Returnless FI?

[svyolo]

Here is a picture of AFP's boost pump. It is rated at 32 gph, at 30psi. In their manual it states it is rated for continuous duty. At the pump outlet, there is a non-adjustable fuel pressure regulator. Fuel is fed, and pressure is maintained at 30 psi, by returning unused fuel to the pump inlet, as I believe Dave wants to do. EFII's boost pump looks pretty similar. Andair's is quite a bit different. It is British, after all.

AFP's docs also state that if the fuel pickup ever becomes un-ported, or a tank is run dry, neither the boost pump or engine driven pump would be able to create enough pressure to open the valve at the fuel divider. I guess that is engineer/lawyer speak for "your engine will cease functioning as a propulsion device."

[Archer39J]

I see AFP put gascolators and pumps fwf in schematic #2, huh. I don't know much about fuel dividers, but why wouldn't a pump be able to re prime the system?

[Archer39J]

Ah, from their manual it looks like it's just an issue on pumps that are pulling fuel UP from a tank.

[svyolo]

Dave, I am glad you started this thread. I think I have almost talked myself into doing a return-less system, even though there seems to be consensus amongst fuel system experts that if you CAN return, it is better.

The BH fuel system has many thousands of hours gravity feeding the inlet side of a pump in a return-less fuel system, feeding a Lycoming engine. The parts downstream of that, shouldn't matter to the aircrafts' fuel system.

Fuel heating ins't supposed to be an issue with SDS's fuel system, and they run at a much higher pressure than the Bendix, which helps with vapors.

[Archer39J]

And I almost agree with them too, but then if a system is unnecessary it's just added failure points, cost, and weight. To reiterate your earlier point doing a header or pump return is sticking with a stock BH fuel system as far as flow is concerned, if I were to do a full return I'd want better flow of 1/2" lines like whee did. The high pressures pretty significantly preclude vapor lock, and that was with high-RVP winter E10 data from the EPA. My goal is to run E10 (auto), regular too, but then have the option to switch to other fuels on the road

[svyolo]

I have been trying to understand the Bendix style stuff, and am surprised how complicated it is, even the flow divider. It is more complex than just a manifold. I found a link to TCM's version of FI, and their engine driven pump looks a little more robust as far as removing air. There is a chamber in the pump to eliminate vapors.

This also kind of got me thinking of just creating a little manifold, right before the inlet of the electric pump, to separate air, and vent it back to the tanks. It wouldn't have to be much. Half a Pint. Tall and skinny with a 1/4 inch or smaller vent at the top. An upside down gascolator might do it, if you could replace the drain filter with a AN4 male fitting.

Return fuel from wherever you want to to just upstream of this little manifold. If you returned it from the fuel block, the whole thing would be self bleeding on first start up. Or just from the outlet from the pump. But, then I am adding complexity back in.

[Bcone1381]

I'm uncomfortable. This thread is starting to scare me, but I don't know enough to give input. At work I've been trained to ask "Does everything look right and feel right?" to my colleagues when we've been rushed or weird stuff happens and we get out of our routine.

Things are starting to feel "Not quite right" to me. I think its lack of data, abundance of unanswered questions, mixed in with a misunderstanding I'm seeing here and there.

[robcaldwell]

You're not alone. Been having the same feeling since this thread initiated. I could say more, but I will just leave it at that...

[Archer39J]

I've noticed a lot of misunderstanding myself actually too, and you're right it is unnerving...

Not sure what feelings have to do with designing aircraft systems, I'll stick to engineering and testing with that.

I'd say if you don't know enough to give input, maybe ask questions first before criticizing?

[svyolo]

I have read enough stuff in the last few days that I am actually getting quite comfortable, much more than before. There is probably nothing that hasn't been tried. Folks with Bendix FI injection have created a return line, some have dumped the engine driven pump for 2 electric ones. Folks with EFI have used an engine driven pump in the belief that these are more reliable. Some folks return to the tank, some to the inlet side of the fuel pump. Some have used a T, with a tiny orifice, to return a tiny amount of fuel to get rid of fuel vapor. Some plumbed them to the fuel tank vents.

Below is cut and pasted. It is a great description.

"What most folks refer to as "Vapour lock" is not vapour lock at all. VL is when the fuel pump sucks on the fuel (that has been warmed and/or is at a low enough pressure) causing it to boil at the inlet of the pump. This is relatively deadly...I.e it can kill you! This is most likely to happen whenever the fuel is hot, at low pressure (high altitude) and is being sucked upon. Autofuel is even worse because it has a higher vapour pressure which means it is more prone to boiling.

