Running a tank dry in flight

[gsxrpilot]

I've intentionally run tanks dry at altitude in both a 1964 M20C and a much later M20K 252. It was a non-event in both. The K does take a little longer to relight in the flight levels, but again, the high-boost pump is there to ensure it does come back even in the thin air.

I'll maintain that it's much safer to know where my fuel is and how much there is, than to be low in BOTH tanks at the end of a long flight. Of course this can be mitigated by always landing with full reserves in both tanks. But that certainly and needlessly reduces the rang of these cross country machines.

[salty]

I don't understand the fear of running a tank dry. How is it any different than switching without a dry tank. I'm scared to even switch, but I recognize it's irrational. 

What could happen by waiting until the line is empty before switching that can't happen if you do it sooner?

if there is crap in your tank it's going into the line long before the tank is empty. I don't get the issue. 

Edited March 28, 2017 by salty

[ArtVandelay]

I don't understand the fear of running a tank dry. How is it any different than switching without a dry tank. I'm scared to even switch, but I recognize it's irrational. 
What could happen by waiting until the line is empty before switching that can't happen if you do it sooner?
if there is crap in your tank it's going into the line long before the tank is empty. I don't get the issue. 

I thought it was from the difficulty of hot starts. I have never done it but wouldn't want do it at lower altitudes.

[salty]

What would stop an NA engine from starting right up when you give it fuel again?

i haven't run a tank dry yet, and probably never will because I don't know that I can sit in the plane that long anyway, but my plan has been to run one tank for an hour, then run the other dry. I have no FF monitor, so I'd be pretty likely to let the engine die before I got the tank switched. I've not yet flown more than 36 gallons worth in one setting yet, so I've just switched to fullest tank for landing long before I'm empty. 

[Shadrach]

3 hours ago, gsxrpilot said:

I've intentionally run tanks dry at altitude in both a 1964 M20C and a much later M20K 252. It was a non-event in both. The K does take a little longer to relight in the flight levels, but again, the high-boost pump is there to ensure it does come back even in the thin air.

I'll maintain that it's much safer to know where my fuel is and how much there is, than to be low in BOTH tanks at the end of a long flight. Of course this can be mitigated by always landing with full reserves in both tanks. But that certainly and needlessly reduces the rang of these cross country machines.

Why do you need high boost?  I would think excess fuel would be the bigger inhibitor to restart once the turbo spools down.

[gsxrpilot]

3 minutes ago, Shadrach said:

Why do you need high boost?  I would think excess fuel would be the bigger inhibitor to restart once the turbo spools down.

I'm not 100% sure of this answer, but I know with a fuel injected engine the lines have to be pressurized. When I've run my 252 dry in the flight levels, it usually takes about 5 to 8 seconds of boost for it to start. I believe the POH says it could take 12 to 14 seconds to start. I'm certainly not worried about it. Up that high I've got plenty of options if it were to fail to restart. But I really can't imagine how it would fail. 

[Shadrach]

19 minutes ago, salty said:

What would stop an NA engine from starting right up when you give it fuel again?

i haven't run a tank dry yet, and probably never will because I don't know that I can sit in the plane that long anyway, but my plan has been to run one tank for an hour, then run the other dry. I have no FF monitor, so I'd be pretty likely to let the engine die before I got the tank switched. I've not yet flown more than 36 gallons worth in one setting yet, so I've just switched to fullest tank for landing long before I'm empty. 

Nothing in my experience.  I try to catch the fuel pressure drop prior to the engine quitting, but if I miss it it's no big deal. Good idea to pull the throttle back a bit to hedge against an engine over speed event. Other than that, the engine easily restarts and always has.  I switch tanks upon reaching cruise altitude and then every hour.

[Shadrach]

Just now, gsxrpilot said:

I'm not 100% sure of this answer, but I know with a fuel injected engine the lines have to be pressurized. When I've run my 252 dry in the flight levels, it usually takes about 5 to 8 seconds of boost for it to start. I believe the POH says it could take 12 to 14 seconds to start. I'm certainly not worried about it. Up that high I've got plenty of options if it were to fail to restart. But I really can't imagine how it would fail. 

I'm not very familiar with the Continental fuel system. I suspect the need for th boost pump has to do with a drop in upper deck pressure. Do you remember what happened to MP while the engine was not running?

[gsxrpilot]

I prefer to switch tanks based on fuel burned rather than time like @201er does. After all, time has nothing to do with when the fuel runs out .

