EI vs EFI and the types of EFI

[Ed.Meyer]

I have tested this. I intentionally un- ported one tank by flying very uncoordinated, fuel selector on both, and EFII system. No fuel starvation occurred. This was in a thread and Beartracks a couple years ago.

[schu]

I have tested this. I intentionally un- ported one tank by flying very uncoordinated, fuel selector on both, and EFII system. No fuel starvation occurred. This was in a thread and Beartracks a couple years ago.

Ed,

When you post your experience are you saying that you believe that EFI is perfectly safe on a bearhawk with the mostly stock fuel system and a both duplex valve or are you just adding one data point to the discussion? Are you able to point to why your system works when the SDS and EFII documentation they both recommend a header tank for high wing aircraft, and both mention using a duplex valve that returns fuel into the tank feeding from? It seems like either their recommendations are wrong, or you got lucky, or there is something you changed that mitigates their concerns, or some combination of the above.

Anyway, I know you are returning fuel to both tanks but it's hard to understand if the fuel is actually going to both tanks equally or if your system is balancing itself through the valve quick enough to not matter. If you are in an extended slip it seems like you could find yourself in a situation where return fuel is going to the opposite tank feeding from. Is there a way to test this? Does it matter?

As for unporting tanks, the EFII documentation says "This is a fuel system that is capable of supplying fuel a minimum of 35gph to the engine area and returning excess fuel back to the selected fuel tank or to a header tank." Which means that at idle, you are probably returning at least 33gph to the tank. Looking at some gravity fuel flow tests to the inlet of the mechanical pump Rob's airplane flowed 28.1GPH on each side and 32.8GPH with both. Given that the EFII numbers are the minimum, and Rob's gravity numbers are below that, it seems reasonable that the EFII pump would be sucking in his installation.

Reading through your comments in bear tracks you said that you tested with a single tank and got it to drop fuel pressure but didn't hold it until engine quit, then switched to both and never got it to show low fuel pressure. The test was with 9 gallons in each side and 8 GPH fuel consumption. That points to some interesting observations. You were able to get a tank to unport enough to drop fuel pressure, but not stop fuel flowing when 32% full otherwise the engine would have died, and when on both with a pretty good slip pushing the fuel to the pump you were able to supply more than the 35GPH needed to feed your engine and the return.

Looking at all of this data, I theorize that your system works because the fuel flow (even with the return) is low enough that the system can keep the pump flooded from gravity and the return to tank on both works because the system can balance the tanks at the same rate or greater than whatever imbalance occurs in the dupex valve. If my theory is true then your airplane might be susceptible to an issue where you have the VFR 30 minute reserve (4 gallons) completely unport a tank, suck some air in, and can't clear the air bubble.

I suspect this is what happened to MattS. He only had 2.6 gallons in the left tank, sucked some air, and the tank with fuel couldn't burp itself fast enough to flood the pump. The only way to know would be to run one tank empty, then switch to both.

[Ed.Meyer]

I don't recall seeing anything from EFII, which is what I have, that said I needed a header tank when I was researching all this and I did not install one. One concern I had with a header tank is the need to vent it and unknown issues that might create. The stock system seems really simple and with the four fuel ports at least one will be feeding fuel in any flight condition that isn't inverted. I also suspect that having the two ports in each tank helps to clear any air than might get in a line.

I had read another thread about this concern for un-porting which caused me to wonder and do the test. When I got the pump to suck air the fuel pressure dropped to zero. I expected the engine to quit but it kept going. Don't know if manifold vacuum sucked enough fuel to run for a bit or what. But once fuel pressure dropped I new it sucked air. At that point i went back to coordinated flight and fuel pressure quickly came back up. When fuel selector was on both, I could not get it to suck air and fuel pressure never varied which I was very pleased to find. It was apparently feeding enough fuel from one tank by gravity to satisfy the pump which is below the floorboards in my airplane. I am sure if gravity fuel flow was measured at the pump inlet it would be grater than measured at the fuel rail of the engine. I am sure that if someone was motivated to do it, more thorough testing could be done but I don't want to unnecessarily create engine failures in flight. I came away satisfied with what I have.

As far as fuel returning, you are correct than I have know way of knowing how much is going to each tank. In 250 hours there has not been enough imbalance to create problems and I run on both nearly all the time.

[schu]

Thanks for the feedback.

