/ . . . . . . .I think we can comfortably assume that
/the primary consideration is to be able to switch a circuit
/with as near to 100% reliability as practically possible.
Why is this a necessary part of a sport airplane
system consideration? Since military specifications
for all things began pouring out of Washington, their
#1 goal has and always will be, to fight a war. In
times of battle, minimize the hassles with the machine
so as to not distract and/or cripple the combatant.
This quest for 100% (or even 99.99%) dribbled over into
aircraft because the major drivers of aircraft requirements
-AND- financers of new designs were military or commercial
transport.
As long as the researchers could tap that very deep well
of taxpayer's pockets, a quest for ultimate device(s)
was well financed and justifiable under the rules of
weapons development.
/That means it opens or closes as many times as necessary under
/any conditions of flight (altitude, moisture, lightning strike
/etc etc) and of electrical health (low volts, high volts...)
/we are likely to find. I would also add 'for the life of the
/aircraft' but this in itself may be a design consideration
/alone. We also need a clear indication of the switch status at all
/times.
Of all single-engine piston airplanes setting on the ramps of
US airports, what percentage of installed switches have
been subjected to extremes of any of the stresses you mention?
Moisture and temperature cycles just sitting on the ramp between
flights seems to be the greatest stress. Switch failures I've analyzed
seem to cluster around old airplanes with relatively low
utilization at one end and, of course, airplanes with very
high utilization at the other end. In the first case, switches
died of old age with realtively few electrical operations
on them. In the later case, switches simply reached end of
service life.
/1. Can the switch handle the current?
/Considerations include: average current, inrush current,
/current under unfavourable supply or loading conditions,
/after deterioration of the contact condition or the mechanical
/latching conditions, during switch on contact bounce, switch
/off arcing.
/Average current....that bit is easy, in fact it is the only
/characteristic commonly published for most switches fitting
/the small toggle bill.
Okay, let's partition off the tasks here. There are very
few loads in airplanes that raise concerns for contact
life due to electrical loading. The obvious ones are
landing lights, pitot heat, and maybe nav lights . . .
Hmmmm . . . we're now down to 6 amps.
/Inrush current .. ahhh .. this is a real demon. Dammed
/hard to measure, the only way I know is with an HF current
/probe and an Oscilloscope. Insert any resistance and the
/measured value can be decimated. Another factor is that
/the inrush magnitude is a function of the supply impedance.
/An interesting exercise would be to calculate the supply
/impedance summing (algebraically as necessary) wire
/resistance, battery int imp, alternator, fuse/cb etc
/and measure the cold resistance of a typical incandescent
/bulb/s for say the nav lights or landing lights.
/I would be interested n the result but my guess is that
/it would be way above the 5x normally quoted.
Actually, not true. 10x is easily hypothesized by measuring
lamp cold resistance but airplane wiring rapidly drops the
multiplier . . . 5 times the running value is right in the
ballpark on most light planes.
/One trap here is the common practice of using a bench
/supply for testing. A bench supply usually exhibits
/significant current limiting, either by design or
/accident. An installation testing OK on the bench,
/but when a well charged battery and alternator is put
/behind it, the transient behaviour can be quite different.
I've found that the alternator doesn't contribute much to
the equation. It's dynamic impedance is pretty high and
even fully loaded, it's good for 50-100 amps. Modern
RG or NiCad batteries can support inrush transients
to 1000 amps or more; but not when driving a load from
the far end of a piece of wire. Loads that toggle switches
are expected to control in light airplanes are just not
very nasty.
/Additionally, modern electronics often has a varistor
/capacitor input circuit. The varistor has only marginal
/impact on the charge current. I myself have had to resort
/to a soft start circuit with a 140Volt DC (rectified mains)
/supply into a 2200uF filter capacitor in an application
/where the switch contacts welded after the first closure.
But that's a whole different ball game. AC mains at the
ordinary wall socket can be charaterized with potential
fault currents in the hundreds of amps behind a 117 VAC
push . . . very hot arcs compared to tens of amps pushed
by 14 volts . . . this just isn't an apples for apples
comparison. The heat generated at during closure-bounce
in your example is at least an order of magnitude greater
if not two magnitudes.
/2. ...under unfavourable loading conditions - stalled
/motor etc hopefully the fuse/cct breaker will protect
/this. But a stiff bearing might cause a 100% variation
/in the continuous load current.
But name me one motor in a light aircraft application
that has any chance of pushing very hard on a switch?
Flaps? Okay, let's assume 5 amps (Big flap motor) fused
at 7 amps. Inrush to this motor could be less than that
of a 10 amp landing light system. A 100% increase in
continuous load due to bearing failure (5 amps
x 15 volts is 70 watts . . . going into two bearings?
They are going to warm them up VERY quick with mucho whining
and complaining).
/I'll bet you have had to change a number of switches
/because the mechanism has failed due 'no apparent
/reason'. I'll wager it was heat related.
