Fuel Flow Discussion, Moved from Float Mounting

[Battson]

short story; Running on "Both" without having interconnecting tanks. Pressure differential between the tanks caused fuel flow stoppage.

I have gone full circle on this and I'm not convinced that interconnecting the tanks will actually prevent this; having the benefit of clarity via Bob.
I am really sorry to "flip-flop" on this.
Having had the event at our home airfield I got very interested and I cannot understand some of the logic.

The facts are:

• This issue seems to only affect engines with a fuel pump.
• Any fuel pump will rather suck air than fuel, if air is available.
• Low fuel in one tank is universal in the incidents discussed.
• If the engine sucks for air long enough it will stop, whether the selector is on both or L/R

Occam's razor: It seems much more likely that sucking air bubbles due to prolonged unporting or running a tank dry caused these incidents. A cross vent will not prevent that.

The whole mysterious scenario of one tank robbing fuel from the other just seems unlikely. Sorry to the supporters of this idea.... but nobody has offered an adequate explanation, the necessary pressure difference is too great, and the experts don't support it.

Also the hypothesis of any plane with a "Both" selector needing a cross vent, we haven't provided evidence. I THINK the reason for the cross vent is probably as a backup in case one of the external vent blocks. That way, you can leave the selector on both and if there is fuel you're likely to get it, even if a vent blocks. That was the whole point of the both selector, avoiding accidents due to fuel mismanagement.

I understand Bob is considering providing some additional guidance about fuel system design.

[JimParker256]

The Rotax powered RANS S-6ES I'm currently flying was designed with two wing tanks (each with a sight gauge and a set of fore and aft fuel pickups) plumbed to a single "on/off" fuel valve, then to a gascolator, and finally to the engine-driven fuel pump. It's functionally very similar to Bob's design, if you were feeding from "BOTH" tanks all the time.

It is very common to hear reports about one wing tank draining much quicker than the other tank – sometimes to the point where there is no longer any fuel visible in the sight gauge on one side while the other is indicating nearly full. The owners that report this issue say that it's always the same tank that "empties" first – which would seem to me a clear indication that their wing tank vents are producing different "pressures." The designer's recommendation was to check the length of the vent and to compare the angle of the cutoff. (The vent line starts at the top of the tank, and runs back and down, ending in a tube that sticks out below the wing, near the jury strut, with a 45º forward-facing cut-off. If the vent is closer / farther from the jury strut, or if the extension of the tube below the wing surface is different, or if the cutoff angle was different, then the "pressure" due to air flow might also be different.)

However, there have been NO actual reports of fuel starvation resulting from this setup. The "going assumption" is that the pressure head from the fuel in the "mostly full" tank is apparently enough higher than the pressure head from the "air" in the lines and tank on the other side that the engine-driven fuel pump does not "suck air" in those situations. Even those who say one of their tanks "always empties" well before the other say they have no issues utilizing the entire useful fuel load in flight.

I did make a change to my S-6's fuel system design, however, in keeping with RANS' latest design changes to the S-20 and S-21. I added a small (2.75 gallon) header tank behind the seats, and redirected the wing tank outlet lines to flow into the header tank. The header tank vents only through the wing fuel lines. There's a "float switch" (connected to a low fuel warning light on the panel) that triggers when there is 2.5 gallons remaining (roughly 20-30 minutes) of fuel remaining – that's my "Final warning, you idiot! Get this plane on the ground NOW!" light...) The header tank's single outlet flows to the fuel cutoff valve (On/Off), the gascolator, and the engine. At some point, I'll probably add an electrical fuel pump inline between the header and fuel valve, just in case the engine-driven pump ever fails. (Rare, but it can happen.)

With this change, I see very even fuel burn from my two wing tanks – I rarely see anything more than a 1/2 gallon difference in the sight gauges. I'm happy with the setup.

[zkelley2]

The Rotax powered RANS S-6ES I'm currently flying was designed with two wing tanks (each with a sight gauge and a set of fore and aft fuel pickups) plumbed to a single "on/off" fuel valve, then to a gascolator, and finally to the engine-driven fuel pump. It's functionally very similar to Bob's design, if you were feeding from "BOTH" tanks all the time.