[svyolo]

I mentioned this in a thread a couple years ago, but small header tanks are commercially available and usually inexpensive. I think I paid 75 or 80 dollars for mine. Their are fancier ones that cost several hundred dollars. They are usually called "surge" tanks, and sometimes "swirl pots". They come in various sizes and shapes. They are used for exactly the purpose I am using it for - to convert a carberated fuel system to EFI, in cars, boats, etc. The slickest ones have internal pumps, but I couldn't get the right size/orientation to fit. Someday maybe I will custom make one.

The type I have, which is common, has ports on just about every side to make installation easy. There is a port on the bottom which I have a fuel strainer installed in to remove water.

I had bungs welded on (my confidence in my Al TIG skills is low) to the tanks on the inboard side, middle, and cross vent the tanks thru that. I know the outboard end of the tanks was preferable, but with only 1 degree of dihedral, I thought the inboard end was good enough. I vented to header tank to a T in this cross vent.

[kestrel]

I had bungs welded on (my confidence in my Al TIG skills is low) to the tanks on the inboard side, middle, and cross vent the tanks thru that. I know the outboard end of the tanks was preferable, but with only 1 degree of dihedral, I thought the inboard end was good enough. I vented to header tank to a T in this cross vent.

I may not be following your description correctly, but I think you have a setup that will cross feed and overflow the low wing tank with full tanks when not sitting level. If I am correct, consider routing the cross feed above the fuselage (yes, up in the air flow, almost like an antenna) and then back down to create a high point between the tanks. Another option might be a shutoff valve in the cross vent. ...but I'd worry about forgetting to open that one if turned off.

Cross vents that tap into the outboard of the tank will not cross feed because one end of the vent will be at the high end of the high tank so the fuel will not run downhill to the low tank.

[schu]

A shutoff valve that closes when power is removed? What could possibly go wrong?

[svyolo]

I couldn't find a way to do an outboard vent that I was happy with. I think the best solution going forward, would be for the factory to install an internal vent line that vents from the outboard end, but terminates at a fitting on the inboard end of the tank.

[kestrel]

> internal vent line that vents from the outboard end, but terminates at a fitting on the inboard end of the tank.

That is what our RV-4 tanks do.

[Nev]

If having any port unport creates a problem then using BOTH is a liability when low on fuel as you have more ports with less fuel over them.

I just wanted to clarify the fuel system on MY Bearhawk is a Bendix returnless FI. As far as I can see, and according to the tests I did by running a tank completely dry, mine will run fine if only one port is covered by fuel. What it boils down to is so long as one port is covered, the fuel will flood the inlet port of the pump under gravity. Battson clarifies the logic to using BOTH in this case in his lost above. That said, while there are unanswered questions around the fuel system, I still choose to keep one hour for reserves wherever possible.

However with a high flow return system if I understand your thoughts correctly, we aren't sure if one or two ports are sufficient to keep up with the pump demand and prevent cavitation etc. In my mind there's also a question mark over that area.

[Nev]

I had read another thread about this concern for un-porting which caused me to wonder and do the test. When I got the pump to suck air the fuel pressure dropped to zero. I expected the engine to quit but it kept going.

Ed, I did a test where I ran the left tank completely dry until the engine quit. It took much longer than I expected, and it didn't quit all at once, rather it surged uncomfortably for several minutes first. Clearly there was still some fuel sloshing around in there, and I flew with the ball out to the right to keep what little fuel was left over the ports until it was completely empty. Then I selected BOTH and it started very quickly. I repeated this, checking that it would restart in R, etc, in a steep climb, ball out to the R in BOTH etc etc. The only thing I couldn't satisfactorily replicate was to unport the Right tank, with no fuel in the Left. I ended up with 25 litres remaining in the Right, and wasn't comfortable at that stage flying with less. I hypothesised that at some point it would be possible to have the remaining fuel in the outboard section of the tank, and stop the engine that way.

I performed this at 8500ft above my local airport to give several minutes to sort out any issues if needed. Actually it wasn't needed, but I would have been very comfortable making a glide approach to land on a grass runway. Each time the Engine stopped it was predicated by a FUEL LOW PRESSURE alert giving about 5 seconds warning.

I've often thought it would be very interesting to see the results of the same test on an engine fitted with a return to tank system, when completely unported. It would also give peace of mind. If you're ever interested, feel free to contact me and I can tell you specifics of how I went about mine, what precautions I took, and a few small things I'd do differently if I did it again.

[Ed.Meyer]

You took the test further than I did with the same results it seems. With my return system I imagine there is somewhat more fuel flowing through the pump but gravity flow from one side was enough to keep the pump satisfied. Thanks for sharing.

[schu]

However with a high flow return system if I understand your thoughts correctly, we aren't sure if one or two ports are sufficient to keep up with the pump demand and prevent cavitation etc. In my mind there's also a question mark over that area.

