Dogfight: Run the tank dry

"Running on empty, momentarily" vs. "Keep the engine(s) running"

March 1, 2011

Editor at Large Tom Horne and Senior Editor Dave Hirschman have a lot of things in common: lots of ratings, lots of experience in lots of airplane models—and lots of opinions (as well as similar haircuts). We last turned them loose on the topic of pattern entry (see “Dogfight: Pattern Entry,” January 2011 AOPA Pilot) and the response to two different schools of thought on this topic garnered some of our most interesting musings—and the largest amount of letters—from a large number of readers ( see “Letters: Dogfight,”). So we launch them into another “Dogfight”—we hope you’ll enjoy these two takes on a topic, and keep those cards and letters coming. —Ed.

Running on empty, momentarily

There are times when it's a smart move

Editor at Large Thomas A. HorneBy Thomas A. Horne >

As a general rule, it’s best to steer well clear of running a fuel tank dry. There are good reasons for this. Air drawn into the engine(s) may prolong a quick restart, for example. In addition, sediment or water that may be resting in the fuel-tank sumps can be sucked into the fuel system plumbing, causing rough engine operations. And yes, there’s the anxiety factor involved with any loss of power. But there are times when running a tank dry can be justified.

Let’s say you are flying a long leg over the desert Southwest. Fuel availability is scarce over terrain like this, and when your exact fuel consumption rate is an unknown—even with your best efforts at flight planning—then running a tank dry may be the best way to confirm your fuel status and thus be able to plan a timely diversion, or perform a 180-degree turn.

It usually takes only a few seconds to restore fuel flow after a tank has run dry.

Of course, much depends on your airplane’s performance and equipment. In an airplane capable of flying 600 nm, there’s no reason to run a tank dry on a trip of, say, 300 nm. But if the same airplane was to attempt a 500-nm leg, then it may be advisable. Unforeseen strong headwinds—but not so strong that the airplane’s endurance is challenged beyond legal limits—may call for knowing exactly how much fuel remains aboard.

Even ferry pilots with ferry fuel pay extremely close attention to fuel quantity. On several long over-water legs in Cessna 172s, I had the standard fuel aboard, plus a 50-gallon tank in the backseat, plus another 15 gallons in the baggage compartment. Transparent fuel lines traveled from the tanks, down the center of the airplane (where I could view the flowing fuel) and through the firewall to the engine. After leveling off after takeoff, the drill was to switch to the aft tank—and watch the fuel line. When that first slug of air made its way backwards up the fuel line, I knew that the engine wouldn’t be running for long. As soon as the engine began to surge from fuel starvation I noted the elapsed time for the tank. That gave me an exact fuel-consumption rate.

It usually takes only a few seconds to restore fuel flow after a tank has run dry. You can switch to a fuller tank and hit the fuel boost pump in a remarkably short period of time. True, fuel-injected engines may continue to surge for a bit, but it usually takes just 10 to 20 seconds or so for the engine to resume smooth operation. This is not a time for the faint of constitution, so running a tank dry is something you’d want to avoid with passengers. I prefer to do it when I’m flying solo.

That procedure I described involved an airplane without a fuel-flow gauge, engine analyzer, or accurate fuel-measurement system. Pilots with this equipment have an advantage—and probably won’t ever have to run a tank dry in order to know their exact fuel condition.

Most GA air-craft don’t have fancy aftermarket fuel-management gear. Their only fuel gauges are the notoriously unreliable fuel quantity indicators that look as though they’d be more at home on a lawn tractor than an airplane. Factor in potential inaccuracies in manifold pressure, rpm, and original-equipment fuel-flow indicators, and it’s easy to see how actual fuel burns—and ranges and endurances—can vary widely from those published in the airplane flight manual. Running a tank dry may be the only way to get a handle on what’s really happening under the cowl.

Run a tank dry? Plan to avoid it. But understand when it may make sense. Better to know a tank’s running dry than to be surprised by complete fuel starvation.

E-mail the author at

Keep the engine(s) running

Why scare your passengers?

Senior Editor Dave Hirschman< By Dave Hirschman

There are few better ways to terrify passengers than shutting down an engine in flight. Even if you warn them in advance and explain what’s going to happen and why, an engine suddenly going silent is enough to make a formerly calm companion jump out of his or her skin—and that’s especially true when the airplane you’re flying has only one engine.

In the old days—before digital fuel computers—running tanks dry was a standard practice because it let pilots know exactly the rate at which their airplanes were consuming fuel, and how much endurance they could expect from a known quantity of fuel. That may still be necessary on some long-distance ferry flights where pilots are seeking maximum range, but running tanks dry is rarely called for under normal conditions. (And if it is, you may be cutting your margins too thin.)

Even when flying solo, running a tank dry can cause unwelcome complications. On an IFR flight plan, for example, a sudden absence of engine power and a restart can quickly put the airplane significantly off an assigned speed, heading, or altitude. Leaving an airplane on autopilot (with altitude hold) could lead to a stall if the restart takes longer than anticipated, and autopilot-commanded stalls are a very bad idea.

An engine going silent is enough to make a formerly calm companion jump out of his or her skin.

Digital fuel computers—especially those with pre-set timers and low-fuel warnings—allow pilots to reliably burn down to the last few minutes of fuel remaining in a tank before switching. Pilots flying airplanes without that handy technology can still calculate their fuel flow and endurance with great precision by using constant power settings and keeping close tabs on actual fuel consumption on all trips.

It’s true that piston aircraft engines usually restart almost instantly as soon as the fuel supply is restored, and a momentary loss of engine power is highly unlikely to cause any harm. Props continue spinning at relatively high rpm, even on high-compression engines, and moving the fuel selector (and sometimes a boost pump) is the only pilot action necessary to restore power.

But repositioning the fuel selector—even when it’s done correctly—is no guarantee that the engine will come back to life, and there have been many fuel starvation accidents in airplanes with lots of fuel on board. If there’s going to be a fuel-flow problem on one tank, it would be nice to know it while there’s still fuel in the other(s). Whenever possible, switch tanks over an airport—just in case.

I confess to trying to have it both ways on this issue by waiting until the last moment to switch tanks. I’ll be flying along with a hand on the fuel selector, staring at the fuel-pressure gauge and waiting for the first flicker that shows the pump is running dry. But invariably, ATC calls with a frequency change, or a new routing, and I look away at the precise moment that the tank runs dry. Better to just switch tanks at an appropriate time and get it over with.

Airline managers track aviation fuel consumption closely, and they hate carrying a drop more than is absolutely necessary—because it requires fuel to carry fuel. If there was an advantage to running tanks dry, you can be sure they’d do it. The fact that they prefer to keep engines running makes this one of those rare times I heartily agree with them.

An uninterrupted supply of fuel keeps engine rpm up, and pilot heart rates down. Save your glider flying for the glider port. It’s seldom necessary to turn powered airplanes into gliders—even if only for a few (very long) seconds at a time.

E-mail the author at