With 36 gal per side, I usually burn 16 gal out of the right tank, then switch and run the left dry for the full 36 gal. Then switch back to the right where I know I have 20 gal. If I'm close to landing before I run dry, I'll switch to make sure I'm flying the pattern, shooting the approach, etc. on the fullest tank.

[gsxrpilot]

9 minutes ago, Shadrach said:

I'm not very familiar with the Continental fuel system. I suspect the need for th boost pump has to do with a drop in upper deck pressure. Do you remember what happened to MP while the engine was not running?

I don't, I should pay attention to that next time.

[201er]

56 minutes ago, gsxrpilot said:I prefer to switch tanks based on fuel burned rather than time like @201er does. After all, time has nothing to do with when the fuel runs out .

Who said I switch based on time?? I do it by fuel flow every 20 gallons until I have 20 left on one side, then I burn out the other and then milk the rest of the 20. It is very rare that I'm tapping into that last 20 but I make preparations along the way so that I know where it all is if I need it.

[gsxrpilot]

4 minutes ago, 201er said:

Who said I switch based on time?? I do it by fuel flow every 20 gallons until I have 20 left on one side, then I burn out the other and then milk the rest of the 20. It is very rare that I'm tapping into that last 20 but I make preparations along the way so that I know where it all is if I need it.

Sorry, the comma was in the wrong place. I meant that I switch based on gallons just like you do.

[fuellevel]

As a student of this issue and having read through most if not all fuel starvation events,  I believe the author is looking for a planned example of running the tank dry

I will have to reference specifics for the author, but as we also do automated systems for switching tanks so we have looked a failure modes and effects 

True a "new" piston aircraft was demonstrated to be restarted in the Type Certificate program - For CAR 3 as well.   

There are several failure modes or potential failure to consider.   While diaphragm pumps are typically reliable exposing them to a condition where the damping effect of pumping fuel vs air introduces a potential for failure.   In an automated system we run the boost pump prior to initiating a switch with a monitor on fuel pressure for this case.  Most POH recommend the same in a manual mode. 

Restart assumes straight and level slightly descending flight due to loss of power and the use of potential energy.  

Running a tank down to the screen level assumes a clean filter screen - something typically found on a conformed aircraft test article,  and running a tank dry consistently insures you have un restricted flow - i believe this was pointed out.  But it is a failure mode potential and it can be combined.

New aircraft use a substantial header tank to prevent anomaly issues related to uncoordinated flight.

Most Cessna pilots using BOTH fuel selection can change tanks with the application of rudder pressure   - coordinated flight or an aircraft out of trim matters. 

For several reason and because we can - we leave a gallon of usable or so in the wing so combined with unusable -

The concept being that a finger error where a known "empty" tank is for whatever reason manually selected does not interrupt flow to the engine 

If necessary most aircraft will draw unusable fuel - so you can maximize range with available fuel - you can also use it as a buffer in a failure condition. 

But the professional pilot question is why -

Do we in normal flying investigate the structural limits of the aircraft - because it met a Type Certificate condition 

Or do we flare early and drop the aircraft on the runway because the aircraft gear was tested to do so 

More importantly do professional pilots plan to use unusable fuel in a cross country flight - or is it just a buffer for the unknown. 

Typically where running a tank dry becomes an issue - is when it happens inadvertently 

What is frustrating is the lack of detail on fuel starvation / exhaustion events in the USA - therese are typically FSDO inspections with NTSB oversight.  There is little or typically no investigation of the aircraft in these accident events.  They are typically rubber-stamped pilot error.   So using US data it is difficult to prove the original poster wrong - because there isn't the investigative rigor applied.   We don't know.  There are several interviews with accident pilots where I have asked if the aircraft was investigated - and the answer was no and they were told universally it would be on their dime if they wanted to see what might have caused the issue.  

However - this is not the case worldwide 

http://www.havarikommissionen.dk/images/Bibliotek_luftfart/General_Aviation/Fuel.pdf

I do have references for why not  - and for examples where running a tank dry wasn't a good solution - i will try to find them 

[salty]

22 minutes ago, fuellevel said:

As a student of this issue and having read through most if not all fuel starvation events,  I believe the author is looking for a planned example of running the tank dry

I will have to reference specifics for the author, but as we also do automated systems for switching tanks so we have looked a failure modes and effects 

True a "new" piston aircraft was demonstrated to be restarted in the Type Certificate program - For CAR 3 as well.   

There are several failure modes or potential failure to consider.   While diaphragm pumps are typically reliable exposing them to a condition where the damping effect of pumping fuel vs air introduces a potential for failure.   In an automated system we run the boost pump prior to initiating a switch with a monitor on fuel pressure for this case.  Most POH recommend the same in a manual mode. 

