Level switches in the fuel system for alarms - Yes, No, Why?

Discussion in 'Firewall Forward / Props / Fuel system' started by wsimpso1, Dec 14, 2018.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Dec 15, 2018 #41

    BJC

    BJC

    BJC

    Well-Known Member

    Joined:
    Oct 7, 2013
    Messages:
    8,879
    Likes Received:
    5,732
    Location:
    97FL, Florida, USA
    lots of good discussion here. Two comments:

    With thin, low dihedral, wings and multiple tanks, the venting system design is challenging and very important.

    When my EFIS is powered up, a discrepancy (I don’t recall the triggering amount / percentage) between the total measured fuel in four tanks and the totalizer-computed fuel on board triggers an alarm to check fuel.


    BJC
     
  2. Dec 15, 2018 #42

    Dan Thomas

    Dan Thomas

    Dan Thomas

    Well-Known Member

    Joined:
    Sep 18, 2008
    Messages:
    4,712
    Likes Received:
    1,941
    Large airplanes with big tanks that run a long way out in the wings will use multiple level sensors. The signals are summed for the gauges. If I were Bill I'd install three resistive sensors along the length of the tank, connect them in series, and feed that to a calibratable digital gauge.
     
    wsimpso1 likes this.
  3. Dec 15, 2018 #43

    rv6ejguy

    rv6ejguy

    rv6ejguy

    Well-Known Member

    Joined:
    Jun 26, 2012
    Messages:
    3,546
    Likes Received:
    2,572
    Location:
    Calgary, Alberta, Canada
    I took most of my flight training in PA-28-151s which is where my distrust of float types fuel gauges came from. Later flying a Grumman Tiger and my RV, same thing. They give you a rough guide of fuel remaining only and dance around pretty good in anything but glass smooth air. Certainly not like my BMWs which are accurate to less than a half liter and give you very accurate range remaining through the ECU.

    Some time in PA-31Ps and King Airs and these had pretty accurate gauges which you could rely on.

    Yes, it seems some pretty poor stuff made it through certification on light aircraft. Notice the regs don't specify accuracy of said gauges...

    I only cared that my RV gauges read reasonably accurate at the 1/4 mark and they do when I tested them. I know I have 4.5 - 5 US gallons remaining there and I never feed fuel from them below that point. When flying though, I rarely look at them, just using time since TO to switch tanks and calculate rough fuel remaining.
     
  4. Dec 16, 2018 #44

    Dan Thomas

    Dan Thomas

    Dan Thomas

    Well-Known Member

    Joined:
    Sep 18, 2008
    Messages:
    4,712
    Likes Received:
    1,941
    The dancing fuel gauge is because the sensor is measuring fuel level in only one spot in the tank, and it isn't the middle. The middle is the most stable spot, but sensors are almost always at the lower end of the tank. Multiple sensors, wired in series, will cancel the sloshing out. If the fuel surges to one end of the tank and raises that float, the other end drops its float and the total resistance stays the same. Airplanes like the King Air will have multiple sensors. The Lockheed Electras I worked on had five per tank, and there were two tanks per wing. The whole wing, between the front and rear spars, from about five feet out from fuselage centerline, was tanks.

    Cars put their sensors in the middle of the tank, and the tank isn't usually long and shallow and skinny like some aircraft tanks. I'd bet they also use damping circuitry to stabilize the gauge needle. Some old cars (60 and 70 years ago) used a thermally-driven gauge; its temperature was controlled by the tank sensor's resistance, and a bimetal strip moved the needle. Very stable. My '51 International had a thermal float sensor, too: the float moved a lever that drove a tiny cam that moved a metal reed toward a bimetal strip; the strip would heat and bend away from the reed and break the circuit. It thereby send an on-off signal to the gauge. When the tank was full the cam forced the reed harder against the bimetal strip and it had to heat longer before it would bend far enough to open the switch, and the gauge's strip got current longer, got hotter and the needle went up farther. Empty was the opposite.
     
  5. Dec 16, 2018 #45

    rv6ejguy

    rv6ejguy

    rv6ejguy

    Well-Known Member

    Joined:
    Jun 26, 2012
    Messages:
    3,546
    Likes Received:
    2,572
    Location:
    Calgary, Alberta, Canada
    BMWs have shallow, flat tanks as they are under the back seat which is why I used that example. Whatever they do, they work very well compared to light aircraft. Could be lots of digital averaging between the sensing unit and gauge via the ECU I'd guess.
     
  6. Dec 16, 2018 #46

    Toobuilder

    Toobuilder

    Toobuilder

    Well-Known Member

    Joined:
    Jan 20, 2010
    Messages:
    4,286
    Likes Received:
    3,029
    Location:
    Mojave, Ca
    We should keep in mind that this is Bill's airplane and he says he has already installed capacitance senders in the wing tanks. That element of the design is locked down. The discussion concerns the header tank.
     
    wsimpso1 and rv6ejguy like this.
  7. Dec 16, 2018 #47

    wsimpso1

    wsimpso1

    wsimpso1

    Well-Known Member

    Joined:
    Oct 19, 2003
    Messages:
    5,627
    Likes Received:
    2,866
    Location:
    Saline Michigan
    It is large enough to use as part of your enroute fuel. Sounds like another Princeton gage is appropriate.
     
