Header tank design and location

[JimParker256]

Personally, I don't think anything I've said should (or even "could") be construed as arguing against Zzz's desire for an EFI engine versus a carbureted one. EFI is more efficient, burns less fuel, can often be run LOP for even more fuel efficiency – though sometimes requiring tuned injectors for LOP operation. I have personally experienced some issues with difficulty re-starting at hot Lycoming engine, due to vapor lock following a "quick-turnaround" fuel stop. But that was with the stock Lycoming fuel injection system – not the one Zzz is talking about using. Even the "stock" Continental fuel injection design uses a fuel return line (as do several aftermarket EFI vendors), and there is apparently far less likelihood of vapor-lock in those engines that circulate excess fuel back to the tank(s). Turning on the fuel boost pump for a few seconds pushes cooler fuel through the lines, displacing the "gas" bubbles (vaporized fuel) from those lines pretty quickly. It's a far more elegant and pilot-friendly design than Lycoming's...

But if I go the EFI route, with a goal of not pushing the design envelope while adding a fuel return line (or lines to both tanks with a duplexing fuel valve) would be to use both fuel pickup points as Bob designed them to deliver the fuel to the fuel pump(s), and then add a 3rd port to the tank for the return line, as someone else discussed earlier in this thread. That would seem to eliminate any possibility of unporting a fuel pickup in pretty much any "normal" attitude, while also providing a port for the returned fuel. Yeah, you're going to have to run a 3rd fuel line, but that seems like small potatoes to me...

But let's assume I really didn't want to run that 3rd line for return fuel... Again using information gathered at the OSH seminar on fuel system design, it would seem reasonable to use a header tank of sufficient size to support a full throttle run for a reasonable period. And if I do that, the return fuel could go to the header tank. It doesn't take a LOT of fuel to cool the returning fuel – even if it contained "vapor lock" bubbles... A few gallons would be plenty to cool it below vapor lock state. The sizing of the header tank is more about the fuel demands of the engine than solving the vapor lock issue...

Personally, I think the additional weight of a header tank would likely offset the additional weight of running a third fuel return line, so I probably would not use a header tank, due to the increased risk and maintenance requirements. But that is an admittedly personal bias against carrying fuel inside the fuselage – or at least any more than HAS to be there because of the fuel lines... (I've read way too many NTSB reports about post-accident fires caused by ruptured fuselage and/or header tanks. If I put one in my plane, it would have to be pretty stout!)

But even with a header tank, the flow from the wings to the header tank is likely to be gravity feed, with no intermediate pumps. The electric boost pump is used to supply sufficient pressure to keep the engine running in the event the "other" pump (which may be engine-driven or electric) quits, right? So that pump would be located downstream from the header tank, between the header and the engine. The fuel lines would still need to be able to supply sufficient volume of fuel to the header tanks to sustain an engine running at full power. If it cannot, you have to add a "boost pump" before the header tank (even more complexity and maintenance issues).

I like Bob's KISS fuel system design... But then, I'll probably wind up using a "boring" carburetor for cost and maintenance reasons... LOL

[Ed.Meyer]

There is some pretty good discussion of this in Glastar service bulletin 43. Apparently they use only rear tank pickups and mandate two small header tanks to preclude air getting into fuel feeds.

[Zzz]

Let's consider this: One of the areas where experimental amateur built aircraft fall significantly short of their certified counterparts is in the category of "Fuel Starvation" related incidents. That's where the aircraft had fuel available, but the engine quit operating because the fuel could not get to the engine. Some of these fall into the "stupid pilot tricks" category, where fuel tank selected by the pilot was empty when the engine quit. But a not-insignificant number of them occurred because even though there was fuel in the tanks, there was no way that fuel could possibly be delivered to the engine. These typically get classified as "pilot error", but should really be classified as "builder error" (or in rare cases, "designer error"). In the came of all three Bearhawk designs, Bob has designed a fuel delivery system that eliminates the "designer error" part of the equation, leaving it up to us as builders to either follow his design or go design our own system. If we opt to deviate from Bob's design, we're effectively designing our own system. And unless we're professional engineers competent to design our own fuel systems, we probably should tread cautiously in this area...

