Ensuring there is fuel aboard the airplane adequate for the planned flight is a basic, primary responsibility for any pilot. Unfortunately, the accident record tells us it’s one responsibility not all pilots fulfill. In our experience, fuel exhaustion, starvation or simple mismanagement accounts for a healthy proportion of engine-failure mishaps. This is true despite the event being resolved without anyone except the pilot knowing about it, or the NTSB opening an accident investigation. While fuel-system complexity can be an argument in defense of any failure to meet this basic requirement, it ultimately falls down because ensuring enough of the stuff is aboard and gets to the engine(s) is always the pilot’s responsibility. And you did read the POH/AFM, and perform a preflight inspection before climbing in, right?
Meanwhile, one reason multi-engine airplanes exist is for the peace of mind the additional powerplant(s) bring to the table. Of course, that only works well when we maintain our skills at flying with one engine inoperative, especially in a high-workload environment. Like when shooting an approach in actual IMC. Losing an engine at such a point can end well if everything else comes together. Here’s an example where it didn’t.
Background
On April 22, 2019, at 0851 Central time, a Beech 58 Baron was substantially damaged when it collided with terrain while executing the RNAV (GPS) approach to Runway 12 at the Kerrville (Texas) Municipal Airport (ERV). The airline transport pilot (male, 65) and five passengers died. Instrument conditions prevailed; the flight operated on an IFR flight plan, departing the West Houston (Texas) Airport (IWS) at about 0730. The straight-line distance between the two airports is 178 NM.
At about 0839, the pilot was cleared for the approach and to maintain 4000 feet MSL to the initial fix. According to ADS-B data, the airplane maintained 3900 feet until about 0844:59, when it began a steady descent. At the time, the airplane was about 13 miles from the runway. Data from the airplane’s engine monitor indicate the left engine lost power about 0845, followed by the right engine about 10 seconds later.
The airplane steadily descended well below the approach profile. About 40 seconds after losing power, the left engine regained nearly full power, which it maintained until the end of recorded data at 0851. The airplane slowed below its minimum controllable airspeed (VMC) of 83 knots as it descended from about 500 to 300 feet AGL. A witness saw the airplane on final approach at a low altitude, when it entered a right turn, began a right spiral and disappeared behind a ridge line.
Investigation
A flight instructor who frequently flew with the pilot and conducted his most recent flight review stated the pilot’s mechanical flying skills were very good but, on occasion, his understanding of technical issues was not as strong. A few times, the pilot did not perform well during unexpected in-flight issues.
Based on passenger weights, the airplane’s takeoff weight was calculated as 5598 lbs. with 50 gallons of usable fuel and 5526 lbs. with 38 gallons of usable fuel. Its maximum gross weight was 5500 lbs. Performance charts indicated a one-engine-inoperative climb capability of about 300 fpm with the inoperative engine’s propeller feathered, flaps up and a gross weight of 5300 lbs. Failing to feather the prop or extending 15 degrees of flaps reduced single-engine rate of climb by 150 and 400 fpm, respectively.
The landing gear was retracted. The left-wing flap actuator corresponded to a 15-degree flap setting; the right-wing flap actuator was fractured. Both electric fuel boost pump switches were in the high position. The throttle, propeller and mixture controls were all near the full-forward position. The left propeller’s blades displayed evidence of rotation at the time of impact. The right propeller was not feathered and was in or near the low-pitch stop position.
The airplane’s fuel system comprised three interconnected fuel cells and one internal tip tank for each wing, for a total capacity of 200 gallons, six of which were unusable. Each wing had two fuel filler caps: one located in each outboard leading-edge fuel cell and one internal tip tank. The airplane’s instrumentation indicated the airplane consumed about 28 gph while at cruise power during the accident flight. According to the NTSB, “high resistances in both fuel quantity transmitters would have caused the cockpit fuel quantity indicators for both wings to read about 5 gallons higher each than the actual fuel present.”
About one gallon of fuel was drained from the left-wing fuel cells. No fuel was observed in the right-wing wet tip tank or the right-wing fuel cells. There was no evidence of fuel blight on the area surrounding the airplane. The left-engine fuel selector was near the ON position, about ¼ toward OFF. The right selector was in the ON position.
The airplane was fully fueled on April 14 and flew on April 15 and 17, after which 30 more gallons were added. Although the pilot kept a fueling log, the NTSB’s investigation found the fuel on board at the beginning of the accident flight was 12 gallons fewer than the pilot’s records indicated.
Probable Cause
The NTSB determined the probable cause(s) of this accident to include: “The pilot’s inadequate preflight fuel planning and fuel management, which resulted in a loss of engine power due to fuel exhaustion. Also causal was the pilot’s failure to follow the one-engine inoperative checklist and maintain the airplane’s minimum controllable airspeed by properly configuring the airplane, which resulted in a loss of airplane controllability.”
According to the NTSB, “multiple errors before takeoff led to a loss of engine power due to fuel exhaustion” and “once the right engine lost power, the pilot failed to properly configure the airplane…and allowed the airspeed to drop below the point at which the airplane could maintain flight.”
Even if the “missing” 12 gallons of fuel had been aboard, the pilot’s flight planning was marginal, and didn’t comply with reserve requirements. Inadequate fuel, coupled with the pilot’s failure to properly manage the engine stoppage, made the flight’s outcome inevitable.