  8. Dec 16, 2018 #48

    Marc Zeitlin

    Marc Zeitlin

    Marc Zeitlin

    Well-Known Member

    Joined:
    Dec 11, 2015
    Messages:
    424
    Likes Received:
    571
    Location:
    Tehachapi, CA
    I agree 100%. And it seems to me that the simplest, easiest, most information dense system that allows for the most flexibility in knowledge is to install a level sender in the header tank, which is NOT long and skinny, and CAN be read accurately. A level switch gives far less info, with no trending. Why - to save $50?

    It's also relatively simple to get (maybe not "accurate", but decent) readings in a long skinny tank that has fuel sloshing around in it with a long capacitance sender that goes from the highest point in the tank to the lowest point in the tank. Then put a 15 second averaging filter on it. You'll get accurate (eh - decent) readings for the whole tank that way. Or, as others have said, multiple senders, with logic to determine what they're telling you. WK2 did that - we had a bunch of senders all over the **** place, with all the tanks and sumps it has, between the wings and two fuselages.
     
    wsimpso1 and rv6ejguy like this.
  9. Dec 16, 2018 #49

    Toobuilder

    Toobuilder

    Toobuilder

    Well-Known Member

    Joined:
    Jan 20, 2010
    Messages:
    4,286
    Likes Received:
    3,029
    Location:
    Mojave, Ca
    In that case, I agree. A fuel level guage will be of use. However, for the majority of the flight the quantity will be maintained at a constant level so a "not normal" warning would be appropriate to break a natural tendency to "expect the normal". An auxiliary warning driven by a float switch will give you an early heads up that the wing tanks are not feeding.

    Both indications should be easy, reliable and inexpensive.
     
    wsimpso1 and rv6ejguy like this.
  10. Dec 16, 2018 #50

    Dan Thomas

    Dan Thomas

    Dan Thomas

    Well-Known Member

    Joined:
    Sep 18, 2008
    Messages:
    4,712
    Likes Received:
    1,941
    Those senders are commonly summed as well.
     
  11. Dec 16, 2018 #51

    Joe Fisher

    Joe Fisher

    Joe Fisher

    Well-Known Member

    Joined:
    Feb 11, 2007
    Messages:
    1,383
    Likes Received:
    374
    Location:
    Galesburg, KS South east Kansas
    Back in January 1985 I flew the family in a 172 from west of St Louis to Fairbury Ne. We stayed 4 days the temp. never got above 0. The tanks were full it was -5 we took off, climbed to 9500 good tail wind (150 ground speed) after an hour it only used fuel out of the right tank. When I tried to switch tanks the valve was stuck, it was obvious the link to the valve was not strong enough to turn the valve. Landed at Columbia Mo., topped off the right tank. Got back in the airplane and the fuel valve worked normally and the tanks fed evenly . A totalizator would not have told me were going be short.
     
  12. Dec 16, 2018 #52

    rv6ejguy

    rv6ejguy

    rv6ejguy

    Well-Known Member

    Joined:
    Jun 26, 2012
    Messages:
    3,546
    Likes Received:
    2,572
    Location:
    Calgary, Alberta, Canada
    Neither would any gauge or light in this case.

    I've seen pilots run tanks dry before switching. Bad idea if you only have 2 tanks and the new one doesn't feed for whatever reason.

    In my RV, runup checklist requires you select tank 2 (startup and taxi on tank 1). In flight I switch tanks at the 1/2 hour mark, 1 hour mark, 1 1/2 hours, 2 1/2 hours etc. This gives you some fuel remaining for a good portion of the flight in case the next tank doesn't feed.
     
    wsimpso1 likes this.
  13. Dec 16, 2018 #53

    Dan Thomas

    Dan Thomas

    Dan Thomas

    Well-Known Member

    Joined:
    Sep 18, 2008
    Messages:
    4,712
    Likes Received:
    1,941
    The selector in the 172 is at the lowest point in the whole system. The factory put an eighth-inch pipe plug in the bottom of the selector, extending through the belly skin, and demanded its removal and draining of contaminants at 200-hour intervals. 99% of the airplanes never got it done, and I would find those plugs seized in there solidly. Once I got them out there would be water and some ugly junk coming out. I installed quick-drains there and told the students to use them at each preflight.

    I would bet that your airplane had water in there that not only froze the valve in one position, but water in the line from the left tank was also frozen, preventing any flow. Be thankful there wasnt just enough ice in both lines to kill the engine on takeoff, when the flow demands are the highest.
     
  14. Dec 16, 2018 #54

    Joe Fisher

    Joe Fisher

    Joe Fisher

    Well-Known Member

    Joined:
    Feb 11, 2007
    Messages:
    1,383
    Likes Received:
    374
    Location:
    Galesburg, KS South east Kansas
    With Pipers only run on one tank at a time. Pipers are what I fly most. Cessnas they just leave the selector on both all of the time. I expect if it was stopped in the right tank it would not have made it to the runway.
     