Restart assumes straight and level slightly descending flight due to loss of power and the use of potential energy.  

Running a tank down to the screen level assumes a clean filter screen - something typically found on a conformed aircraft test article,  and running a tank dry consistently insures you have un restricted flow - i believe this was pointed out.  But it is a failure mode potential and it can be combined.

New aircraft use a substantial header tank to prevent anomaly issues related to uncoordinated flight.

Most Cessna pilots using BOTH fuel selection can change tanks with the application of rudder pressure   - coordinated flight or an aircraft out of trim matters. 

For several reason and because we can - we leave a gallon of usable or so in the wing so combined with unusable -

The concept being that a finger error where a known "empty" tank is for whatever reason manually selected does not interrupt flow to the engine 

If necessary most aircraft will draw unusable fuel - so you can maximize range with available fuel - you can also use it as a buffer in a failure condition. 

But the professional pilot question is why -

Do we in normal flying investigate the structural limits of the aircraft - because it met a Type Certificate condition 

Or do we flare early and drop the aircraft on the runway because the aircraft gear was tested to do so 

More importantly do professional pilots plan to use unusable fuel in a cross country flight - or is it just a buffer for the unknown. 

Typically where running a tank dry becomes an issue - is when it happens inadvertently 

What is frustrating is the lack of detail on fuel starvation / exhaustion events in the USA - therese are typically FSDO inspections with NTSB oversight.  There is little or typically no investigation of the aircraft in these accident events.  They are typically rubber-stamped pilot error.   So using US data it is difficult to prove the original poster wrong - because there isn't the investigative rigor applied.   We don't know.  There are several interviews with accident pilots where I have asked if the aircraft was investigated - and the answer was no and they were told universally it would be on their dime if they wanted to see what might have caused the issue.  

However - this is not the case worldwide 

http://www.havarikommissionen.dk/images/Bibliotek_luftfart/General_Aviation/Fuel.pdf

I do have references for why not  - and for examples where running a tank dry wasn't a good solution - i will try to find them 

So far, the only answer to my original question  "What could go wrong by letting the tank go dry that wouldn't already have gone wrong when wet?" has been the fear of hot start. And I don't understand how that could be an issue with a NA engine.

I'm still trying to understand what people think could happen. Crap in the screen is crap in the screen regardless of fuel level. Wouldn't it have gotten into the fuel line or clogged the line when fuel was still present? Why would it wait until there's no fuel to suddenly be a problem, and what problem would it cause?

[fuellevel]

Transport Canada  Aviation Investigation Report A11Q0136

1.18.1.3  Fuel Management

To prolong cruise flight through fuel management, section 7 of the AOM recommends using the fuel selector valve to alternate between the left and right tanks in order to completely empty one tank before selecting the other. Section 7 also indicates that if the pilot wishes to completely empty a tank in flight, the auxiliary electric fuel pump may be required to help restart the engine. In this situation, it is recommended the pilot check that the auxiliary electric fuel pump is functioning properly before emptying a tank, by briefly activating the pump and checking the indicator for a slight increase in fuel flow.

The Pilot Safety and Warning Supplement, published by Cessna and amended in March 1998, contains additional information that, according to Cessna, acts as an important complement to the AOM and provides good safety practices. However, the Pilot Safety and Warning Supplement does not replace the aircraft’s certified AOM.

Chapter 6 of the Pilot Safety and Warning Supplement describes how improper fuel management often causes airplane accidents. It also mentions that pilots often employ improper fuel management techniques. The supplement recommends that pilots become completely familiar with the fuel supply system and fuel tank switching procedures. It emphasizes that running a fuel tank dry as a routine procedure is an unsafe technique, although there are exceptions. Any sediment or water not drained from the fuel tank could be drawn into the fuel system and cause erratic operation or even power loss.

[salty]

3 minutes ago, fuellevel said:

Transport Canada  Aviation Investigation Report A11Q0136

1.18.1.3  Fuel Management

To prolong cruise flight through fuel management, section 7 of the AOM recommends using the fuel selector valve to alternate between the left and right tanks in order to completely empty one tank before selecting the other. Section 7 also indicates that if the pilot wishes to completely empty a tank in flight, the auxiliary electric fuel pump may be required to help restart the engine. In this situation, it is recommended the pilot check that the auxiliary electric fuel pump is functioning properly before emptying a tank, by briefly activating the pump and checking the indicator for a slight increase in fuel flow.

The Pilot Safety and Warning Supplement, published by Cessna and amended in March 1998, contains additional information that, according to Cessna, acts as an important complement to the AOM and provides good safety practices. However, the Pilot Safety and Warning Supplement does not replace the aircraft’s certified AOM.

