Return fuel lines

F

FFTravism

Well-known member
For those that ran return fuel lines to the tanks for fuel injection did you run the lines up the front at the window or in the rear? I can see benefits to both. Front is defiantly shorter and less tubing. The rear location would be potentially less visible.
Thanks
 
Disclaimer - I didn't run return lines, but I did run a lot of stuff down the front door post and wanted it to be out of sight and well protected.

I just used offcut aluminium from the kit to make these box sections you see in the image. They are one piece and slip over the door post, very rigid and strong once secured to the window frame tabs. It also makes a tidy finish for the window and the door seals.

As pictured, the screws aren't installed to hold the frames to the window. I used nut plates on the window tabs, with short truss-head screws.
20130608_163150.jpg - Click image for larger version Name: 20130608_163150.jpg Views: 0 Size: 1.03 MB ID: 81760
 
Travis, only some fuel injection systems require a return line to the tank. The legacy injection systems, Bendix & Airflow Performance, do not. While we are free in our planes to do largely whatever we want - I prefer the legacy fuel injection systems for the Bearhawks. If you want fuel injection.

Not trying to start an argument here. Or even a discussion about fuel injection systems and which are best. But I know you are a first time builder. So I thought to chime in. Mark
 
It's a good comment, simple fuel systems are so important for safety, and return lines do not make the system simpler.
Not to mention the challenges with not overflowing one side, when you want to fill both tanks.
 
Some passive thoughts to share from my personal decision process, this is not to discourage you..just to share.

My driving reason originally to chase return type fuel injection AND ignition was fuel efficiency, engine longevity(less carbon build up), and cooler fuel (because it returns, so vapor lock is reduced).
-Fuel efficiency is taken care of with a lambda gauge if you want to be precise, this will also reduce chances of carbon buildup from overly rich mixtures(there are tests done comparing carb to fuel injection for efficiency, where the BSFC was essentially equivalent). From what I gathered, the primary factor that makes us “think” fuel injection is way better becomes clearer when there’s a control for the ignition type used for the comparison. So theoretically, after a baseline, I’d like to eventually swap one ignition out to get the majority benefit while keeping a non-electrically dependent setup(mags for the 2nd). This is all thrown out the window if you like negative Gs or are trying to have a performance setup (max hp available no matter the increase in fuel usage)
-cooler fuel/vapor lock risk reduction can be achieved with proper down hill flowing lines, insulation, heat shielding, and a carb if you’re extra concerned (b/c of mechanical injection worries of fuel heating around the spider in the heat of summer after shutdown or at a low idle/taxi)

The appraised value of these thoughts is 2 cents. Just trying to relate, injection is cool though but ended up with too many caveats for my comfort. I think discussion is healthy and reduces the risk of choices made in a vacuum.
I do not have a completed Bearhawk.
 
Last edited:
I ran my return lines up the front door post. Shorter lines, away from passengers that might accidentally damage them, easier routing and easy to cover. Routing my rear supply lines was a little challenging and sharing that space with return lines would have made it more so. I intended to cover the fuel lines and wiring I ran up the front door posts but I never have.
 
One consideration: I believe the bank attitude (in exclusion) doesn’t drive the scenario you describe. In the left turn wouldn’t you would generally be in balanced flight (ball centered)? Theoretically then, the remaining fuel would remain level in the tanks, regardless of the bank attitude. It would require a slipping turn to cause fuel to move to the down-side and un-port the pickup at the wing root for a sustained period.

The slipping turn is of course a common technique, which might need to be avoided at low fuel levels. Conversely, a skidding left turn (which pilots are taught to avoid) would potentially un-port the right pickup. In fact, any unbalanced flight in any normal flight attitude could cause these effects.

Perhaps this is a case for using a “both” selection? Im not arguing against compelling reasons not to use both-just theorizing. Personally, I’ll be careful if I am doing much maneuvering at low fuel levels. I’m going carbureted as another part of my mitigation strategy.
 
You feel me on this. Not to offend any of the answers as the info is good but yes. I know fuel injection is a topic with different opinions. I tried to ask it without creating a big debate.
 
I am looking at both options. I evaluated my mission with the plane and what we want to do and really the option of running any fuel would be really nice. Mogas/Rec Fuel/100LL and 100GAMA.
 
Thank you. There is a few somethings I didn't think about. They do offer the newton dual valves with the both option now. I really wasn't trying to start a big discussion about EFI but I have received some info I have not thought about.
 
