Fuel Flow Discussion, Moved from Float Mounting

[svyolo]

I believe I read the requirement in AC43. I will try to look it up again. I was planning on return lines to the mains for my EFI so I had 1/8 NPT ports welded to the top center of the inboard edge of each tank. I am probably going to use it to cross vent the tanks.

[jaredyates]

I would like to understand the science here.. for me it seems that when the valve is on "both" the left tank can drain to the right tank, and vice versa. Does that not allow the fuel to cross over?

[Archer39J]

It does, if you were in prolonged uncoordinated flight you'd be "transferring" fuel from the wing high tank to the wing low one.

Just having the tanks connected at the fuel valve means they'll seek the same level and necessarily move fuel.

[svyolo]

It wasn't in AC43, it was in a different FAA (FAA-H-8083031A page 14-10) pub describing a single engine, gravity flow fuel system with 2 wing tanks and a fuel selector valve. It shows a different diagram for high wing/fuel injected. That diagram shows header tanks and no BOTH on the fuel valve. I believe this pub is an educational pub, not regs.

High-wing aircraft with a fuel tank in each wing are common. With the tanks above the engine, gravity is used to deliver the fuel. A simple gravity feed fuel system is shown in Figure 14-12.
Figure 14-12. The gravity-feed fuel system in a single- engine high-wing aircraft is the simplest aircraft fuel system.
The space above the fuel level is vented to maintain atmospheric pressure as the tank empties. The two tanks are also vented to each other to ensure equal pressure when both tanks feed the engine. A single screened outlet on each tank feeds lines that connect to either a fuel shutoff valve or multiposition selector valve. The shutoff valve has two positions: fuel ON and fuel OFF. If installed, the selector valve provides four options: fuel shutoff to the engine; fuel feed from the right-wing tank only; fuel feed from the left fuel tank only; fuel feed to the engine from both tanks simultaneously.

[svyolo]

Other than the little vent tidbit, there really isn't enough info on his particular fuel system to speculate what happened. It would be nice to learn something here.

[whee]

The reg is 23.975(a)(4). It offers no discription so is not particularly helpful. If we look back to when the regs were written in blood, as they say, we can see the same requirement in CAR 3.446 but no reasoning. CAR 4a has some helpful info.

§ 4a.606 Tank installation. No fuel tank shall be
placed closer to an engine than the remote side of a
fire wall. At least one-half inch clear air space shall be
allowed between the tank and the fire wall. Spaces
adjacent to the surfaces of the tank shall be ventilated
so that fumes cannot accumulate or reach the crew or
passengers in case of leakage. If two or more tanks
have their outlets interconnected they shall be
considered as one tank and the air space in the tanks
shall also be interconnected to prevent difference in
pressure at the air vents of each tank of sufficient
magnitude to cause fuel flow between tanks.
Mechanical pump systems shall not feed from more
than one tank at a time except by special ruling from
the Administrator.

Also helpful is this NTSB finding. Note that Luscombe fuel system is essentially the same as those found on Bearhawks except that there is no unusable fuel in a Luscombe.

http://www.aopa.org/asf/ntsb/narrati...20090810X75326

[zkelley2]

So you need to have a fuel pump for this to be a thing. It does not apply to gravity feed. If you have it on both and run 1 tank dry, it will suck the air from the empty tank and not the fuel from the fuller tank. If this happened in flight you can just switch to the tank with fuel in it, but you'll need to remember to do that and know which one that is. If you're not aware of this, you might think, it's on both, so clearly it has access to all the fuel.


The relevant FAR is 23.951b - Each fuel system must be arranged so that— (1) No fuel pump can draw fuel from more than one tank at a time; or (2) There are means to prevent introducing air into the system.

Then there's the differential head pressure issue as well. Which I think you could do by installing 1 fuel cap backwards and probably sideways and possibly some extent between 90 and 0 degrees.


Here's an experiment you can do at home that is the same phenomenon. Get 2 straws and a glass of water, put both straws in your mouth, one in the glass of water and one outside the water. Suck and see how much water you get in your mouth and how much air.

[Battson]

Try that experiment again with the glasses above your head and straws out the bottom of the cup...

This would be an issue with a low wing, which typically don't have a both position on the tank selector. Whereas most high wings do.

See Jon's post below.

[jaredyates]

I moved the discussion out of the floats thread, but it looks like some of the comments didn't survive the move?

[Sir Newton]

Thank-you
I am reading the thread now. I would have missed this completely had you not moved it.

[jaredyates]

Here's an experiment you can do at home that is the same phenomenon. Get 2 straws and a glass of water, put both straws in your mouth, one in the glass of water and one outside the water. Suck and see how much water you get in your mouth and how much air.

