generator. When both generators are making power, both relays
close. When they do, they allow current to flow between the
paralelling coils of the voltage regulators, balancing the load. At
other times, the voltage regulators are fully conventional. The
design is actually 1940's tractor technology. At the time, generators
were simply not built to handle 100 ampere loads. When one
particularly large tractor was built that needed 100 amps, the
designers had a choice - design an oversize generator (with all the
tooling this involved) or find a way to make two stock generators work
together. They chose the latter. Redundancy was never the issue.

Piper took this design and transplanted it wholesale into the PA-30.
We pay a lot of money in certification costs to ensure that designs
are not simply lifted from industrial equipment and modified
willy-nilly. There is supposed to be failure analysis and such. Of
course this is the FAA running the show, so that doesn't happen.
Really, I blame myself for trusting the certification process. I
should never have assumed that just because the FAA certified it the
design made sense for an airplane - and it doesn't.

You see, the design was not really lifted wholesale. Circuit breakers
were added to each generator - after the regulator. It was the worst
possible place. No circuit breakers would have been better - but the
FAA wouldn't certify that. The proper place to have them would have
been in the armature line - but that's not where piper put them.

Generators have more than one failure mode. They can fail off, but
they can also fail in an overvolt/overamp condition. This occurs
when, for whatever reason, the field line shorts to the case. This
can occur if the technician assembling the generator is a little
sloppy in soldering in the feedthrough for the field. A little solder
goes over the side. However, the inside of the case is painted, so
the solder is covered with an insulator. When the generator is
assembled, the through bolt shouldn't touch the feedthrough anyway -
but the clearance is tiny. Over time, and with vibration, the paint
can wear away. At first the field will short intermittently, but
eventually the short can become permanent.

The PA-30 is not normally equipped with a voltmenter, so there is no
chance of catching the problem while it is still intermittent. Even
when it becomes permanent, the only indication will be battery
overcharge. Since the ammeter is of the +/- 0-100 amp variety,
noticing a 10-20 amp charging current isn't trivial. It's also not
that hard to exceed 50 amps on the bus at that voltage. Normally the
paralelling relay would distrbute the load, but since the field on the
generator is shorted, it will take all or most of it. The breaker
will pop. Of course if the plane is equipped with a voltmeter, the
pilot will likely notice the overvolt condition (as I did). The pilot
will then troubleshoot, and realize that even turning off both field
switches will not solve the problem. He will then attempt to resolve
the problem by pulling generator breakers, and will quickly identify
the misbehaving generator, pull its breaker, turn off its field switch
(which accomplishes precisely nothing) and use the other generator to
complete the flight.

Either way (whether the breaker pops or the pilot pulls it) the
outcome is the same (minus any damage to avionics). Because the
breaker is downstream of the regulator, the armature from the
misbehaving generator is still delivering juice to the paralelling
relay. Thus the relay is attempting to parallel the working generator
(which is now carrying the entire load) and the idling generator with
the shorted field. THERE IS NO CURRENT LIMIT THROUGH THE PARALELLING
RELAY! There is no fuse, no breaker, nothing. In short order, the
voltage regulator coils will be fried, and both generators will go
offline.

However, it can be even worse. As the regulator on the still-normal
generator begins to malfunction, this regulator will deliver an
overcurrent. This can weld the contactor shut. Then, when the field
goes offline, the armature will short the battery to ground, putting
about a 40 amp load on the battery. This, in addition to the normal
load, will discharge the battery in short order unless the pilot pulls
the other generator circuit breaker. Did I mention that these live
under a floor panel?

Go ahead, ask me how I know...

The fix is simple. Add a fuse to the paralelling line (and a switch
would be better) and move the circuit breakers to the armature lines
where they belong. Now what are the chances the FAA will let me do
it?

Michael