Does anyone know the gravity feed flow rate of a Bearhawk with 3/8 inch fuel lines? A 540 burns about 23-24 gal/hr at full power and I think that some people run that without a fuel pump? Mine is equipped with fuel pumps so I can't make any assumptions from aircraft.

EDIT: Maybe I can make some assumptions?

The spec is 125% of flow rate to a pump. On a Bearhawk with feeds from both front and rear of the tank, it really should be 150%, but I should be able to expect that it was tested to at least 30 gph?
 
Last edited:
I never tested it without a fuel flow meter in the circuit, but found on both airplanes around 30gph as I recall. Bob ran his 540 without any pumps but I didn't feel like I could meet the spec without them. Maybe not having a fuel flow meter would help, but for me it's a required item to manage the engine.
 
Maybe someone can check my numbers:

A 3/8 inch OD fuel line has a wall thickness of 0.035" (I'm uncertain here, I found the spec for 1/2 line) so and ID of 0.305", an ID cross section area of 0.0731" and an ID circumference of 0.958".

A 1/2 inch OD fuel line has an ID of 0.430", a cross section area of 0.1452" and an ID circumference of 1.351"

The cross sectional area of the 1/2" line is 1.99 times that of the 3/8" line. So, for the same flow "speed", it should flow about twice the GPH.

However, it is going to be somewhat better than that because the circumference is "only" 1.4 times that of the 3/8 line so while it has twice the cross section area, it has only 1.4 times the surface area to slow it down. So, a fuel system made from 1/2 inch line should flow more than 60 gph.

Did I miss anything?
 
My system flows 78 gpm with 3/8 in lines.
Aircraft at 17 deg nose up with 5 gal in each tank, selector in both, feeding only from the aft feed ports on the tank.

Kevin D
#272
KCHD
 
Friction loss for each one. It doesn't go up or down at a proportional rate typically. Calculating friction loss is something I have to do for my job.
 
FFTravism said:
Friction loss for each one. It doesn't go up or down at a proportional rate typically. Calculating friction loss is something I have to do for my job.
Click to expand...
I didn't forget friction, but I'm also not expert in it. I tried to address it by pointing out that the ratio of cross section area vs. circumference was more favorable for the 1/2 line.
 
kestrel said:
I didn't forget friction, but I'm also not expert in it. I tried to address it by pointing out that the ratio of cross section area vs. circumference was more favorable for the 1/2 line.
Click to expand...

That makes sense. I may try to find it but it may not be readily available.
 
AZBearhawk272 said:
My system flows 78 gpm with 3/8 in lines.
Aircraft at 17 deg nose up with 5 gal in each tank, selector in both, feeding only from the aft feed ports on the tank.

Kevin D
#272
KCHD
Click to expand...

Wow! That's higher than I would have expected. Can you expand what the fuel was flowing through for this? Any chance you tested with fuel in one tank only? ...thinking of uncoordinated flight and or one tank run dry before the other.

Jared's test seemed to include at least the fuel selector valve and fuel flow sensor? I'm not sure where his test point was and if the fuel was flowing through any idle fuel pumps?

For an EFII system, the fuel pump is often under the floor or front seats, so that would be a shorter, less restrictive run than all the way through the firewall to the engine.
 
No point calculating it, theory and practice are only the same in theory. Just go with a simple system with 3/8s lines and old school injection - everyone goes through the exact thought process you're doing now, at the start. They all end in the same place... Practically, it just makes sense.
 
kestrel said:
Wow! That's higher than I would have expected. Can you expand what the fuel was flowing through for this? Any chance you tested with fuel in one tank only? ...thinking of uncoordinated flight and or one tank run dry before the other.

Jared's test seemed to include at least the fuel selector valve and fuel flow sensor? I'm not sure where his test point was and if the fuel was flowing through any idle fuel pumps?

For an EFII system, the fuel pump is often under the floor or front seats, so that would be a shorter, less restrictive run than all the way through the firewall to the engine.
Click to expand...

I suppose if you tested it at the output port of the fuel selector, that number would probably be fairly universal. It's all the other stuff downstream which slows down the flow rate, and which varies between builds.

I got something like 125 L/hr from memory (33 Gal/hr, very similar to Jared's numbers) - less than half what Kevin measured - after flowing through all the usual obstructions (selector, gascolator, filter, transducer, electric pump, flexible hose, plus countless corners / elbows, fittings). That's without the electric pump running, so it's flowing through the pump's bypass valve.