For a CERTIFIED airplane that uses a fuel pump to deliver the fuel to the engine, the manufacturer is REQUIRED by 14 CFR 23.995 to demonstrate that the fuel system can deliver fuel at a rate equal to 125% of the engine's maximum full-throttle fuel consumption. Gravity-only fuel systems (with neither an electric boost pump or an engine-driven fuel pump) would need to be able to flow 150% of the max full-throttle fuel consumption.

Both "pump" and "gravity" types are required to demonstrate they can meet the stated fuel flow capability while in the most critical flight attitude for that aircraft design, and the testing must be done with the tanks at very low "fill level" – only the "unusable" fuel plus whatever is needed to perform the test. My reference for these statements is 14 CFR 23.995, which can be found at https://www.law.cornell.edu/cfr/text/14/23.955.

While that requirement does NOT technically apply to Experimental Amateur Built aircraft, I think we would all agree that this fuel flow test is a really good idea. (And as a side note, my local DAR will NOT sign off an EAB aircraft that has not had this test performed.) You'll also note that the requirement to perform the test at the most critical aircraft attitude would generally mean that the airplane would be in a very nose-high attitude, since max-performance takeoffs are generally the most critical situations for fuel delivery.

But if you're leaving off the forward pickup point for the fuel tanks, I would suggest that in addition to the nose-high attitude test, you might want to consider that your actual "most-critical attitude" might well be a nose-down attitude approximating the final approach profile. That's because a go-around would be initiated from that attitude, and if you're been in a nose-low descent for a while, it is possible that the aft fuel pickup point has been "unported" for however long that nose-low flight has gone on. Translation: The fuel line from the aft pickup point to the fuel pump could be bone dry when your engine quits due to fuel starvation. You might have sufficient fuel on board to be able to restart, but if you're flying using "Both" tanks, it's entirely possible that both lines could be bone dry. How long would it take to refill those lines so the engine could be restarted? If you're on final approach when this happens, would you have enough altitude remaining to maintain a level (or better yet, nose-up) attitude so the aft pickup point would again be awash in fuel and begin to fill the line?

So, far from being a misunderstanding about differences between FI and Carbureted engine fuel delivery systems, my concerns are about whether or not your altered fuel system design can still safely meet the criteria of 14 CFR 23.995, regardless of the fact that it may not be required to do so... We kind of like having you around, Zane, and don't want to read about about you in an NTSB report!

Nobody wants to not die more than me, but a life lived with a boring carburetor and nothing to provoke old EAA guys is a slow death of sorts. If you guys think that making modifications automatically means eschewing FAR 23.955, or testing, that MUST be entertaining and get the cushion sweaty.

I'm not pushing the envelope of design. Several have come before me implementing similar designs, many RV guys. Even the beloved old 170 only has a single mid-tank pickup. I guess I've never flown one low enough on fuel to make it an issue.

My design of rear port only is predicated on avoiding unporting the front pickup. The fuel system will be circulating fuel at 35-45 GPH. What happens when either of the pickups, that are T-ed or Y-ed together downstream, is unported? Will it suck air? If you don't line the word "suck" to describe fuel or air filling the void behind the supply side of a high pressure fuel pump, please suggest another term. It does create a lower pressure region, compressible or incompressible, right?

So, a few scenarios:

1. For some reason my imagination has failed me for the better and an unported front(or rear) pickup doesn't actually "suck" air. Everything remains primed nicely, the pump never cavitates.

2. Air DOES get pulled into the supply line upon unporting or one of the fuel tank pickups. Fuel pump manages to push it through into the fuel rail and the engine dies or runs like crap long enough to vent the air through the injectors, if it even will. I've heard reports where it was a non-event.

3. The air is circulated back the tank quickly via the return lines, and is removed from the system. Or the fuel pressure regulator purges it through a bypass.

As for in-flight fuel starvation, yeah, it sucks to unport front or rear pickups in any pitch attitude. But climbing steeply and maybe slowly is when I'd least like it to happen. it would suck in a descent too but usually you're carrying an abundance of airspeed. And for approaches,I don't approach any runway or airstrip pitched down aggressively. But one shouldn't have to use any flying technique to compensate for engineering flaws.

Any situation where you have unusable fuel is undesirable, though, I agree. I have a few emails out to other high wing EFII builders to see if they've intelligently mitigated the risk of unporting. But as far as I know, you either go header tank or full return with single tank pickup.

