by Lyle S. Powell, Jr.
117 El Camino Corto Walnut Creek, CA 94596
I'm one of those people who reads accident reports, and
subscribes to several publications where they're featured. Over the years it
keeps re-occurring to me that fuel system related accidents are by far the most
frequent single type of non-pilot-error accidents, especially the serious ones,
and especially in homebuilt airplanes. Many fuel system related accidents also
involve pilot error, such as inaccurate knowledge of fuel quantity and
mismanagement of fuel selector valves. However, it is my feeling that these are
chiefly system errors more than pilot errors, because such systems often invite
There are essentially three types of fuel system
emergencies. The first and most important are those occurring on take-off or
initial climb, when there is insufficient time, however well managed, to correct
the situation. The second can best be called running out of fuel. Some of these
are also system-invited by such things as having tanks whose capacity varies
with the attitude of the airplane when fueled, or gauging systems of poor
accuracy. The third major type are those that occur on approaches and go-arounds.
Most of these are definitely system related.
On take-off and initial climb, sudden engine stopping or
serious power loss occurs more frequently in homebuilt than factory airplanes.
This is probably a reflection of repeated experience and standardization by the
factory airplane builders. However, some of these employ antiquated systems and
non-ergonomic practices, fuel selector valves being the most prominent. It is
absolutely amazing how often fuel selector valves are mismanaged under stress by
even the most experienced pilots. Another frequent one is the vapor-lock
incident or accident. Almost all of these are system related. One of the many
causes is the engine driven fuel pump which is such a good 'teapot" for boiling
fuel. When the boost pump is plumbed in series, rather than parallel, and when
the engine (and pump) is hot from waiting for take-off, or from lean cruise
followed by descent many "carburetor ice" accidents occur, both on approach or
go-around, and on liftoff. The only reliable thing that carries the calories out
of that hot fuel pump on the engine is the flow of fuel itself. When the
throttle is at idle (for descent or waiting for takeoff) there is precious
little flow, so bubbles form and have a hard time getting through the small
openings into the carburetor needle valve orifice or fuel injection servo unit.
When power loss occurs on go around or take-off, even it
proper and immediate valve switching is done (if that is what is needed), the
time required for reestablishment of sufficient flow into carburetor or servo
unit is often too long. Part of the problem is due to the tremendous demand of
the engine for fuel at full throttle . . . it's usually 2-1/2 to 4 times the
usually thought-of cruise flow, and catch-up in this circumstance is hard to
accomplish. Also, this is a time when boost pumps cavitate from air inhalation
from an empty or near-empty tank.
Homebuilt airplanes have a tendency to have fuel system
accidents early in their careers, during the learning and sorting-out phase of
the pilot/builder as well as the airplane. Fuel system accidents include not
only outright failures of devices in the system, but things we rarely think
about, such as:
- Not knowing how much fuel you can put into a tank
because of attitude sensitivity or venting.
- Small vent tubes easily obstructed by a single drop of
water or an insect.
- High pressure boost pumps cavitating with interruption
or surge in flow.
- Fuel selector valves sticking or not having clearly
defined detent positions.
- A leaking gascolator gasket admitting air bubbles into
the system, yet leaks little or no fuel.
- Vibration-induced cracking and leakage of spare fuel
- Split flares in metal tubes producing leaks, or inlet
of air into system.
- Inadvertent flap valves in fuel hoses produced by
improper insertion of connectors.
- Foreign bodies in the tank jamming boost pump piston or
breaking carbon vanes of high pressure pumps.
- Inadequate sized elbows or other fittings in the system
producing bubbles in the flow of fuel.
- Foreign bodies obstructing finger screens, gascolator
screens or filters that are too small.
- Leaky carburetor floats.
- Leaky fuel injector servo diaphragm (beware of the
- Leaks in diaphragms and edges of diaphragm in engine
- An unsupported vibrating fuel hose that partially
- Worn or grooved connector fittings that leak air or
- High pressure systems (fuel injection) are considerably
more critical with respect to leaks and obstructions than low pressure
systems, for obvious reasons, and experience bears this out. Also fire
hazards are greater with high pressure systems due not only to the higher
pressure but because of the increased footage of plumbing and larger number
of connections in the engine compartment. Boost pump failures and pump
priming failures are also more prevalent here.