It is very common to hear reports about one wing tank draining much quicker than the other tank – sometimes to the point where there is no longer any fuel visible in the sight gauge on one side while the other is indicating nearly full. The owners that report this issue say that it's always the same tank that "empties" first – which would seem to me a clear indication that their wing tank vents are producing different "pressures." The designer's recommendation was to check the length of the vent and to compare the angle of the cutoff. (The vent line starts at the top of the tank, and runs back and down, ending in a tube that sticks out below the wing, near the jury strut, with a 45º forward-facing cut-off. If the vent is closer / farther from the jury strut, or if the extension of the tube below the wing surface is different, or if the cutoff angle was different, then the "pressure" due to air flow might also be different.)

However, there have been NO actual reports of fuel starvation resulting from this setup. The "going assumption" is that the pressure head from the fuel in the "mostly full" tank is apparently enough higher than the pressure head from the "air" in the lines and tank on the other side that the engine-driven fuel pump does not "suck air" in those situations. Even those who say one of their tanks "always empties" well before the other say they have no issues utilizing the entire useful fuel load in flight.

I did make a change to my S-6's fuel system design, however, in keeping with RANS' latest design changes to the S-20 and S-21. I added a small (2.75 gallon) header tank behind the seats, and redirected the wing tank outlet lines to flow into the header tank. The header tank vents only through the wing fuel lines. There's a "float switch" (connected to a low fuel warning light on the panel) that triggers when there is 2.5 gallons remaining (roughly 20-30 minutes) of fuel remaining – that's my "Final warning, you idiot! Get this plane on the ground NOW!" light...) The header tank's single outlet flows to the fuel cutoff valve (On/Off), the gascolator, and the engine. At some point, I'll probably add an electrical fuel pump inline between the header and fuel valve, just in case the engine-driven pump ever fails. (Rare, but it can happen.)

With this change, I see very even fuel burn from my two wing tanks – I rarely see anything more than a 1/2 gallon difference in the sight gauges. I'm happy with the setup.

That's how my rotax powered airplane was plumbed to. I would suggest getting that electric backup in there asap. The fuel pump on the rotax has gone through something like 4 service bulletins "recommending" replacement of the earlier design. They are not the strong point and the engine quits without fuel pressure despite being carb.

[Bcone1381]

What questions should be asked to lead this towards a scientific based solution. Does this issue need a leader, spokesperson, authority, or a facilitator? Who should that person be? Has anyone plumbed altimeters into upper sight gage plumbing to measure data? Does anyone have a aircraft/circumstances that can make this a e repeatable event?

[svyolo]

I think running a tank dry and killing the engine while the other is full is primarily a fuel management error, even the there is also a fuel flow problem. Two problems simultaneously.

Sometimes I will try new stuff, even if it is different than convention. Sometimes it works, sometimes not. Sometimes I actively seek out wisdom from the "herd". Despite being a pilot for work for over 30 years, I am one of the very least experienced here with light aircraft, and their systems.

Sometimes logic is wonderful, but sometimes it can't identify threats or the risk of threats because of external "unknowns". That is where the "herd" is king. Experience in the real world. That includes external threats, and the experience of how often they happen.

Since I chose EFI and that is a choice of the vast minority (going against the herd) I kind of did of a deep dive into fuel systems in light aircraft. Not theoretically based, but I wanted to see what the "herd" did. Luckily Cessna built thousands of high wing aircraft with injected Continentals for me to see what they did, including their change and development over the decades. Piper built more than a couple but I think they were mostly carburated.

Bob's stock fuel system is perfectly acceptable. It is redundant. You have 2 sides. If you run one dry, whether or not one stops feeding, you failed to monitor your fuel state properly. If it results in a mishap, that fact will become at least a "contributing factor" if not a "causal factor".

More redundant is usually better. Not always, especially if it increases complexity or workload. It also may not help long-term if it makes the pilot complacent.

The "herd" went with a cross-vent in high wing aircraft with a fuel valve "both" setting. It is also specified in rules of thumb in a bunch of fuel system textual descriptions for the same application. In this case, I am going with the herd.

[JimParker256]

Note that your "herd" (Continental engine power) also designed the fuel return to always go to a single tank, thus saving money over a full-duplexed fuel valve and additional return line to the other tank. I think that fact alone would argue for cross-venting the tanks. BWTHDIK?