[AZBearhawk272]

Just to fan some air across the coals of this
thread.

There had been a comment about pumps and sucking a while back that needs some clarity.

Most positive displacement pumps are not
capable of operating successfully without a flooded or pressurized inlet. They do not suck well and basically quit pumping once they
cavitate or drop below vapor pressure at the inlet. If pumps sucked well every water well with a submersible at the bottom of the hole
would be unnecessary.
The typical engine driven diaphragm pump is designed to suck but the pressure capability
is minimal and the durability questionable.

With respect to feeding fuel from the tanks to
the engine. Half the conversation is unporting a line feeding the engine.
Liquids are funny things, they only go where you guide them. Many different factors determine how they get down the pipe.
one thing is sure. After unporting a feed line,
air has to come out for fuel to go in.
Hence the dual feeds from each tank.
The line with the highest static pressure becomes the feed, the other becomes the vent. No vent, fuel does not like to flow down where air is heading up.

[JimParker256]

Let's consider this: One of the areas where experimental amateur built aircraft fall significantly short of their certified counterparts is in the category of "Fuel Starvation" related incidents. That's where the aircraft had fuel available, but the engine quit operating because the fuel could not get to the engine. Some of these fall into the "stupid pilot tricks" category, where fuel tank selected by the pilot was empty when the engine quit. But a not-insignificant number of them occurred because even though there was fuel in the tanks, there was no way that fuel could possibly be delivered to the engine. These typically get classified as "pilot error", but should really be classified as "builder error" (or in rare cases, "designer error"). In the came of all three Bearhawk designs, Bob has designed a fuel delivery system that eliminates the "designer error" part of the equation, leaving it up to us as builders to either follow his design or go design our own system. If we opt to deviate from Bob's design, we're effectively designing our own system. And unless we're professional engineers competent to design our own fuel systems, we probably should tread cautiously in this area...

For a CERTIFIED airplane that uses a fuel pump to deliver the fuel to the engine, the manufacturer is REQUIRED by 14 CFR 23.995 to demonstrate that the fuel system can deliver fuel at a rate equal to 125% of the engine's maximum full-throttle fuel consumption. Gravity-only fuel systems (with neither an electric boost pump or an engine-driven fuel pump) would need to be able to flow 150% of the max full-throttle fuel consumption.

Both "pump" and "gravity" types are required to demonstrate they can meet the stated fuel flow capability while in the most critical flight attitude for that aircraft design, and the testing must be done with the tanks at very low "fill level" – only the "unusable" fuel plus whatever is needed to perform the test. My reference for these statements is 14 CFR 23.995, which can be found at https://www.law.cornell.edu/cfr/text/14/23.955.

While that requirement does NOT technically apply to Experimental Amateur Built aircraft, I think we would all agree that this fuel flow test is a really good idea. (And as a side note, my local DAR will NOT sign off an EAB aircraft that has not had this test performed.) You'll also note that the requirement to perform the test at the most critical aircraft attitude would generally mean that the airplane would be in a very nose-high attitude, since max-performance takeoffs are generally the most critical situations for fuel delivery.

But if you're leaving off the forward pickup point for the fuel tanks, I would suggest that in addition to the nose-high attitude test, you might want to consider that your actual "most-critical attitude" might well be a nose-down attitude approximating the final approach profile. That's because a go-around would be initiated from that attitude, and if you're been in a nose-low descent for a while, it is possible that the aft fuel pickup point has been "unported" for however long that nose-low flight has gone on. Translation: The fuel line from the aft pickup point to the fuel pump could be bone dry when your engine quits due to fuel starvation. You might have sufficient fuel on board to be able to restart, but if you're flying using "Both" tanks, it's entirely possible that both lines could be bone dry. How long would it take to refill those lines so the engine could be restarted? If you're on final approach when this happens, would you have enough altitude remaining to maintain a level (or better yet, nose-up) attitude so the aft pickup point would again be awash in fuel and begin to fill the line?

So, far from being a misunderstanding about differences between FI and Carbureted engine fuel delivery systems, my concerns are about whether or not your altered fuel system design can still safely meet the criteria of 14 CFR 23.995, regardless of the fact that it may not be required to do so... We kind of like having you around, Zane, and don't want to read about about you in an NTSB report!

