Pelican’s Perch #7: Run That Fuel Tank Dry!

AVweb's John Deakin takes aim at yet another OWT (Old Wives' Tale). While running a fuel tank dry in your recip powered plane may serve to increase your heart rate, John explains why it's not such a bad thing at all, and it is probably a really good idea for most of us. In fact, John explains why it's one of the first things you ought to do with a new plane and how it could save your life someday.


If you’ve been reading any of this stuff, you might have figured out by now that the headlines are written to grab some attention, in the hopes you’ll read on. These “leaders” or “teasers” are written by someone else, I’m too lazy. They may or may not resemble the actual gist of the article, if there even is a gist. So, sue me. Better, sue AVweb, they’ve got more money. [Editor’s note: We do? Where’d they hide it?] If you want really serious, go read William F. Buckley, we’re trying to have a little fun here, poke a few holes in some egos, puncture a few OWTs (Old Wives’ Tales), and with luck, maybe exercise a few brain cells.

Fuel Management

I’d like to take a look at fuel management, and since my method sometimes calls for running a tank dry, let’s get that out of the way first.

I hate OWTs. There are new ones, and there are old ones, and they seem to have a life of their own, with no way to kill them. “Beware the downwind turn,” “never turn into the dead engine,” and “don’t run lean of peak” are all so silly, yet they persist, even among people who should know better. Other stupidities are “The only time you have too much fuel is when you’re on fire, “the three most useless things in aviation are runway behind you, fuel in the truck, and altitude above you” (or variations on that theme), and the endless inane argument over whether pitch controls speed or altitude. I hope to shoot at some of those in future columns.

How about Running a Tank Dry?

“Never run a fuel tank dry” is one of the newer ones, because it used to be a standard method of fuel management. We used to do it on virtually EVERY flight, before the jets came along! I still do it regularly, in non-turbine aircraft, and it works great. What happened to change this?

Now please, folks. I’m not talking about running a tank dry when it’s the only tank in the airplane! Nor when it’s the last tank with fuel in it! Some readers have blasted me on the checklist columns about single-pilot operations, saying “Well, how about the A-26, or the single-pilot Citation?” Folks, there are exceptions to every rule, and if you haven’t got the smarts to figure that out, you not only shouldn’t be reading this column, you probably shouldn’t be flying (or even walking). As I tried to point out in “What Really Counts,” you ought to read this, think about it, figure out what applies to you and what doesn’t, and act accordingly. You don’t like something? Argue about it! Find some data to prove I’m wrong! Please! I want to know more than anyone.

Who Started This Insanity?

In my opinion, the change in attitude was mostly due to turbines (jets and turboprops), in which even a tiny bit of air in the fuel supply may kill the engine. Turbines, as wonderful as they are, can be miserable or impossible to restart in the air, often requiring a descent into “thicker” air, specific airspeeds and windmilling RPM, and several system shave to function properly to get a re-light. The turbine is like a campfire on a rainy day. Keep it going, and it’s fine, but let it go out, it can be really tough to get it going again. I suspect that as many pilots started flying turbines, the idea developed that running a tank dry was “a bad thing,” and this attitude permeated down to the recips, as well.

This is simply not true of recips. When a recip runs out of fuel, nothing else has changed. The spark is still there on every power stroke, the piston is still pumping air, driven by the prop, which is nearly impossible to stop in flight even when you want to. Two of the “three necessities” (fuel, air, spark) remain, totally unaffected by the lack of fuel.

Stopping a Prop

Some pilots seem scared the prop will stop. It won’t, it generally takes a lot of effort to stop the prop, though there is some variation between airplanes.

For example, in order to stop the prop on my Bonanza (IO-550, three bladed prop), I have to kill the mags, or stop the fuel (mixture or fuel valve), pull the prop control all the way back, extend full flaps, and fly very near the stall for many seconds with the throttle closed (makes the engine have to suck a little harder). Once the prop stops, I can fly at about 95 knots without the prop starting to windmill again. It will slowly work its way past the compression stroke, then come up against the next one, and stop again. Anything above 95, it will spin up past that next compression stroke, and the next, and then increase to whatever RPM I’ve set with the prop control (about 1200, on mine).

So, even if the fuel stops, the “suck, squeeze, bang, blow” of the normal internal combustion process continues. All we have to do is introduce fuel again, and the engine simply has no choice, it MUST start producing power again.

Will It Start Again?