Thanks for this... It seems to apply especially in a low-wing situation. But whether our system is entirely gravity or not, it is still partially gravity-fed. If the pump is forward of the fuel valve, how can it be possible for one straw to be in the air? Wouldn't tank gravity pressure provide flow through the valve and up to the pump? Or is the issue that the pump is pulling more than the gravity pressure provides? Not trying to argue, but rather just to understand how this could happen in our case.

[Sir Newton]

FYI, I have lots to say regarding fuel systems designs & choices of pump types. Different types of pumps will act differently in given situations. This is VERY common knowledge in oil & gas processing industry.

[Archer39J]

Oh, this is the situation that only applies if your fuel pump is sucking fuel. We went over this already in the returnless FI thread.

In a properly constructed Bob fuel system this shouldn't be an issue, unless you're running FI with a full return. Remember that whole discussion about not over taxing the Bob systems ability to supply fuel?

[jaredyates]

I see, so it is a matter of pulling more than the gravity feed supplies.

[Archer39J]

At least for what zkelley was referencing. The NTSB report whee linked has to do with a plugged vent cap, but I'm not sure I follow their reasoning in why one plugged vent will cause an interruption in fuel from the unencumbered tank and vent, doesn't make sense to me at least.

[whee]

This is an important topic and I don't want anyone to get confused. Zach and I are talking about two different failure modes. The one he is referring to does require a fuel pump for it to be an issue. The failure mode I am referring to affects both gravity feed and pumped fuel systems in the same way.

If you read the CAR4a reg I quoted and the NTSB report I linked you should have a reasonable grasp of the issue.

[Bcone1381]

My head is spinning. Lets not guess or take anything for granted.

Facts....A standard Continental IO-360's engine fuel system supplies excess fuel to the engine and requires a fuel return system of some kind. In contrast, a Bendix RSA type of FI system does not return fuel.

Do we know if the accident aircraft....
-had the standard Continental fuel system installed?
-had one fuel tank empty when it crashed?
-had a tank vent plugged?
-had the tanks vented via caps or some other alternate means?
-what is the Fuel supply requirement for a Continental io-360 with a standard fuel system installation......ie how many gph does the system demand to be supplied from the tanks.
-Do we know how the fuel return system was designed on the accident aircraft?

Data communicated so far (not necessarily a fact)-
-Post #4 "Running on "Both" without having interconnecting tanks. Pressure differential between the tanks caused fuel flow stoppage."

My comments.
-I understand a pressure differential could develop if one tank vent is plugged or the cap is not aligned into the airstream-.
-I don't understand how a differential pressure that develops between two tanks when in BOTH could suck the good tank dry.
-I don't understand how a pressure differential could stop fuel from flowing from a properly vented tank that has fuel in it.
-A pressure differential could change how the fuel is returned from the engine to the tanks.

[MattS]

I am a missionary pilot here in the Amazon Basin of Brazil. I have 1100 hours TT with over half of those occurring here in Brazil (270 in an RV6, 170 in the BH on wheels, and 160 hrs in a Murphy Rebel on straight floats). As mentioned in another forum, I reported that in 2018 we mounted the BH on Clamar amphibs. I had about 10 hours total flying the newly configured plane when I had the accident. The plane is equipped with the standard BH vented fuel caps which i checked before the flight and after the accident. They were not plugged. Someone asked about one tank being empty. There was fuel in both tanks although a limited quantity in one. Could have unported a fuel line. However, this is the very reason I always felt running on Both was safer- who cares if you unport a fuel line if there is fuel in the other tank, right? Why were the quantities of fuel not more balanced when they were connected while running on Both? I did note during the flight that is is harder to stay coordinated when on these large floats. Here is a portion of the accident report as I submitted to the mission board:

7:30AM; Take-off for Apui (425 NM) for a 4.3 hour flight before re-fueling and continuing to Humaita (2.0 hours). Diverted slightly for weather but mainly stayed about the Tapajos River for the first 2.5 hours. After Jacarecanga, headed west to Apui following the TransAmazonia highway. Flight level 6,500ft.

During the flight I left the fuel selector in both. However, I noticed about 3 hours in that it was pulling more fuel from the left tank. This did not happen immediately because at one point early in the flight the left tank had more fuel.

10:30AM Switched to using just the right tank to get the fuel more balanced. No change detected in Fuel pressure or usage between “both” or “right”.

11:00AM; Switched back to using both tanks. 11:05AM; Began power-on descent for Apui.

Flight heading 265 with a tailwind so I was set-up to enter a left down wind. The runway at Apui is asphalt with many potholes. With minimal experience in landing an amphib aircraft I wanted a extended, very controlled final. However, I also wanted to do a good visual of the runway because of the poor condition of the runway.