Without getting too involved in this, think about the mechanical fuel pump location.....In a hot cowl. and sucking from the tank...Not good!..Thats why my fuel pumps are in the wingroots which is much cooler than the cowl and its not sucking a long way from the tanks either. Mind you with a 3/8ths line you could suck it a afair distance without causing an issue. Being an engineer that designs pumping systems for high VP liquids I wanted them in the hydraulically correct place.

Anyway, lots of folks call localised boiling in the injector lines as vapour lock....This is not correct. The only reason it happens is when you shut down the fuel lines get heat soaked and because the pump is not making pressure (hot fuel won't vapourise when its under significant pressure) the fuel boils...When you got to start the engine has to be cranked to pump all the hoses free of vapour and get wet fuel in there. Hence the classic hot start...Crank crank crank..."

By my reading this, the worst possible combination is an engine driven fuel pump, suction feeding. The scenario that sounded the worst was flying along, pull the power back, rpm drops, fuel pressure drops from the engine pump, and you get vapor lock that stops the engine.

Almost all newer cars run return-less. Most Lycoming using Bendix style FI run return-less. AFP includes a purge valve to return for 15-20 seconds for hot starting. After start, it runs return-less. TCM FI, by the diagram I saw, returns fuel from the servo (not the divider) to the pump, and the pump has a vapor return to the tanks (or somewhere downstream). After the servo, it looks return-less to the divider and injectors.

AFP, EFII, and Andair boost pumps all bypass and return fuel from the pump outlet, to the pump inlet. At least AFP rates theirs for continuous operation. Not sure about the other 2.

EFI has nothing between the pump outlet and the fuel injectors. 15 minutes after a hot shutdown, you might get some vapors on top of the engine, like any engine. Turn the electric pump on, and the 45 psi slams the vapors right back into suspension. No more fuel vapors. The pumps are mounted aft of the firewall. On startup they get cool fuel, gravity fed, at the pump inlet. No vapor.

The hot start problem is a mechanical FI problem. The fuel pump and servo are mounted on the engine, and get hot. The fuel in them gets hot after shutdown, and vapors form.

If I was going to use Bendix style FI, I think I would use all electric pumps as well. No hot start issues. No boost pump needed. It is cheaper than a boost pump plus an engine driven pump. Electric pumps might last longer, and are far cheaper to replace. It would be easier to start, easier to operate, and cheaper to maintain.

Everything that we can think of has already been tried.

[Archer39J]

Yeah if anything this has shown I'm quite unoriginal, this has already been done before. We're just making sure it doesn't over tax the BH system.

[zkelley2]

The only reason there is a mechanical pump is because it's far more redundant to have a mechanical and electric than two electric. From a pump performance and cost standpoint electric is superior, but it's not redundant in the same way different drive mechanisms is. From an aviation engineering perspective, this is how things are normally designed to be safe. Not just 2 pumps(whether it's fuel, pneumatic, hydraulics) but also different ways of powering them.

Now the hot start thing. I have quite a few thousand hours behind Bendix injection. Mostly with turbos under the cowls as well for extra heat. If you understand the system and how to correctly start it, you'll never have a hot start. It's not a thing(It is a thing, but you shouldn't experience it). It's always operator error. I'll even go as far as to say Bendix injection will always start, down to temperatures you should be way past pre-heating, where carbs won't even start anymore.
That said, I don't personally see the benefit of removing redundancy from a system(2 electric pumps in lieu of 1 mech and 1 electric) for a flaw that is purely in the operator, not the equipment.

[svyolo]

Bendix is a proven and tested system. No arguments. On a light aircraft, with limited electrical redundancy, I would say that 1 engine driven and 1 electrical pump offers more redundancy in some ways, mostly while airborne.

Almost everything I have flown has 2 or 3 of everything, if not more. The 2 or 3 were all identical, but the electrical redundancy/separation was far greater.

I agree with the operator error part. But deep into 5 digits in flight time, operator error is unfortunately part of my DNA. I would rather just turn the key, and have it start. Less for me to screw up.

[Archer39J]

I was hoping there would be a better discussion of the technical reasoning behind using a header tank. But it's apparent that knowledge isn't here.

I'm doing CFD analysis on gascolators and fittings and reading EPA tank design guides for equations on extrapolating TVP from RVP. So "I don't know what you're talking about, but neither do you" is an especially rediculous response from my perspective.