  15. Dec 17, 2018 #55

    proppastie

    proppastie

    proppastie

    Well-Known Member

    Joined:
    Feb 19, 2012
    Messages:
    3,505
    Likes Received:
    1,006
    Location:
    NJ
    I put my Cherokie in a field that way. Switched tanks before take off to one with a frozen line. Quit at 200 ft. At the far end.

    Start taxi and take off on the same tank
     
  16. Dec 17, 2018 #56

    Marc Zeitlin

    Marc Zeitlin

    Marc Zeitlin

    Well-Known Member

    Joined:
    Dec 11, 2015
    Messages:
    424
    Likes Received:
    571
    Location:
    Tehachapi, CA
    Agreed. My system is to taxi and takeoff on a known good tank (with FI, if I can taxi for a few minutes and do a 1900 RPM runup, then taxi onto the runway, I know the tank is feeding and will last through the climbout). Then, after an hour or so, I switch tanks, but ONLY if I'm within gliding distance of an airport. If the second tank works, now I know that I've got two good tanks and can stop worrying about fuel flow. But I will only switch tanks when I'm within gliding distance of an airport.
     
    rv6ejguy and Joe Fisher like this.
  17. Dec 17, 2018 #57

    Chilton

    Chilton

    Chilton

    Well-Known Member

    Joined:
    Jun 28, 2014
    Messages:
    63
    Likes Received:
    15
    Location:
    Arusha
    My reasoning is that if your header tank holds 50 minutets at cruise and you have configured the aircraft for IFR then you should be able to run both main tanks fully dry and still have a little over IFR reserves in the header. At that point if you are still airborne due to adverse circumstances I would certainly want accurate gauges so I knew what was left, for headers with very small capacity i would not worry but with the amount you have in here I would.
     
    wsimpso1 likes this.
  18. Dec 17, 2018 #58

    rv6ejguy

    rv6ejguy

    rv6ejguy

    Well-Known Member

    Joined:
    Jun 26, 2012
    Messages:
    3,546
    Likes Received:
    2,572
    Location:
    Calgary, Alberta, Canada
    Runup is completed on the 2nd tank so no chance of that, especially with EFI where the engine would stop immediately if there was no feed. If you don't check that both tanks feed before takeoff, how will you know they will in flight?
     
    wsimpso1 likes this.
  19. Dec 17, 2018 #59

    blane.c

    blane.c

    blane.c

    Well-Known Member HBA Supporter

    Joined:
    Jun 27, 2015
    Messages:
    2,881
    Likes Received:
    492
    Location:
    capital district NY
    Warning bells and lights quickly become ignored like your wife's voice during your favorite sports game if they come on all the time. I am saying if during what turns out in the end to be normal operation or procedure the lights and sound start activating they are going to start being ignored, and once the habit of ignoring them is developed they will be ignored when it really does matter.

    Excluding leakage, if you start with 5 hrs worth of fuel in the tank, you have 5 hrs to fly. Simple.
     
  20. Dec 17, 2018 #60

    Marc Zeitlin

    Marc Zeitlin

    Marc Zeitlin

    Well-Known Member

    Joined:
    Dec 11, 2015
    Messages:
    424
    Likes Received:
    571
    Location:
    Tehachapi, CA
    That's true. False alarms only train you to ignore the alarm. However...

    Not simple. I don't know about you, but my fuel burn is not always exactly the same. If I'm at 7500 ft., it might be 8.8 gph, leaving me with 6.6 hrs of fuel with full tanks, reserves notwithstanding. But if I'm at 15.5K ft., and burning 6.2 gph, then I've got 9.3 hrs of fuel, no reserve. And at 15.5K ft, my reserve will last a lot longer than at 7500 ft. My altitude changes based on winds, weather, and throttle/mixture setting. I don't have a fixed amount of fuel TIME on board, I have a fixed amount of fuel VOLUME on board. They're NOT the same thing; they don't measure the same thing, and they don't represent the same thing. You use a clock to determine how much fuel you have if you don't have a good level gauge. If you CAN have a level gauge, it's going to be a lot more accurate than a clock.

    Rather than indirectly know how much fuel you have based on how much you put in and how long you've been flying, or based on a totalized fuel burn, why not just read the level directly, with a device that costs $140? No guessing, no calculating, alarms on the EMS to tell you when one or more tanks (that are measured individually) are low, at whatever level YOU determine is appropriate for alarms to go off.

    I have my yellow alarm set at 6 gallons on each side. My red alarm comes on at 4 gallons. My unusable fuel in a Vx climb (worst case, per my measurements) is about 3 gallons, so I can still do a go-around without switching tanks even if the alarm has just started going off.

    I don't understand why anyone WOULDN'T want to know how much fuel they have on their plane... How the information could be helpful is obvious - how it could be harmful is not.
     
    wsimpso1 likes this.

Share This Page

arrow_white