[JimParker256]

Edited to move it to a "stand-alone" comment for better thread flow...

[svyolo]

Any fuel system, carb, MFI, or EFI, needs to be "fed" fuel. Since Bob's design works, I will start with that. Downstream of that, fuel will enter the fuel pump to be pressurized. But any fuel system needs to be reliably "fed". No type of fuel injection changes that.

There will only be a short section of pressurized fuel line aft of the firewall. Maybe 18 inches, and 2 fittings. The rest of the pressurized fuel system will be ahead of the firewall.

[PaulSA]

I'm with you, Mark. With 2 pick-ups per tank I really don't see the need to modify anything. Ok, with a FI system, you need to put in a secondary electric pump but the basic system is sound. I know it is called "Experimental" for a reason, but highly complex FI systems that require dual batteries and return lines etc don't really - in my mind- seem to have a place in what, is after all, a bush aeroplane........

Of course, in the end, it's your aeroplane so do whatever you want.

[Ed.Meyer]

Yes. SPRL 4. It is possible that it returns more to one side than the other when set to both. If so, it has not been an issue.

[Mark Goldberg]

Aircraft designers generally do not like homebuilders modifying the design. While some things are obviously not going to effect safety or airworthiness, nothing is usually more serious to "redesign" than the fuel system. While I know that some you builders posting in this thread are quite smart and capable people, I still urge caution changing the basic fuel system from what Bob designed - which has worked and proven safe. If you do, then you become the test pilot. Mark

[Zzz]

The duplex valve we have does have a 'both' setting for supply and return. Fly set to both most of the time.

It's certainly possible, I was advised to not do it because it's not reliably metered between the tanks, and you can fill one more than the other. You have the SPRL?

[Ed.Meyer]

The duplex valve we have does have a 'both' setting for supply and return. Fly set to both most of the time.

[Zzz]

This has been a funny discussion, kind of like "Who's on first." We're not talking about the same things.

EFII systems are much more akin to car electronic fuel injection, which utilize a pressurized fuel rail to provide supply to the electric injectors. If you decide to go down this road, you have to throw what you know about the Bob fuel system out the window. Like schu said, it is great for a carb or mechanical drip fuel injection. But when you go with a system that utilizes high pressure fuel pumps to pressurize your fuel rail to 35 psi, there is no consideration of "will my engine continue to run on gravity feed alone?" It will not. So why worry about it? At that point you focus on the reliability and failover of your electric pumps.

The fuel return is mostly to keep fuel cool that is running over the top of, or along side hot cylinders. You could also use a header tank to accomplish that, but you need a minimum volume tank to effect the cooling. I have chosen to return to the tanks because I just don't want to use a header tank in my aircraft.

You cannot return fuel to both tanks at once using a BOTH selector. You have to use a duplex valve and operate on LEFT or RIGHT; supply and return are set to one tank at a time. The header tank can get you around this limitation, so that's a plus for the header tank. But if you've flown old Pipers or Bonanzas, you get used to managing your tanks. No big deal.

Those of us choosing to go with EFII aren't modifying our fuel systems just for the hell of it. I would stick with the Bob design if I was using a carburetor. But if you commit to EFII, it is a different approach and so you modify the fuel system to support the paradigm of electronic everything. The guidance on the fuel system design has come more from the system engineers, and is the same for RVs, Cubs, etc.

To Kevin D: I'm not sure I understand your claim that an electric fuel pump doesn't "suck." ? With fuel upstream and downstream of the pump regulated to 35 psi, how can you avoid drawing fuel at an equivalent flow rate? There's always going to be a positive pressure, given the tanks are above the pumps (in the Bearhawk), so it's not fighting gravity.

[Gerhard Rieger]

I chose to have Left, Off, Right. With the FI system And my header tank with two 1/2 “ vents back to the wing tanks. I placed one vent on the rear left and the other front right., no matter what attitude your in you always have a vent facing up for any air to escape back to the main tanks. The Header tank will fill up super fast with the two 1/2 vents. No matter if you slip for an extended period with low fuel it will not cause fuel starvation or air to enter the FI system. I only added the header tank and two 1/2 “ vents to what Bob shows on the plans.

[jim.mclaughlin924]

I welded in an extra bung for the return line.near the top of the tank.