Turns out that a header tank isn't necessary with the SDS system configured as I intend, backed by the fact other systems are configured the same. If you have technical justification for why this isn't the case by all means let's discuss it. Else all these opinions are just that, and mostly uninformed it seems...
​​​​​

​​​​​

[Bcone1381]

I was hoping there would be a better discussion of the technical reasoning behind using a header tank.
​​​​​

The header tank is a choice some designs use, but not all. It may simplify a fuel system on a FI aircraft that requires engine return fuel. On carbureted engines, its used to prevent fuel starvation if a wing tank is un-ported for a period of time. Many taildraggers with little or no wing dihedral and a wing fuel tank with only one fuel pick were manufactured. Even a crowned runway with 1/4 tank can un-port the fuel pick up and starve the engine so the header tank was installed.

Fuel flows into a header tank via gravity at the rate that an engine consumes fuel. It is vented, and so it always remains full until the main tanks are empty. The fuel port out of the header tank to the engine is alway is covered with fuel. The header tank is a good place to plumb excess engine return fuel to.

Cessna 310R, C-185, and Beech Bonanza's use Continental's Fi system which returns excess fuel from the engine back to a tank. The 310 always returns the fuel back to the respective main tank, even when the Aux tank is selected. So the C-310 main tanks fill back up when burning off the Aux tank. The Bonanza returns fuel back to the selected tank via a duplex valve. The C-185 has a header tank that is labeled a "Fuel Accumulator Tank".

The Continental FI system design requires a fuel return because that system pumps excess fuel, like SDS. Having a header tank simplifies the fuel system.
-The header tank is a destination for the engine return fuel and the tank supply fuel,
-The Header tank is always full even if one tank is run dry,
-A simple ON-OFF fuel valve may be used to control engine supply to the engine.
-It allows for a L-BOTH-R fuel selector.
-Multiple return fuel lines do not have to be plumbed back to the wings all the way back to the wings,
-A duplex fuel valve is not required.

The original Super Cub fuel system is gravity fed with a header tank located aft of the baggage area. Each wing tank in a Cub had one fuel line supplying fuel to the header tank. (Bob's design has two, one near the front, one near the rear to allow fuel supply with near empty tanks in a nose low or nose high attitude)

The header tank allowed take-off with 1/4 tanks without worry of unsporting a tank. The C-140 did not have a header tank, but had a placard "Do not take off with less than 1/4 tanks." (I un-ported a C-140 tank on takeoff once and momentarily loss power even though we honored the placard.) A Super Cub STC does away with the header tank if the system is modified to what the Bob Barrows original design is.

So, the technical reason for having a header tank is to maintain a positive supply of fuel to the engine in all normal aircraft states, to simplify fuel system plumbing, and in some cases to have a place for fuel to cool after traveling through an engine compartment.

[svyolo]

The later 172's (R model and later) had IO-360's. Their fuel systems also used a return line to a header tank. The diagram wasn't great, but it looked like the return came from somewhere on the servo. If that is correct then from the servo to divider and injectors is still return-less.

I never flew one, just found it online. I think I have an hour or 2 in a 172 in 1981.

[Archer39J]

So, the technical reason for having a header tank is to maintain a positive supply of fuel to the engine in all normal aircraft states, to simplify fuel system plumbing, and in some cases to have a place for fuel to cool after traveling through an engine compartment.

Except in the BH fuel system, for the reasons I've explained multiple times, these goals can be achieved without a reservoir of fuel. As proven by analysis and real world examples.

You forgot about serving as a place for gasses to collect (but I won't use that as a basis for claiming you don't know what you're talking about), which is really the only concern left going returnless, but again this is how other systems do it.

​​​​

[Bcone1381]

I tried to supply what you asked for in regards to technical discussion of header tanks. I am rooting for you.

[Archer39J]

I appreciate that. It's a decent review of header tank fundamentals for those who don't already know or don't do any research beyond this forum.

[whee]

I'm uncomfortable. This thread is starting to scare me, but I don't know enough to give input.

I know the feeling. Years ago my wife and I were relaxing in a natural hot spring after enjoying a day of winter whitewater kayaking. Two attractive females appeared out of the darkness and asked if they could join us. Of course we said yes. They disrobed and slide into the pool next to me. There I was sitting in a hot pool with 3 attractive and naked women... My engineering professors didn't teach me how to handle such a situation so I leaned over to me wife and said "I'm uncomfortable."

[svyolo]

I think my response would have been "hey honey, what do you think?" But then again, I am divorced.

[Chewie]

Didn't see that coming reading this thread! I should keep going.