- Gravity fuel systems, while seemingly simple and
reliable, are plagued by very small supply pressures and ease of
interruption. For instance, the minimum pressure required by most current
carburetors is 1/2 lb./sq. in. This requires a gravity column of 18 inches
-not counting any losses for tubes, filters, valves, elbows, connectors,
etc. - or the occasional sticking of a float needle valve. For small engine
applications only, where small flow requirements prevail.
- Air being sucked into the flow of fuel can be as
obstructive as vapor lock bubbles. This is another reason to have little or
no suction component to the fuel system. Fuel leaks are much
easier to find than air leaks because air leaks don't always leak fuel.
This is only a partial list of potential problems, but it
is a sufficient list to illustrate the character and magnitude of the problem.
Often the homebuilder (and the homebuilt designer) gives the fuel system
inadequate consideration, or simply follows one of several standard production
examples. Too often they don't realize the pitfalls of small variations from
specific applications, or lack an overall understanding of the problems, and/or
the shortcuts that homebuilders are likely to take for their own convenience. I
believe that the underestimation of the critical nature of the fuel system is
the largest single source of poorly designed or fabricated fuel systems.
Following is an outline-type summary of fuel system items
to observe when designing or building your fuel system.
Into which a known amount of fuel can be put each and
every time (not attitude, tilt or vent sensitive). Of reliable mechanical
construction, unlikely to develop leaks with time and vibration and unlikely to
present an unusual hazard in a crash landing. This requires substantial
resistance to rupture on impact or deformation. Must not have low spots behind
baffles and in comers for collection of water. Vibration is worse in 4-cylinder
airplanes than others, and must receive generous respect as a destroyer of
structures and producer of leaks.
Fuel Tank Caps
Must not leak fuel, air or water. Expensive, but available
and necessary. Look at those caps - take them apart and examine them. Small
details are important here.
Adequate depth and size, with screen and drain valve as
necessary. Do not tolerate a main tank without a real sump. Auxiliary tanks,
with good lowpoint drains and a "no take-off" restriction 0. K. without sump.
Protection Prevention of fuel being thrown away from sump outlet by
uncoordinated flight or turn just before take off, by use of slosh gates or
check valves and baffles in tank. Necessary.
3/8" tubes or larger to prevent a frozen drop of water
from obstructing. As short a run as possible, especially if horizontal (because
of water droplet precipitation), with non-icing opening (any one of several
types). Backup second vent highly desirable.
As simple a system as possible with all on or off if
possible. An amazing number of accidents occur from pilot misplacement of valve
handle or valve sticking, even from handles breaking off. Also even when
properly changed, a long interval is required before engine starts. Selector
valves are inherently dangerous and should be recognized as such. One
alternative is a separate ball-type valve for each tank, arranged so that the
handle position is obvious. Also these valves are more reliable and don't stick.
Must be inside sump or have short gravity-fed inlet,
otherwise very often will not reprime if run dry, especially fuel injection
boost pumps. Do not try to suck fuel uphill or forward. It pulls bubbles into
the fuel inviting cavitation. Acceleration occurs forward and upward on take-off
and climb for a sustained period of time and fuel moves backward and down, and
that's where the inlet of the boost pump should be if it is not in the sump.
Protect inlet of pump with screen or filter adequate to protect the pump from
jamming due to foreign body. Such filter must be inspectable and cleanable.
These are often provided in the pump body by the manufacturer.
Should be direct from boost pump through filter to
carburetor or fuel injector servo. Have engine driven pump plumbed in parallel,
not series, so that possible vapor lock in engine driven pump will be bypassed.
A check valve may be necessary, depending on pump type.
Engine Driven Pump
Requires shroud for positive-pressure ventilation to cool
it, thus minimize fuel boiling (due to accessory case and oil temperatures which
Fuel Lines and Devices
Should not be exposed to heat anywhere, for two reasons:
To prevent vapor lock (bubbles whose surface tension
make them resistant to going through small holes).
To prevent fire in case of accident, or fuel leak in
Particularly avoid proximity of fuel lines or carburetors
to exhaust pipes radiated heat is more intense than most people imagine. This
heat acts as an ignition source in case of accident, or a fuel leak, or a crack
in an exhaust pipe in flight. Metal heat shields are often necessary because
most heat transfer in cowling is by radiation, not convection or conduction.