[Mark Goldberg]

If you fly along in a slight slip or skid - one tank will drain first. Very easy to do. My point is that it is not different pressures in the tanks that might make one drain faster than the other - but how we are not flying wings level. Mark

[Nev]

This is probably the most common cause Mark. We used to sometimes balance tank quantities when needed by opening a cross feed valve and deliberately flying out of balance to get fuel across to the other tank. In aircraft without a cross feed valve the fuel can be free to flow through the fuel selector in BOTH to the other tank.

[Nev]

Cessna identified another cause many years ago - internal fuel line pressure differences coming from the slight variations in fuel line plumbing (different radiused bends etc). These sometimes led to vapor locks on one side, and caused an uneven fuel feed rate. So while closely monitoring your tanks is always a great idea, in some cases it can be treating a symptom without realizing there’s a cause.

While Cessna were unable to replicate the exact problem consistently, they did find that tank cross-venting largely eliminated this particular issue. They also speculated that very specific atmospheric conditions were playing a part in the vapor-loc scenario, meaning you could fly without incident for a long time while always having the potential issue lurking.

[Battson]

In our case, we have two Bearhawk 4-place's in the hanger at the moment, and we flew both yesterday. The first machine sucked 60 L (15 gal) out of one tank and didn't touch the other. Mine draws perfectly equally on both tanks, as always.

Same two pilots flew in both machines at once, same day, almost identical setups apart from:
- the plumbing from the tanks to the selector - one has many more fuel drains and lots more AN hardware
- use of aux tanks vs no aux tanks (with gas cap vents)

Literally everything else about the fuel system is the same. Same gascoalator, same pumps, same injection system, same engine, same venting.

It seems to be very fiddly to pin down why the tanks draw at different rates. Perhaps a cross vent would help THAT symptom.

[jaredyates]

I wonder if swapping the caps would make any difference?

[Nev]

It’d be interesting to disconnect the fuel lines and compare flow rates L vs R after the selector, before the pump.

[Bcone1381]

Fuel Cap position....would the pressure differential exist between each cap's vent if each is placed slightly differently at the wing, or the alignment is not exact.......is one cap slightly higher, or slightly more forward slightly angled?

The only explanation I can see at this point is at best theoretical, but might be worth to investigate.

Lets say the L vent pressure is 4 units and the R is a 6 in cruise and causes an imbalance. Then one slows for landing, and extends flaps....it would seem to change the pressure distribution on the wing, and maybe change the distribution at each caps vent, inducing crossflow of fuel thru the selector valve. Would not the crossflow then limit flow to the stem of the valve's "tee" going to the engine?

[Battson]

Good questions - we tried swapping fuel caps, it appeared to make a difference at first, but then it transpired that it made no difference whatsoever.

We haven't tried swapping fuel caps on the aux tanks yet.

[Ed.Meyer]

Having followed this thread and then seeing the safety notice from Bob regarding fuel pumps with his fuel system design I decided to do a test.

As some of you know, we have a Patrol with the EFII ignition and injection system which includes fuel pumps (main and secondary) to boost fuel pressure up for the injectors. Also, return lines were installed to return the excess fuel from the continuously running pump. All the details of our system installation in the Patrol are in a Beartracks article I wrote a while back.

Not knowing positively whether un-porting of the fuel outlets on one tank would allow air to enter the fuel system and starve the injection system,a concern expressed in this thread; I decided to do a test:

I flew up over my home airport about 4500 AGL with about 9 gallons of fuel in each side. I was using a cruise power setting with about 8GPH fuel flow showing. I turned the fuel valve to the right tank (I usually run both) and then put the right wing low in a pretty hard slip/skid to intentionally un-port the fuel outlets. I expected that fuel pressure would quickly drop and the engine would quit. After about 10 seconds or so, the fuel pressure did drop and the automatic switching to the secondary fuel pump occurred but the pressure stayed low. At this point I knew I had succeeded in un-porting the fuel outlets on that side. I then went back to normal coordinated flight and the fuel pressure slowly recovered in about 10 - 15 seconds. Much to my surprise, the engine never missed a beat through this.

Following this, I switched the fuel selector to both and again did the right wing low slip/skid. I held this for probably about two minutes monitoring the fuel pressure closely. It never wavered in the slightest.