Chapter 6 of the Pilot Safety and Warning Supplement describes how improper fuel management often causes airplane accidents. It also mentions that pilots often employ improper fuel management techniques. The supplement recommends that pilots become completely familiar with the fuel supply system and fuel tank switching procedures. It emphasizes that running a fuel tank dry as a routine procedure is an unsafe technique, although there are exceptions. Any sediment or water not drained from the fuel tank could be drawn into the fuel system and cause erratic operation or even power loss.

Interesting. But both water and sediment are heavier than fuel. 

[gsxrpilot]

The problem here is that some of us have been to Ada and have taken the APS class where everything is based on data and Old Pilots Tales are summarily debunked. On balance it's safer to know how much useable fuel you actually have and not just the number written in the POH. And the only way to know that for sure is by running the tank dry while in cruise. It's good enough for me, and has worked very well over the last several years.

[Hank]

I've had two inadvertent dry tanks:

• First time, I was distracted and forgot to change; it coughed once, ran fine for a while, coughed again and sputtered while I franticly scanned the panel, as I was descending and about 2500 agl over the bumpy part of central WV. Saw the clock hands 20 minutes past each other, threw the selector and had a smooth landing 15-20 nm away.
• Second time, my fuel stop had pump problems and couldn't deliver fuel. I was sure I had plenty for the trip, but wanted it all in one tank for descent and landing. I was watching the clock, the fuel gage and the fuel pressure. Reached my last-figured change point, thought I could make the extra 3 minutes to Top of Descent. Nope. No cough, no sputter, or fluctuations of any kind, just silence and a significant pitch down at 9500 msl that my wife didn't appreciate. So I threw the selector and it restarted almost immediately. Filled up with 40 gallons, meaning I landed with 12 gallons in the one tank, good for 1:20 more flight time.
• Last weekend, I thought I had magneto issues in cruise (9500 in clear but darkening skies). So I tested it, turned the switch one click left and it quit--silence and that pitch down again. Turned it one click back to Both and she was immediately running again. Turned it two clicks left, she kept running, and I left it there.

Three restarts, two at 9500 msl, the third one maybe half or a little less. Nothing to it, no delays, no fuel pumps, no drama.

Then again, I don't panic about changing tanks anyway. It's always nice to do within a reasonable distance of an airport, if there's one around, but I wouldn't deviate or wait to reach one; I also lean down and turn the selector (fairly rapidly) without turning on the electric fuel pump. My Owners Manual doesn't say anything about that, and recommends running the first tank for an hour, then running the second tank dry as a regular practice. No thanks, I'll keep changing hourly.

[carusoam]

See if I can collect a few ideas....

1) the purpose of draining all the fuel out of one tank, leaves the most fuel for the last leg of the flight in one known place...

2) Some people change fuel at the TOC, because it aids in calculating fuel used and where it is.  It makes a lot of sense when you don't have a FF gauge with a totalizer.

3) Some people use the FF/Totalizer to switch tanks to balance the weight on the wings.  This works with a lot of precision.  With a good wing leveler you won't easily notice how good your wing balance really is... switching every hour or so is a compromise of simplicity and balance accuracy.

4) Some fear comes from restarting issues...

5) Restarting a turbocharged engine at altitude can lead to drama if the Turbo stops compressing air.  There is a known altitude for restarting in the event the compressor can't be brought back to life.  Turbo normalizers are better for this issue.  Broken turbos, and intake system tubes have a tendency of falling off of experimental/home built set-ups.  Not usually a Mooney issue.

6) Usual ground start fears don't apply at altitude.  We are not using the starter.  We are not running down the battery... hot starts don't stay hot for very long with air running through the cowl at 90+ knots...

7) One thing that does happen when we run a tank dry is we draw air into the fuel lines which can be problematic.

8) Running the aux fuel pump continues the fuel flow so that air bubbles are a non-issue and will pass through the fuel lines.

9) dirt and water in the tanks are resident bellow the fuel pick-up.  Gravity usually keeps it away from the pick-up.  A filter on the pick-up usually keeps the dirt out.

10) When the gas has all but drained completely, anything like water left in the tank has more of a chance of sloshing to the right place.  The tank may be quite a high percentage of water to fuel at this point.  Keep in mid the lowest point of the tank is still lower than the fuel drain.

11) like air in the lines, water in the lines probably won't last very long either, with the aux fuel pump running...

12) as for running the Hi pressure fuel pump option... This is usually for when the mechanical fuel pump is suspected of not working.  Check your POH to see what it says to use and when...