[whee]

A thread on another forum lead me to read this article: https://generalaviationnews.com/2011...-gas-is-mogas/

I found the following quote interesting:

Todd Petersen, owner of Petersen Aviation, commented recently on Lycoming’s assertion that autogas airworthiness from the FAA is a matter of luck, not by design:

“I must take issue with Michael Kraft’s statement that using pump gas is achieving airworthiness not by design but by luck. The STC’s were not developed through luck. Detonation and endurance testing was required in order to develop the STC’s in the first place, as well as flight testing for airframe STC’s which make the engine STC’s worthwhile. In certificated airplanes engine approval by itself is, without airframe approval, largely irrelevant.

Auto fuel specifications are indeed designed for ground vehicles, which is why testing was done to approve airplanes to use it. Yes these specifications change frequently. If they changed to the point of creating adverse circumstances then Mr. Kraft’s point would be well taken, but with one exception (ethanol) they have changed only for the better, making autogas (without ethanol) more like avgas than it ever was.

Mr. Kraft also stated the following: “93 AKI is “Super Premium” fuel. That AKI rated octane is what is needed to achieve the exact same power performance on the approved engines as 100LL avgas. 93AKI is produced worldwide, but not necessarily distributed worldwide.”

This statement implies that an engine burning 91 octane doesn’t give the same amount of power as 93 octane. In fact, octane has to do with improving the detonation margins. Power is more a function of BTU’s per gallon. You do not need 93 AKI to achieve the same amount of “power” as on 100LL. Our detonation testing demonstrated that you get about 3% more power on auto fuel than on 100LL because there are more BTU’s per gallon. If it were less, the FAA could never have issued the STC’s.

While limiting volatility to class A-4 may reduce the chances of vapor lock it will not eliminate it, and it will not make a Bendix injection equipped Lycoming safe to operate on mogas or pump gas. I have been personally been involved with hot fuel flight testing of airplanes using Bendix injection systems. Regardless of the type of airplane, it failed. I’ve spoken to owners of Bendix equipped airplane who have it happen frequently on avgas.

Lycoming erred when they approved Bendix injection. I’m sure they feel that it’s an airframe issue and therefore not their problem, and indeed that may be the case — in certificated airplanes. Owners of Experimentals though are using Lycoming engines without any sort of testing of the airframe fuel system against vapor lock. Those guys are test pilots whether they know it or not. Let’s hope that Lycoming makes a change to the injection system which will make their proposed I0-233 safe against vapor lock. If they fail to do so, then it’s Airworthiness by Luck.

I welcome Lycoming’s revised stance on the use of autogas in general. It’s especially refreshing given the repeated attempts by Lycoming throughout the 1980’s to discredit auto fuel STC’s.

Here is another interesting article that talking about the testing necessary to prove a fuel system is safe with E0 mogas:https://generalaviationnews.com/2011...eryone-part-i/ None of us are going to so thoroughly test our fuel systems so as Mr Petersen said in the quote above, we are truly test pilots when we start changing fuels or fuel systems. Makes since why fuel issues are the #1 cause of EAB wrecks.

Except in the BH fuel system, for the reasons I've explained multiple times, these goals can be achieved without a reservoir of fuel. As proven by analysis and real world examples.

There are also real world examples proving the opposite and no doubt someone else could prove by analysis the same. As with any technical analysis, the data we use and assumptions we make affect the result. I'm not saying your wrong but I do think you have your mind made up and are dismissing valid points. However, I am all for others doing exactly what they want to do. I just hope some of them don't fly over my house.

Literally everyone told me I couldn't run my Continental IO360 on mogas with I asked on some forums about it. Everything seemed to check out when I looked into it. It was clear that the detonation margin was going to be close and that vapor formation at the intake of the electric boost pump could be an issue but there was not glaring issue telling me it wouldn't be worth the effort. I called Mr. Petersen, discussed it with him. I called another outfit that currently handles the big bore fuel injected Continental STCs and discussed it with them. I determined that I may have to install pumps in the wings as is required on Cessna 210, 185, 180s when running the FI engines on mogas but I could perform some low risk tests to determine if the pumps were necessary on my installation. So far I've flown 50hrs on mogas. Detonation margins are indeed quite minimal but it is manageable, vapor formation doesn't seem to be an issue but it is something I'm still watching, and unexpectedly I've discovered that my engine runs like crap LOP on mogas.

[Archer39J]

What valid points am I dismissing? Those who insist the BH system may not provide adequate flow despite the fleet history and actual system design? The ones who insist vapor lock can occur despite apparently not having any understand of fluid vapor pressures or bothering to do any analysis themselves? The people who continue to hold up examples of systems that heat soak the engine side fuel lines as reasons why this doesn't work?

[Archer39J]

I feel with your story about your experience trying to discuss your design, thanks for sharing that.