Is there a problem restoring the fuel flow? In a word, NO. The FAA certification rules require testing in this area. For example:

FAR 23.955(e) Multiple fuel tanks. For reciprocating engines that are supplied with fuel from more than one tank, if engine power loss becomes apparent due to fuel depletion from the tank selected, it must be possible after switching to any full tank, in level flight, to obtain 75 percent maximum continuous power on that engine in not more than –

(1) 10 seconds for naturally aspirated single engine airplanes;

(2) 20 seconds for turbocharged single engine airplanes, provided that 75 percent maximum continuous naturally aspirated power is regained within 10 seconds; or

(3) 20 seconds for multi-engine airplanes.

FAR 25.951(a) Each fuel system must be constructed and arranged to ensure a flow of fuel at a rate and pressure established for proper engine and auxiliary power unit functioning under each likely operating condition, including any maneuver for which certification is requested and during which the engine or auxiliary power unit is permitted to be in operation.

(b) Each fuel system must be arranged so that any air which is introduced into the system will not result in –

(1) Power interruption for more than 20 seconds for reciprocating engines; or

(2) Flameout for turbine engines.

I don’t have a copy of the old CARs (CAR 3, and 4b) here, but I’m pretty sure that language is lifted right out of those early, common-sense rules.

I know of no accidents that have occurred because an engine would not restart when supplied with fuel in flight. I have personally done this literally thousands of times myself, and never seen more than a couple of seconds of interruption, even when I was completely unaware the engine was about to quit. If we count all the people I know who routinely did it, there are literally millions of such events.

The Passenger Factor

There is also the perception that running a tank dry scares passengers. Maybe, but I don’t think so. I’ve done it many times with all types of passengers watching, and if it’s treated matter-of-factly, with a few words in advance, a chuckle when it does happen, none have ever expressed anything but curiosity. In other words, no big deal. Some of the less-aware don’t even realize anything happened. A frantic scramble to turn the fuel valve, swearing, praying, screaming, or sweating does not induce confidence, however.

The Creeping Crud OWT

Then there is the secondary OWT that says something like “But what if some crud gets sucked into the system from the tank bottom?” Give me a break! Think about this, for a moment. There are three areas where “crud” might be a concern. Crud lying on the bottom, crud suspended in the fuel, and crud floating on the surface. When we fuel the airplane, fuel is injected rather violently, stirring up the whole tank. When we fly in turbulence, fuel sloshes rather violently around the tank. Do you really think anything will be peacefully lying on the bottom, year after year? If it were, why would running the tank dry stir it up, and if it’s that tenacious, how on earth is running the tank dry going to magically pick it up?

How about suspended crud? It is no more, or less likely to be sucked into the fuel lines at any fuel level. Floating crud, on the surface? Well, maybe, but can you name me something that will do that? And if there is, well, how much of it will you allow, before you rip the tank out for “cleaning,” or how WILL you get it out, someday? Just how, and when, will you identify it, detect it, and get rid of it? And, how much fuel do you want as a “buffer” below the floating crud, to keep from sucking it in? In fact, if there is a little something floating on the surface, I WANT it to be sucked into the fuel lines, preferably a little at a time, so that the strainers and filters can catch it, and alert me that something is going on in there. The likelihood of there being enough to cause a problem is remote, at best, and if running a tank dry will pick up a little crud, then running a tank dry often is a very good thing, because you’ll catch it a little at a time, and drain it out the strainer.

Of course, if you keep the tank full most of the time, and the cap on, and drain the sumps often, there isn’t any way for crud to get into, or stay in the tank in the first place.

Shock Cooling?

I’m beginning to think this business of “shock cooling” is perhaps the worst of the “modern” OWTs around, but it’s one that sounds very reasonable at first glance. I’m still not utterly certain myself. But consider, which airplanes get the longest TBO? Why, it’s TRAINERS! They CONSTANTLY jam the power on, and yank it off, many times an hour, all day, every day. No regard or thought of temperature control, at all. If shock cooling is so bad, why do they last so long? Does anyone know how many hours Bob Hoover gets from his Shrike engines? They go from full power to a full stop (feathered) with NO interval, and right back to full power again. I’ve tried to get to him to ask, but he’s always mobbed. Does anyone know?

Now please, I’m not advocating abuse of any precision machinery, here! I prefer gentle engine management, it’s just good airmanship. But “shock cooling” maybe a non-issue.

In any event, if properly done, running a tank dry does NOT cause “shock cooling” at all! Once the pilot knows what he is doing, he should know precisely when the tank will run dry, and will detect the early signs well before the engine is affected (fuel pressure, mostly). Even after the engine begins to falter, there is time to change tanks and get things back to normal before the CHTs drop. We now know this, with the modern digital engine monitors. The EGTs do drop a couple of units, but they recover very quickly, so even turbocharged engines shouldn’t have a problem. I would be more careful with a turbo, however, because other issues arise, like surging, bootstrapping, etc.