Fuel Mismanagement
One of the nasty little secrets of personal airplanes is the inadequacy of their fuel quantity indicating systems. The only real way to know how much fuel is aboard is to fill the tanks, and even then some quirks can creep in. According to the NTSB’s Safety Alert SA-067, Flying On Empty, “From 2011 to 2015, an average of more than 50 accidents per year occurred due to fuel management issues. Fuel exhaustion accounted for 56% of fuel-related accidents while fuel starvation was responsible for 35% of these accidents.” The NTSB’s recommendations include:
- Know how much fuel you have onboard AT ALL TIMES.
- Visually confirm the fuel quantity in your tanks.
- Know how much fuel you will need for a given flight.
- Know your aircraft’s fuel system and how it works.
- Don’t stretch your available fuel supply. Stop and get gas!
Aircraft Profile: Beechcraft Model 58 Baron
OEM Engines: Continental IO-550-C
Empty Weight: 3443 lbs.
Maximum Gross Takeoff Weight: 5500 lbs.
Typical Cruise Speed: 195 KTAS
Standard Fuel Capacity: 136 gal.
Service Ceiling: 20,688 feet
Range: 1109 NM
VS0: 74 KCAS
This article originally appeared in the April 2022 issue of Aviation Safety magazine.
For more great content like this, subscribe to Aviation Safety!
A $1 calibrated dowel dipstick can save your life.
This dowel method won’t work with a Baron. The Baron does have mechanical sight gauges located in each wing that work quite well.
I knew the accident pilot (Jeff) of this Baron very well. I checked Jeff out in his Aviat Husky, in Houston, a few years before this accident. Jeff was a very successful business man in the Houston area. He loved aviation. His love for aviation was infectious. He would rather fly than work. He started his aviation career later in life and couldn’t get enough flying. He flew every chance he got and many times for fuel (134.5). One thing I can say about this flight is Jeff was concerned about people and would not have wanted to hurt anyone. Yet it happened……
I don’t want to get into this accident and or the pages of other problems associated with this flight. The greater question is why we as pilots continue to make the same mistakes in airplanes that we made 100 years ago. We don’t seem to learn from the past nor the mistakes of others. Certainly more regulations isn’t the answer to accident reduction.
Perhaps part of the solution to reducing accidents rests in this stat. 6% of the pilot population is female, but only 2% of the accidents are caused by females. Almost never fatal.
N/A Night circling approaches to Gillespie, Scud running helos (K. Bryant), Circling at Truckee in a Challenger 604, etc etc. Take a look at N733CD a Cirrus accident that happened in AR two years ago today. I call this accident Gavin’s story. Gavin was only 7 years old and paid the ultimate price. All of these accidents have a common link and we as an industry need to have a serious conversation about the real issue.
God bless.
It seems as we age that simpler fuel systems save lives, as long as we maintain speed, and naturally, it is essential to feather the prop(s) if the engine(s) stall.
If your aircraft has a tank arrangement like a Cub’s fires are almost unheard of in a crash (according to Collins), and very few Cubs crash overall, if we discard flight into IMC, STOL competitions (stall/spin crashes, usually), and overstressed airframes.
While I had my ’46 Navion, each year at annual I’d add or do something to the bird to improve the safety of flight. One of the best upgrades I did was a full-function engine monitor (EDM-930) which included fuel consumption. It took a few flights to get the K-factor set correctly, but when finished it was accurate down to 1/2 gallon in a 60 gallon full load. If flow rate information was available and used correctly, we’d see a lot less of these fuel exhaustion accidents.
Fuel starvation is one of the risks I don’t accept in the 206. Especially since the factory fuel sensors frequently toss up the dreaded “red cross” error on the G1000. So we keep it simple. Every 2 hours we stop to “refresh” and refuel. With the tip tanks we are always good for 1000 mi when the IO540T is babied. Problem solved.
I hear a lot of comments about how fuel gauges in light aircraft are useless. My experience is that this translates to the fact that no effort is made to verify fuel gauge accuracy and fix gauges that are not working properly.
I get it is often impossible to correct fuel gauge inaccuracies in the mid range but the gauge should absolutely read zero or empty when the tank is down to unusable fuel and full when the tank is full. If it doesn’t it is broken and needs to be fixed.
My experience is that if properly set up the gauge will get more accurate at the lower ranges and therefore any low fuel indication needs to be taken seriously.
If you are flying random rental airplanes it can be hard to get a feel for fuel gauge accuracy but if it is your airplane you should know what the gauges indicate vs actual fuel on board across the full indication range. This is a good exercise for the next annual inspection starting with verifying the empty tank reading with the airplane level and adding measured increments of fuel and note the gauge readings.
I also very much endorse the recommendation James C made regarding a fuel computer. My experience mirrors his. A properly set up computer will be extremely accurate
The last paragraph of the article points out that this accident sequence started on the ground at the departure airport. Even if the fuel gauges were perfectly accurate this airplane departed with insufficient fuel to safely and legally completed the planned flight.
Ultimately preventing these kinds of accidents has to focus on the human factors that drove the poor decision making.
Personally I think the problem starts with IFR training. Fuel planning on the average IFR training flight is usually entirely theoretical as the airplane will have much more fuel than required and the practical aspects of real world IFR planning like the effect of ATC reroutes and real vs forecast winds is never really covered