My thoughts are simple:

1. We really don't know exactly how it works in all cases because we aren't fluid design engineers with the ability to test every aspect of this.

2. When using a returnless FI or carb it appears that having any port covered will keep the engine running as long as it's not full throttle as the fuel system has enough gravity to keep the inlet flooded.

3. When using return FI or EFI it appears that any port uncovered could cause it to suck air because the stock bearhawk fuel system can't flood the pump inlet with gravity fast enough. A header tank should in theory solve this.

4. When using gravity, venting is super important. With the stock bearhawk fuel tank setup you don't have two tanks that act as one, you have two tanks that are Y'd into the same pump inlet. If you install the cross vent then you have two tanks that act as one (according to the FAA.)

5. BOTH fuel selector valves give you more ports, but less fuel over those ports. I haven't seen any evidence that 5 gallons of fuel over two ports in both tanks Y'd together is better than 10 gallons of fuel over 2 ports in one tank. I think you and Jono theorize that BOTH is better, I remain unconvinced especially when there isn't a cross vent. (Note: If you are doing a massive slip with less than 30 minutes of fuel and pick the wrong tank, all bets are off.)

6. If you are on BOTH and get some air or vapor in the line then you don't have another tank to switch to as both are inop at that point. If you are on the left tank and get a sputter you can in theory switch to the right which should have a large column of fuel ready to flood the pump. Many old Cessna 170's and 172's are placarded single tank operation at cruise for this reason.

7. It appears at least two fuel starvation issues were when tanks had very different amounts of fuel in them and both was selected and they didn't have cross vents which seems to point to the system not completely acting like a single tank in all situations.

Like many things in aviation, flying at the edge of the design or envelope is where you find most problems. I'm pretty sure anything people have done will work with full tanks, it's when you switch to both with 1 gallon in one and 10 in the other or do a huge slip with a spits worth of fuel where you find your problems, which means that many issues could be there but never found because people won't fly below 1/4 tank.

[schu]

Ed, I did a test where I ran the left tank completely dry until the engine quit. It took much longer than I expected, and it didn't quit all at once, rather it surged uncomfortably for several minutes first. Clearly there was still some fuel sloshing around in there, and I flew with the ball out to the right to keep what little fuel was left over the ports until it was completely empty. Then I selected BOTH and it started very quickly. I repeated this, checking that it would restart in R, etc, in a steep climb, ball out to the R in BOTH etc etc. The only thing I couldn't satisfactorily replicate was to unport the Right tank, with no fuel in the Left. I ended up with 25 litres remaining in the Right, and wasn't comfortable at that stage flying with less. I hypothesised that at some point it would be possible to have the remaining fuel in the outboard section of the tank, and stop the engine that way.

I performed this at 8500ft above my local airport to give several minutes to sort out any issues if needed. Actually it wasn't needed, but I would have been very comfortable making a glide approach to land on a grass runway. Each time the Engine stopped it was predicated by a FUEL LOW PRESSURE alert giving about 5 seconds warning.

I've often thought it would be very interesting to see the results of the same test on an engine fitted with a return to tank system, when completely unported. It would also give peace of mind. If you're ever interested, feel free to contact me and I can tell you specifics of how I went about mine, what precautions I took, and a few small things I'd do differently if I did it again.

This is very good testing. I hope to do the same when I'm flying to validate my system which is a little different than yours.

This is a solid data point that says that mechanical FI is just fine with the stock system as long as the vents are working and caps installed into the wind.

Can you tell us if you were using a boost pump in these tests, where you boost pump is mounted, and where the fuel pressure sensor is?

[Nev]

Can you tell us if you were using a boost pump in these tests, where you boost pump is mounted, and where the fuel pressure sensor is?

I tested it with and without the boost pump running. The boost pump sits immediately downstream of the fuel selector. If I remember correctly the fuel pressure sender is tapped off the mechanical pump. I was also able to climb at full power with the left tank completely empty and 25 litres remaining in the right, boost pump off.

[Battson]

2. When using a returnless FI or carb it appears that having any port covered will keep the engine running as long as it's not full throttle as the fuel system has enough gravity to keep the inlet flooded.

I don't think you need to full throttle caveat, well - not in all cases anyway, if someone designed the mother of all fuel systems, then maybe they could run into a problem.
I know that mine will run at full throttle on just one port. Fuel pressure might not be great if the electric boost pump is off, but it runs smooth and I've never had a stutter.

I don't think the fullness of the tank makes a material difference in terms of gravity pressure head. All the head comes from the fuel lines, not from the tank.