11:25AM; Descended to 500-600ft AGL for visual inspection of the runway. After completion, I powered up (~2200 RPM) and began climbing for an extended downwind. About 2.5 miles from the runway I entered a left base. Still with power and a slow climb. After wings leveled in the left base the engine quit. No sputtering, no signal. Just 2200 rpm to nothing. Remembering I had more fuel in the right tank I switched to “right” and then hit the electric fuel pump. The engine surged once and then quit again. I did not think to look at fuel PSI or flow. Looking towards the runway I decided I couldn’t make it (a floatplane without power descends at 1000-1500 FPM). Between me and the runway were a bunch of hills, trees, and a sawmill so I did not want to land short.

Looking to the right I saw a small lake (later the owner said it was 800 meters). I decided that was my best option and turned towards it. I had already lowered the landing gear so I began retracting them and set-up for a short landing (I saw the orange light indicating the wheels were coming up but don’t remember confirming “blue”). I put in 2 notches of flaps but soon realized I had turned too soon and was too high. I continued to fly just over the lake but with too much altitude. The lake had a 150 ft dogleg that I turned left to try to use as much water available. However I ran out of water/landing area. The owner said he had lowered the water level that month to clean out some tree trunks. Therefore the bank was 10-12 feet high. I ended up hitting the bank with the floats and they started to shear off slowing me down. Around the lake was a driving path. I slid across the path and then nosed the plane into another 15-18 ft embankment.

Later when we began removing the wings we discovered that I had 10 litters in the left tank and 55.5 in the right. This does not included the limited quantity that was spilled during the removal process.

The engine was not rotating when it impacted the ground (one propeller blade is bent backwards) so there was no sudden stoppage.

[svyolo]

Very sorry to hear about your accident with your new build. Must be heartbreaking.

Thanks for taking the time to write it up.

[Sir Newton]

Glad to hear this story. Thank-you.

[Cguy]

Don't want to add to any confusion, but my RANS S-7S has NO selector valves, both tanks feed a 3 gallon header all of the time. There IS a crossover vent line that between the two tanks, plus individual vents. For whatever reason, p factor, prop blast, don't know and don't care.....the right tank ALWAYS empties first. You can tell a new S-7 pilot by him freaking out when this happens, thinking he is going to run out of fuel! Doesn't happen, ever. Nothing wrong, no clogged vents or clogged feed lines, and messing with selector valves, in our case anyway, can add complexity, weight, and expense. I use my left tanks sight gauge as my fuel gauge, when it shows empty, I then start looking at my 3 gallon header tank's sight gauge, and I've landed with less then a gallon showing (oops) and the engine ( Rotax 912) never faltered. After landing, I drained the remaining fuel out, just as a test, to see if perhaps I had more fuel then indicated. Nope, I had exactly what I thought I had in flight, 1/2 gallon. This is such an idiot proof system, also making full use of the fuel on board IF YOU CHOOSE TO, (!) that if I build another plane I'd set it up the exact same way.

[zkelley2]

Thanks for this... It seems to apply especially in a low-wing situation. But whether our system is entirely gravity or not, it is still partially gravity-fed. If the pump is forward of the fuel valve, how can it be possible for one straw to be in the air? Wouldn't tank gravity pressure provide flow through the valve and up to the pump? Or is the issue that the pump is pulling more than the gravity pressure provides? Not trying to argue, but rather just to understand how this could happen in our case.

Yes, it's a much bigger issue in a low wing, but can still happen on a high wing with a pump, hence the FAR. I'd think you'd need to be in a quite high pitch attitude in our case. Also why low wing pipers don't have a both.

You can also get the pressure imbalance like John is talking about, gravity or not, and so if you're going to run both, you NEED to have a vent line between the tanks. To atmosphere does not count.
All of the Cessnas with both have vents between the tanks, but also have tank vents to atmosphere. They started doing this in the C140's. The early 140s have a L/R and no vent between them. When you get to the later 140s with a both, there's a vent. I'ts much easier to make the vent in a high wing than a low wing.
You'll also see, I think starting with the 170 POH's and I think it's even placarded on the 172 near the selector that even though they are gravity feed with tanks vented to each other, they recommend cruise flight in L or R, not both.

[robcaldwell]

Eric Newton's Sample POH only mentions the following about fuel selector positions in section 2:

• To ensure maximum fuel capacity when refueling, place the fuel selector valve in either LEFT, RIGHT or OFF position to prevent cross-feeding.
• Takeoff and land with the fuel selector valve handle in the BOTH position.

Do we need to amend this? I've always flown in BOTH on most of the high wings I've flown and not thought much else about it. I've only switched to LEFT or RIGHT as MattS says he did on a long cross country to level the tanks.

What can we take away from this to better understand fuel selector operation? Not saying he did anything wrong. Just trying to glean a "learning moment" from this incident.

[zkelley2]

It depends on the fuel selector you install and how you build the fuel system, if it has an interconnect and if it has a pump or gravity feed. A problem I already see is most fuel selectors in the off position allow fuel to transfer between tanks, just not out the outlet port. So for fueling if you're worried about transfer, you would want the selector in R or L, not both or off. This applies to parking on an incline as well.