Actually, most of the failures I've seen are environmental
whether the faiure exhibits itself mechanically or
electrically. Oddly enough, not 1 hour ago my wife
and I landed after observing some fantastic cloud
structures and a beautiful sunset over the Kansas
wheatfields. As we were rolling up to the hangar,
I turned on the dome light in the C-150 we were using
and the light didn't come on! I wiggled the switch
a bit and the light could be made to flash. These are
the little 69-cent rockers I've been talking about.
You'd think that the landing light switch would be
the first to go after 30+ years . . . but no, it
was the switch that controlled a single, .08 amp
lamp on the ceiling! Corroded internal contacts
no doubt.
/ Arcing... DC/AC etc etc can only lead to a degradation
/ in the contact condition. Another book on this topic, but
/ lets leave it for now.
The #1 switch problem in single engine Cessnas
is degredation of the alternator side of the
battery master switch. This switch carries 3 amps
MAX but it is a slightly inductive load. With
months to years of little blue flames along
with dust and moisture, contact resistance
goes up. The result is a voltage regulator that
goes nuts trying to figure out what the REAL
bus voltage. Nobody suspects switch "failure"
because it's still turning things on and of and
it feels okay . . . but the voltage regulator
can't abide an additional 100 milliohms of
resistance in the voltage control loop. Thousands
of Cessna owners have suffered thousands of dollars
in expense to replace regulators, ov relays and
alternators without fixing a jumpy bus voltage.
Replacing the $17 switch fixed it.
/So tell me, how can anyone know the real characteristics
/of a toggle switch for which the average rated current is
/the only published data?
Easy, don't worry about it. The circumstances where you might
begin to push a switch to limits simply don't exist in
contemporary light plane designs and that's the way it
SHOULD be. An RV pilot isn't going to risk dying because
his AmRam missiles would't arm due to a single switch failure!
/I agree that if a switch is used within it's design limitations
/it will work almost for ever, and I'll also bet the Cessna's
/electrical designer / switch supplier put a lot of effort into
/the design. The switches may look the same as others but what
/are the real specs.
I was there when the rockers were installed in the Cessnas,
they are catalog-item commercial switches. We did run some
silly little cycle-life tests that the manufacturer had already
done and documented but our DER wanted us to test SOMETHING.
Heaven forbid that we didn't have some document on hand thirty
years later to show that we "tested" the things. The total
effort required to bring those switches on board was less
than 200 person-hours on the part of engineering, drafting,
experimental and service parts cataloging COMBINED. People have
been trained (fooled?) into thinking that a great deal of
analysis goes into EVERY certified airplane part . . . it simply
isn't so.
/In my eyes, the mil spec lever switch of guaranteed quality
/looks real pretty. The cute plastic jobs do not.
Not a thing wrong with these switches. But why spend $25
for a switch to operate a $20 landing light bulb that you
KNOW is going to fail? At some point in time you'll
find yourself landing in the dark no matter how good
your switch is. Okay, how about dual landing lights?
Great . . . now you've got redundant switches too. Again,
why spend extra $ to feel good about the switches when
you can replace it in a few minutes without even dragging
you tie on the floorboards?
/Also, WRT solid state switches, one strike (lightning) and
/they may be 15% damaged, or perhaps 50%, or if you're real
/lucky 100% kaput.
I don't know of anyone offering "solid state" switches
as a replacement for ordinary toggle functions. Solid
state relays (opto-couple/triac/a.c.) devices abound and
DC ones are just around the corner with the advent of
low Rds(on) FETs and photo-voltaic opto couplers. But
you only need two or maybe three of these (landing lights
and pitot heat, nav lights) . . . and then only because
you want to control your airpalne with little minature toggles.
All the things you mention are of intense interest
to some people, just not light airplane builders. They
should be designing for failure tolerant systems that
feature low-cost, easy to replace components with
reasonable service life . . and each builder will have
to come up with his own definition of "reasonable."
Here in the US, the average utilization of a light airplane
is under 100 hours per year with less than 10 hours
of night flight. This means that most switches regularly
see less than 200 cycles per year (landing lights a dozen
or so). Most switches die of old age than from any
degree of electrical stress or utilization. The
war machine/air transport requirements are worlds apart
from sport-lightplane requirements. None-the-less, the
word "aircraft" conjures up images of those same machines
and their requirements when you go out to buy parts.
There's nothing 'wrong' with buying the best you can
afford . . . by all means, buy a Lexus if you really
want one and can afford it.
I counsel my builders that there's no reason to consider
a Chevrolet budget airplane to be less safe because
the switches or any other parts were purchased out
of the Allied Electronics catalog . . . design so that
you don't DEPEND on any single item for comfortable
completion of flight.
Regards,
Bob . . .
AeroElectric Connection
////
(o o)
oOOo(_)oOOo
| |
| Go ahead, make my day . . . |
| Show me where I'm wrong. |
72770.552ompuserve.com
|