13) it also helps to know where your fuel pick-up is located in the tank.

14) Also know that when you point the nose down in descent, fuel runs away from the pick-up....

15) If you have the old analog float fuel level gauges that are in crummy condition...expect that you won't know how much fuel is in either tank unless you have run one of them dry.

16) While descending from the traffic pattern with very low levels of fuel, the tank ran dry, somewhat expectedly.

17) Know when this happens how long it takes to switch tanks, run the pump, have fuel arrive at the engine, and come back to life...  expect a handful of seconds to tick by while the descent rate starts increasing...

18) In this final approach situation...the engine comes back to life, It is now delivering high power on final.  Going around is not the best option since you clearly proved to yourself, you don't know how much and where the fuel is.  Enough fuel to go around?  Will it be able to pull fuel while pointing up for the climb?

19) know that the system is designed for the climb.  And not for the gliding descent...

20) it is a good idea to have working fuel level gauges.  Calibrated FF is extra good. I really appreciate the totalizer. Adding digital fuel level would be extra special...

Thanks to fuellevel for adding the pro wealth of knowledge on this subject. 

Let me know if I have missed something.  The topic is simple, yet complex enough, that it only takes one tiny detail to not align right to become a glider pilot.  It is an amazing feeling you get as you want to arrest the descent by pulling back on the yoke instinctively... you simply can't...

PP thoughts only, not a CFI or mechanic...

Best regards,

-a-

Edited March 29, 2017 by carusoam

[gsxrpilot]

Great summation as usual, but I think you've missed the most important reason why... It's the only way to know exactly how much useable fuel you have in each tank. When my safety depends on having fuel in the tanks, I'd like to have confidence to know exactly how much useable is in each tank. It's not often exactly what the POH says. And also depends how much or to what mark I fill the tanks.  In my previous Mooney, because I would periodically run a tank dry, then top it off after landing, I knew exactly how much fuel I had, and then with my totalizer, I knew exactly when it would run out. I could watch the totalizer and know within +/-15 seconds, when the tank would run dry. 

On an IFR cross country, and after having to divert to an alternate, it was very comforting to know with complete confidence exactly how much fuel I had and where it was. That confidence allows me to calmly and confidently fly the airplane, set up properly and shoot a nice approach to an alternate without sweating over fuel.

[fuellevel]

-a- that is spot on.  

I do see gsxrpilot's position - it is a well founded aviation viewpoint and shared by many successful pilots.   

His core values, are what we are trying to accomplish -

A system so that any pilot of any capability can at any time can look down and with confidence know where the fuel is and how much is available.

Workarounds for this issue are endemic in GA and they speak volumes to the quality of instrumentation supplied to us for this purpose.   I wouldn't repurpose a Cessna fuel gauge on my lawn tractor, but I am newly found picky about the subject. 

The Transport Canada references I used,  speak to a culture of pilots not trusting poor or inadequate fuel gauges and using fuel sticks to complement flight planning - These initial assessments for various human reasons end up, on occasion,  being wrong.   

I have anecdotal pilot comments from Beechtalk mostly,  that given adequate fuel indication,  the need or desire to run a tank completely dry is a mute point.  

 

[gsxrpilot]

And to that end... I've installed the CiES fuel senders in my K. Knowing the fuel situation in flight at all times, increases the overall enjoyment and is frankly, priceless.

[Shadrach]

Well one thing I can say about this thread is that GA pilots (at least MSers) of today are far more conscientious about fuel management then they were decades ago.  

A totalizer is on my list of wants, it's not a need.  Using time and confirming the position of the sight gauges (mechanical sending unit) on the wings, I  always anticipate when a tank will run dry on the conservative side. This engine always runs longer than my calculations but I'm usually within 10 mins.  On one occasion I was surprised that the engine continued much longer than expected (~30 mins), but that was due to a short taxi and very good climb performance against what I figure for fuel usage during taxi, take off and climb (flat 10gl).

[Guitarmaster]

On 3/29/2017 at 9:13 AM, gsxrpilot said:

And to that end... I've installed the CiES fuel senders in my K. Knowing the fuel situation in flight at all times, increases the overall enjoyment and is frankly, priceless.

I am thinking about installing these in my "F" but they are quite expensive.  I DO think they would dovetail nicely with the EDM-900.  

How has your experience been with them?

[gsxrpilot]

1 hour ago, Guitarmaster said:

I am thinking about installing these in my "F" but they are quite expensive.  I DO think they would dovetail nicely with the EDM-900.  

How has your experience been with them?

I'll have to tell you after I fly with them.