Examples are the metal shields between an exhaust pipe and adjacent hoses or
wires seen in many factory airplanes.
Should be of well engineered type and size and protected
by fire-sleeve in engine compartment.
Gascolators are not sacred devices, not even very
efficient ones. They were really designed for use with fuel tanks without sumps
or sump drains. With sump drains they become unnecessary, or at best
supplementary. Often they are sources of fuel leaks and air leaks into fuel
systems. Also they are sometimes vulnerable to rupture in case of accident.
Where tank sump drains are provided, good fuel filters of several types are
better than gascolators and are safer, less prone to leaks and damage. Must be
of adequate size and accessible to inspection, draining, cleaning or
replacement. Beware of very small automotive filters which could obstruct in
flight from a slug of dirt in the fuel. I am using a FRAM HPG-1 fuel filter in
my Glasair. It is commonly used in racing cars and boats, has an excellent
service experience. It has a 13 ounce capacity, a steel case into which you can
put a drain valve. Expensive and bulky, but a good example of what is needed.
Two brands are currently available (Wag-Aero; A.I.R.
Corp., Oakland). A very good idea - either in sump or filter can.
Reliable, backup simple mechanical type gauge or
sight-tube gauge advisable for last few gallons in addition to standard gauges.
Float switch with warning light is another good alternative (Aircraft Spruce).
Fuel gauges are justifiably mistrusted, but they are also usually of low
quality. Reliable separate gauging of the last 1/10 or so of fuel can be very
accurate. Flow meters and totalizers are not a substitute for fuel gauging
because they are so dependent on accurate knowledge of how much you start with.
Be sure to have some back-up gauging or warning system beyond standard gauge
system, or a reliable spare tank.
Spare Fuel Tanks, Header Tanks, Etc.
All have definite problems, including selector valve
hazard, but they are a reasonable alternative it designed well. Using a
vibrating firewall as one wall of header tank is a questionable practice unless
it is specially reinforced and stiffened. (Touch that firewall in flight
sometime.) Again think of a survivable crash landing or an in-flight fire. A
VW-like standpipe in the main tank is one alternative to a spare fuel tank - or
a separate tank within the main tank that fills automatically -or a spare tank
that transfers into the main tank. Outer wing panels are the best location for
spare tanks, for structural as well as safety reasons.
Air Inlet System To Carburetor or Fuel Injector
Must be of adequate size and especially not obstructed by
a too-small air filter. Better no filter than an obstructive one, because the
obstructive one can seriously disturb fuel mixture and produce erratic
throttle-mixture correspondence. Air filters are highly desirable but must
receive the same design consideration as any other system and not simply yield
to what is convenient (frequently seen in homebuilts). Be sure filter will also
act as flame arrestor in case of start-up backfire - it can save your whole
airplane. This is done by containing the filter in a blow-out and suck-in proof
container. A curved elbow type of air entry into a carburetor is poor practice
because of inertial lamination of airflow into the carburetor. A plenum, horn or
diffuser entry is much better, and removes those dead spots at some throttle
settings also higher power.
From engine pump, inlet spider, inlet plenum boxes as well
as drains from filters and gascolators, must be overboarded in a safe place away
from exhaust. A manifold collector and single drain often useful here.
Carburetor Heat Source
Standard and necessary. Can be easily combined with cabin
heat. Pulling carb heat cable turns off cabin heat with two-way flap valve. Be
sure to overboard any unused carb or cabin heat so there's a constant airflow
over shrouded exhaust pipe. Otherwise that segment of pipe will bum through and
become an early carbon monoxide and/or fire hazard.
In case of gear-up landing, or gear collapse accident in
fixed gear aircraft. Any exposed or vulnerable fuel-containing structures such
as sump, filters, drains, gascolators, etc., should have mechanical protection.
No drain valve or such structure should project where it can be easily broken
off in a belly landing. Longeron-like braces in belly pan is an example of a
mechanical protective structure. A strong belly plate under this area is
Putting fuel into aircraft tanks deserves some thought.
The flow of fuel through a hose and nozzle creates static electricity, and a
discharge arc sometimes occurs to the filler neck of the airplane - explosion.
So, in fiberglass or other composite airplanes, it is desirable to ground the
metal fuel filler cap ring. This is true because any mass of metal plus adjacent
semi-conductor fluid (gasoline with some moisture in it) has some capacitance.