The first time any pilot runs a tank dry, the heart rate does increase, but it recovers quickly, too! As experience is gained, it becomes a non-event.

Broken Fuel Selectors

Okay, this one may be a slightly more valid concern. If you are concerned about this, you need to have your mechanic check the fuel tank selector out thoroughly, in order to make sure it moves freely, and that it is not worn. Once this is done, the possibility of it breaking becomes so remote it’s hardly worth considering. If you want to consider it anyway, carry a small pair of pliers, and think of glider pilots. EVERY landing in a glider is a deadstick landing! If you are not utterly confident of your ability to deadstick your airplane into any reasonable airport, allow me to point out the fact that you need some training, or practice. Then you can plan to run that tank dry when within gliding range of an airport. In most parts of this country, it is fairly unusual to be too far from any airports at all. Do I need to point out that if you don’t know the glide capability of your aircraft, it’s time to hit the books, and confirm that with some training? Most GA aircraft can do 10:1, so for each thousand feet above the airport you’re hoping to reach, you can be almost two nautical miles away (give me a little room here, this is not an exact science).

Just How Much Fuel Is Safe?

The FARs require a minimum of 45 minutes fuel remaining at night, 30 minutes in the daytime, and 20 minutes for helicopters. I think most of us can agree those are NOT conservative figures! Only under very unusual circumstances will I approach those limits deliberately. Using the techniques I am about to discuss, I have pushed to those limits, when approaching my home airport (Arlington, Wash.). But only in very good weather, maintaining a high enough altitude to deadstick it into that airport, with good airports within gliding range along the way, and knowing there are several runways available, including the grass alongside the runway. Take away any of those, and my personal minimum fuel starts going up.

You may never want to push this. You need to make up your mind before flying just what your personal “minimum fuel” is, under ideal conditions, and how much should be added for less than ideal conditions. These should be positive calculations, based on reality, not “Grandmother Fuel,” or “Feel Good” fuel. Additions should be based on hard data, not “feel.”

I frequently hear “Aw, I’ve got a six-hour airplane, but only a two-hour bladder, I don’t really care about fuel.” That’s a cop-out, and extremely unprofessional. What about the case where you arrive somewhere and there is no fuel available? Just HOW MUCH fuel DID you have when you took off? Do you KNOW? HOW MUCH fuel did you use during the flight (or flights)? Do you KNOW? All this leads to the all-important question, HOW MUCH do you have left, and how far can you go with it? If you cannot reliably answer those questions within about 3% of the fuel capacity of your airplane, you’re not doing your job as a pilot.

How Many Ways to Cheat?

Then there is the Bonanza pilot who says “Well, I’ve got two 40-gallon tanks, with 37 usable each, so I’ll call that 35 (70 total), to be safe. I burn about 12 gph, so I’ll call that 15, to be safe. It probably takes 5 more gallons to taxi out, takeoff, runup and climb, so I’ll call it 10, to be safe.” That brings our “safe” aviator down to 4 hours of total fuel for flight! Now he starts in again, saying “I don’t think an hour’s reserve is enough, so I’ll add one, leaving me two hours of planning fuel, or 320 NM in still air.” Now he can’t find a fuel stop right at 320 NM, so he looks for anything short of that, how far it’ll be is anyone’s guess. Finally, he’ll go out and fly it, and if he gets a hold, or a diversion, he starts sweating, because he knows he planned “minimum fuel!” Is he safe? Well, yeah, but he’s not making effective use of his tools, and he isn’t very professional.

So, How Much Fuel Do I Have?

I have deliberately used the numbers above, because they apply to my airplane, or so it might seem. The POH, and the fuel tank placards for my Bonanza all state there is only 37 usable gallons per side, of 40 total. The very first long flight I made, I ran the left tank an hour, then ran the right tank dry, “maneuvering” while burning the last few gallons. With an aggressive slip to the right, I could unport the tank, and get a fuel pressure fluctuation, and this is probably the reason for the “37 usable.” I ran it dry, and when I filled it, I found it actually took 41.2 gallons. The next flight, I reversed the pattern, and found the left tank was identical. Does this meet FAA standards for “usable”fuel? No, but the fact remains that I can reliably get 41.2 gallons out of either tank in cruise, where unporting is no issue at all, and if there no turbulence, or low-altitude maneuvering, the other tank is fully usable, too. Of course, I can’t count on “no turbulence,” so some allowance for this is needed.

At the very least I have 78.0 gallons available (41.2 + 37.0) up to 82.4 (41.2 x 2).