[svyolo]

High wing aircraft have the luxury of gravity keeping the pump intake at a slight, positive pressure. Suction feed is much worse, and much more prone to vapor problems, especially with Mogas.

[svyolo]

Dave, I think you are on to something. AFP has a good writeup of what happens on Bendix FI when you shutdown and try to start up a hot engine. It is a selling point for their purge valve, which is nothing but a switchable return line used for 15-20 seconds. Experimental aircraft only.

I can think of one or two minor issues, but the EFI fuel system has none of the engine heated components of the Bendix. We have the dual benefits of a gravity fed fuel pump inlet, and double the operating pressure of Bendix, after the pump. I can't think of a reason not to do it. A return line is theoretically better. Like we used to say in the military, "If the minimum wasn't good enough, it wouldn't be the minimum." Bendix FI return-less seems to be working perfectly with the BH fuel system. Ours should as well.

Possible issues is the fuel pump gets too hot, and it will be harder to prime the system on either a first start up, and if you ever run the tanks dry. I think the hot pump might only be an issue on the ground, at idle. Once you are burning 7-15 gph, there should be enough cool fuel going through the pump to keep the pump cool.

[Archer39J]

It sure looks like it from what I can determine, testing will of course be the proof. A purge valve... now there's a brute force solution lol, but maybe the only one for that application.

Plenty of RVs have trouble with vapor on the ground with stock Bendix, a couple guys had to take out their engine pumps to run E10 and are still plagued with vapor issues. Glad I'm going this route.

[schu]

Dave, I think you are on to something. AFP has a good writeup of what happens on Bendix FI when you shutdown and try to start up a hot engine. It is a selling point for their purge valve, which is nothing but a switchable return line used for 15-20 seconds. Experimental aircraft only.

I can think of one or two minor issues, but the EFI fuel system has none of the engine heated components of the Bendix. We have the dual benefits of a gravity fed fuel pump inlet, and double the operating pressure of Bendix, after the pump. I can't think of a reason not to do it. A return line is theoretically better. Like we used to say in the military, "If the minimum wasn't good enough, it wouldn't be the minimum." Bendix FI return-less seems to be working perfectly with the BH fuel system. Ours should as well.

Possible issues is the fuel pump gets too hot, and it will be harder to prime the system on either a first start up, and if you ever run the tanks dry. I think the hot pump might only be an issue on the ground, at idle. Once you are burning 7-15 gph, there should be enough cool fuel going through the pump to keep the pump cool.

There is a lot of comparison between the bendix system system, the SDS system, a returnless system on a car, and the system proposed, but they are not the same, and saying that it works for bendix so it should for us is dangerous. Here is why:

1. The bendix system is a true returnless system. The fuel isn't bled off somewhere and returned to another part of the system. Thus, the entire fuel system past the pump is pressurized. What is being proposed here is not a returnless system, it's a return system that bleeds back to before the pump. The concern is that fuel will get really hot being cycled around, then that hot fuel will be fed to a very low-pressure point right before the pump which could cause cavitation.

2. The SDS system is a true return system that returns to the tank. Given the volume of the tank, it won't have hot fuel at the inlet of the pump.

3. The returnless system in a car is a true returnless system where fuel isn't fed back to a pump. Instead, the pump is driven by a pump controller that cycles or throttles the pump based on a fuel pressure sensor. Lots of people in the high-performance spaces convert to a return fuel system because a bleed off regulator is generally more consistent than the feedback loop of a pump sensor and controller which will always have some lag.

If you both do build this, then please:

1. Test to see how hot fuel gets running just the pump on the ground for a few hours.

2. Test to see how hot the fuel can get before the pump cavitates.

3. Test to see if lost pressure results in boiling fuel, and if so, how long does it take to hot restart. If it's not instant then consider what could happen in air.

4. First flights on 100LL.

5. Introducing junk fuel (that's pretty much the only reason to do this) at altitude only and from a dedicated tank for testing. IE, run 100LL in the left tank and mogas in the right, and only a few mistakes high until you absolutely know what's going to happen.

I can think of more tests, but you you get the idea. You need to totally understand what is going to happen and be ready for the engine to quit.

schu

[Archer39J]

Let’s try not to misrepresent what’s been discussed, that’s happened too much already. We understand they’re not the same, obviously, the comparisons only relate to the aspects of the systems that are the same and how ours is an improvement from a heat/vaporization standpoint. You can see where we’ve been discussing the drawbacks of having a return, like recirculating trapped/sucked gasses, which I’ve always understood wasn’t on the Bendix systems.

Tests will be performed commensurate with an understanding of the pressures and properties of the fluids involved. “Junk fuel” means nothing from a design standpoint.