As such, it becomes the target for a static electric arc from the fuel hose
nozzle (which may or may not be grounded) or the pouring spout of a jerry can.
This filler-neck grounding should be done with a wire (18
gauge or so is enough) attached with a good AN plated bolt and washer through
the ,aluminum ring to a good plated crimpon fitting. These details are to
minimize dielectric corrosion of the dissimilar metals. Aluminum bolt or rivet
and wire could also be used. However, the experience with aluminum wires and
connectors in the presence of moisture is poor. The wire should be mechanically
supported properly into the cabin area where it is attached to the ground buss
through a resistor (approx. 1 meg OHM 1 watt). This resistor limits the power of
any static discharge. What it actually does is spread out the time of the
discharge from instantaneous to several micro seconds. This then replaces the
arc with a corona-like discharge which is probably below ignition temperature
for gas fumes. In any case, when fueling, it is good practice to keep the nozzle
in contact with the filler neck. If both the filler neck and the nozzle are
grounded, there should be no problem. But you can't be too sure about some gas
hoses and nozzles or ground connectors to airplane from the truck or pump.
Finger screen in sump should be grounded - by grounding aluminum line or
Fueling from plastic (polyethylene) cans should never be
done because these materials have a very high static electric generation
potential when gasoline flows over its surface, or it is rubbed against another
material. Metal cans are much safer.
The preflight ceremony of draining sumps and other fuel
devices should be taken seriously because it is here that you can best prevent
the most terrifying of aircraft accidents - the engine stopping on take-off or
initial climb. Water is the chief enemy, foreign bodies of all kinds, second.
Always use a cup or container to drain fuel, so that you can am any water or
debris that you drain. If you can't see it, you don't ever know how much to
drain, and every once in a while it takes a lot. Fueling from some places can
produce very large amounts of water and debris - much more than a cupful - even
gallons. Old buried tanks with doubtful maintenance are guilty here, even
trucks have pro: produced such events. I know first-hand of several such events.
Cessna's experience in rocking wings and tail to dislodge
water from wrinkled bladder tanks should be remembered - it was successful. This
applies to other airplanes, too, such as taildraggers where low spots due to
attitude become pockets for water -or nose draggers with multiple baffles. Water
droplets on the floor of a gas tank seem reluctant to move toward the low spot
(sump) unless agitated, especially with minimal dihedral wing tanks, apparently
due to surface tension.
Water is soluble in gasoline to a limited extent, and this
is particularly important in winter (see article by Niel Petersen Sport
Aviation, December 1986). As fuel cools in the tanks overnight, some water
precipitates out as droplets. This accounts for some of the .moisture of
condensation' even when the tanks are full. Also if it is cold enough, these
droplets can form ice crystals or slush, which can obstruct fuel outlet screens,
even to the point of collapsing them in flight In the cold winter areas this can
be important not only as a pre-flight consideration, but on long flights at
altitude, where the fuel has time to become cold. Jet-powered aircraft use fuel
heaters or water-dissolving additives ("Prist") in their fuel for this reason.
However, their fuel has a greater solubility for water, and their flight
altitudes are higher - but the problem is essentially the same.
This list is, of course, incomplete. My emphasis has been
on those items which seem to me to be most important from a safe-design
standpoint. Homebuilders suffer from "ran out of room" problems just when
something like a fuel filter or an air filter demands a place, then compromise
occurs, and the last items on the list get the poorest design. Don't yield to
this - be willing to go back and rebuild or rearrange things so that fuel
priority is properly respected.
One last thing. If; you must put fuel system components in
the engine compartment (where all the heat and ignition sources are), group them
together and build a good metal, positively ventilated, box around them. Place
the vent exit as far away from the exhaust pipe as possible.
In fuel systems, the enemies are heat, water, leaks,
foreign bodies, static electricity and devices that invite human error.
ABOUT THE AUTHOR
Lyle S. Powell, Jr., EAA 38012, has been flying for 23
years (Commercial, Mulfi. Instr., 3300 hrs.). Lyle has owned several
factory-built airplanes, built and rebuilt a VariEze, a Glasair (with 750 hrs.
on it) and is now building a Glasair III.
He offers thanks to Andy Marshall of Marshall Consulting,
George Periera (designer of the Osprey and G.P.-4) and Jim Horn (Voyager engine
instrument design consultant) for their help and advice on this article.