Note this cannot be determined without running a tank dry, in flight. I consider it so important to KNOW this figure, that if I regularly flew a single-tank airplane, I would probably fly it to fuel exhaustion over a quiet airport, and deadstick it in, like a glider, then tow it to the gas pump. I know, the FAA probably wouldn’t like that! Draining the fuel won’t be accurate enough, some tanks drain above or below the fuel outlet. Even disconnecting the fuel line to drain it isn’t totally accurate, because you can’t maneuver the airplane in the hangar. Better than doing nothing at all, of course.

Exactly How Much Do We Burn?

On those same early flights, I made sure that for the entire time I was running on the tank I was testing, I ran at a constant power, and constant altitude, and constant fuel flow. By timing that from “full” to “dry tanks,” I knew EXACTLY how accurate my fuel flow indicator was. By noting the fuel quantity indicators during the burn, and noting the time at which they indicated “3/4,” “1/2,”and “1/4,” I calibrated the fuel quantity indicators, too. While I’m lucky to have dead accurate instrumentation, it doesn’t really matter, as long as you know what the instruments are really telling you, at any given reading. Make some notes, so that you can use the data later.

On subsequent flights of more than an hour, but less than about three, I used just one tank for start, taxi, runup, takeoff and climb, and for exactly one hour from liftoff, then switched to the full tank for the remainder of the flight. By noting the fuel used from that “takeoff tank,” I now know with great accuracy that all that requires only about 4 gallons additional fuel, over the normal hourly rate.

At normal altitudes and power settings, I find 12.7 gph to be a reasonable figure. Now, let’s look at those numbers, again. Use a conservative 78.0 gallons. That comes to 6.17 hours of usable fuel. NOW take off your “never use” fuel, perhaps an hour’s worth, leaving 5.17 hours for planning, including fuel for alternates. At 160 knots (140 the first hour), that’s 787 NM in still air, more than double that for our “nervous nellie” above.

All these tests were duplicated at different power settings and altitudes, during normal flights, over the course of the first few months I had the airplane. In fact, I continue to do them, as a systems check. For example,there is no guarantee that a bladder tank will always hold the same fuel. Perhaps it will partially collapse, and take less, leaving you less fuel than you think. If I run the left tank dry today, and it takes 41.2 gallons, I know it’s still okay, and as a side benefit, I’m also checking the accuracy of the FBO’s fuel pump! Next week, I’ll reverse my usage pattern, and check the other tank.

How to Use All This

Normal power settings in my airplane will range from about 9 gph to 15 gph, depending on altitude, and whether I want speed or non-stop distance. But a very common figure keeps popping up, flight after flight, about 12.7 gph, full throttle, 2500 RPM, leaned for “best power” somewhere between 10,000′ and 15,000′. Lower altitudes will be well lean of peak EGT, still full throttle and 2500.

I take off, and shortly after takeoff, I set a timer to alert me at exactly the one-hour point. When that timer beeps, I KNOW I’ve used 12.7 + 4 gallons, and thus have 24.5 gallons remaining in that tank. I know my fuel quantity should show about 2/3 of that “takeoff tank” remaining.

I’m now on the full tank, and if I’m burning 12.7 gph, I know it will take 3+14 to run that tank dry. I set the timer to 3+10 or 3+12, so it will warn me a few minutes before that happens. 48 minutes later (at 1:48 into the flight), that fuel quantity indicator should show “3/4,” at 2:36 into the flight it should show “1/2,” and at 3:24 elapsed, it should show “1/4.” When the timer dings, I have a couple minutes to keep a wary eye on the fuel pressure. When it flickers, I switch back to “The Takeoff Tank.” I have re-validated my instrumentation, my actual fuel capacity and burn, and most important, I know precisely how much fuel remains, 24.5 gallons, or 1+55 to dry tanks. For a final check, when I pull up to the pump, I run the final calculation, and “predict” how much fuel the tanks will take. In several years, I’ve never missed it by more than about 2 gallons on the “safe side,” which I consider an acceptable error.

Know Thine Own Airplane

Some airplanes lend themselves to this sort of thing better than others. My Bonanza is nearly ideal for this, having two tanks that can be filled quickly to the top with great accuracy, and with the vapor vent return line returning to the tank being used. Some older Bonanzas have all the vapor vent return going back to just the left tank, which makes all this a bit more difficult, or a bit less accurate. High-wing Cessnas are particularly difficult, because it’s very hard to know with certainty how much fuel is aboard, that last few gallons goes in very slowly, if at all, and there is often a cross-transfer that happens in flight. Still, these procedures can be modified and used to determine the numbers for almost any airplane. Test flights are seldom necessary, you just need to log the data on normal flights, and within a month or three, you’ll have good, useful data that may just save your life, someday.

Be careful up there!