My conclusion from this is that there is apparently enough fuel flow capacity feeding from only one tank via 3/8" lines to supply all the fuel the engine needs plus whatever amount the fuel regulator returns to the tank even with the other tank feed open to the vent. I recognize that this is not an exhaustive test but it was enough to boost my confidence in the fuel system as installed.

[Nev]

This is an interesting experiment Ed. I think that in theory the way you have put the aircraft into a slip, both fuel out-let’s should have un-ported and the engine should have stopped. Given that it didn’t stop, it may be still running fuel from a gascolator or header (I didn’t look at the details of your system), albeit at a reduced pressure. I suspect that the gascolators may be acting as a mini header tank, along with larger diameter fuel lines where used.

It would be interesting to see the same experiment, but instead of an out of balance situation, try an extended steep climb or steep descent. This should unport one outlet, and give the pump an opportunity to suck air in preference to fuel. It may take quite some time before the engine stops because it has to burn all the fuel stored in the lines and gascolator first.

The aircraft that have experienced engine stoppages so far, what we don’t know is how long the fuel stopped feeding before the engine stopped. It may have been several minutes. This effect has implications where the gascolator is plumbed between the fuel selector and the engine. There have been accidents where a pilot has started the engine with fuel selector in the OFF position, and subsequently become airborne, only to have the engine stop at low altitude on climb out. Many fuel systems were designed to prevent this by only having a very limited amount of fuel “stored” in the system (cause the engine to stop before it had a chance to become airborne), but there’s an obvious trade off between that and having sufficient fuel to cater for a un-porting event.

[zkelley2]

Your thoughts are what I have experienced Mark. I have started with the fuel selector off on purpose. It does start, and it'll idle for a few minutes, but putting power to it for taxi I can't get 100ft.
If I started the airplane on the runway in position I don't think I could put full power to it before it'd stumble.

[Ed.Meyer]

Airplanes that have carburetors have a float bowl that holds enough fuel for a short idle. Ours will not start with the fuel valve off because the pump will not build pressure. Screams loudly in protest. I think the gascolator probably was a factor in supplying a bit of fuel.
I agree that this could be further tested with climb/decent and lower fuel levels. Given what I learned so far, I think that fuel coming from a single port would be enough since each port is a 3/8” line and even with both ports feeding from a single tank, they tee into a single 3/8” line. Doing this test at real low fuel levels I an not willing to do.

[zkelley2]

I think we know, or we all should know that a single 3/8 port and line can feed enough fuel because we've all done the fuel test recommended by the AC and even asked for by most DARs/FAA people.

[Nev]

I think it’s well established that a single 3/8 line will provide sufficient fuel under gravity feed conditions.

The issue is what happens once a pump is added to the system, and one line then becomes un-ported.
Will the pump suck fuel from the other line, or will it suck air in preference to fuel from the un-ported line, thereby resulting in a stoppage.

[Bcone1381]

Having followed this thread and then seeing the safety notice from Bob regarding fuel pumps with his fuel system design I decided to do a test.
...........
My conclusion from this is that there is apparently enough fuel flow capacity feeding from only one tank via 3/8" lines to supply all the fuel the engine needs plus whatever amount the fuel regulator returns to the tank even with the other tank feed open to the vent. I recognize that this is not an exhaustive test but it was enough to boost my confidence in the fuel system as installed.

I think you are right, Ed. You experiment gives good evidence that the first fuel pump in your system is supplied by gravity. Will a change in power change your fuel flow demand? I don't think so, but not sure. I seems to me like your system has a fixed rate of fuel demand out of the tank regardless of the power setting, and system returns unused fuel. Am I right?

[Ed.Meyer]

You are right Brooks that the system returns unused fuel. When I first turn it on prior to engine start, I can hear a short burst of bubbling in the tanks as air is purged. I am not sure if the volume pumped stays the same. Fuel pressure varies some with manifold pressure. Higher MAP has higher FP. The same fuel pumps drive the EFII system for an O540 setup so I think the volume being pump is somewhat higher than the engine demand.than our O360.

[svyolo]

ED;
Do you return to a header, or the main tank(s)?

[Ed.Meyer]

Return is to main ranks via a duplex fuel valve.