The IO-540-A4EXP "Bob" engine requires about 90 L/hr (23 Gal/hr) at full power / sea level, developing around 270 to 275 hp with ram air and fuel injection.
 
Last edited:
Battson said:
No point calculating it, theory and practice are only the same in theory. Just go with a simple system with 3/8s lines and old school injection - everyone goes through the exact thought process you're doing now, at the start. They all end in the same place... Practically, it just makes sense.
Click to expand...

Based on this approach, no one would bother taking math classes because numerical analysis is a complete waste of time. I was going to say that we'd all still be flying Wright Flyers, but that is wrong. There would have been no Wright Flyers. You are flying an airplane with fuel injection and extended wingtips of non-stock shape. Why not a carburetor and stock tips?

Everything we are flying is an advancement over what came before it. We all have different priorities on where to step a little further than the mainstream. I can't recall it in enough detail to find it, but I've read some sort of statement about pigs attacking the one that tries to get out of the pen. ...not a real event, an analogy of human behavior.

It is not true that "they all end in the same place". There are Bearhawks flying with EFII. We don't hear a lot about it because there its a lot of hostility. You'll note that the OP felt the need to reply, "I tried to ask it without creating a big debate."
 
Last edited:
Well,
Made sure tubing runs were straight, smooth,
bends not ovaled in cross section. Used two Bob gascolators, located one on each side behind and
below the struts. Fuel selector modified on the lower body for better fitting position and cleaner tubing hookup with less bends, located in front floor adjacent to flap handle. Then flowing through an electrical pump with bypass, lower firewall mounted and then on to the inlet of the standard engine driven Continental
mechanical fuel pump. Measured at the pump inlet.

Return from the pump goes directly to the left fuel tank.
via a 3/8 line.

And…… each tank has TWO, (2) vents of .280 ID.
Fuel will not flow out well if air cannot get in.

Flow from a single side with fuel level down to about 3 gallons is 48 gph.

OH! Correcting typo in my original post.
fuel flow is 78 GPH! When in both…

Kevin D
#272
KCHD
 
jaredyates, Do you want to move the supply lines discussion, or is it OK here? I didn't expect to get this involved when I high jacked the OPs topic.
 
I did 400 level "theoretical fluid dynamics" amidst a range of other fluid dynamics papers, so I feel qualified to talk about the value of calculation - particularly for flow within a pipe. It is hugely valuable and useful, but not required when deciding on your fuel line diameter in this specific case.

I think you mistake my comments to encompass more than they do, I only refer fuel line diameter.
 
Last edited:
Battson said:
I did 400 level "theoretical fluid dynamics" amidst a range of other fluid dynamics papers, so I feel qualified to talk about the value of calculation - particularly for flow within a pipe. It is hugely valuable and useful, but not required when deciding on your fuel line diameter in this specific case.

I think you mistake my comments to encompass more than they do, I only refer fuel line diameter.​
Click to expand...

Ok, you are close to saying things that I want to hear. Why are calculations (estimations with expected error value) not required (useful?) for fuel line diameter? Am I incorrect that a 1/2 line will flow at twice the gph under otherwise same conditions as a 3/8?

Note that I am not trying to predict flow rate from "scratch". I am trying to estimate the relative performance of the two line sizes and then apply that to experimental/measured data.
 
I once got 400-level science credit for going to the Ozark National UFO conference in Arkansas, so I'm not qualified at all. But I suspect the value of calculating line flow is less useful, if our flow rates are more primarily limited by the non-line members of the system, like fittings, valves, sensors, bends, etc. Sort of like using the straight part of the car racing track to calculate theoretical top speed when it's merely the slowdown prior to the next curve and the speed-up after the last one.
 
jaredyates said:
I once got 400-level science credit for going to the Ozark National UFO conference in Arkansas, so I'm not qualified at all. But I suspect the value of calculating line flow is less useful, if our flow rates are more primarily limited by the non-line members of the system, like fittings, valves, sensors, bends, etc. Sort of like using the straight part of the car racing track to calculate theoretical top speed when it's merely the slowdown prior to the next curve and the speed-up after the last one.
Click to expand...

I've considered that, at least as best I can. The fittings will also scale with the line. For an EFI type system, the fuel pump is often right after the selector valve. So, there are no extra bends, sensors, etc.

Maybe I should back up a little and better define the problem.

I'm not sure of the exact FAA wording on the 125% flow rate requirement. I've read it, but can't recall exactly. I also don't think that is the right requirement to be looking at.

If the pressure in line drops below ambient, then you are "sucking" on the line. In this case, if there is a Y in the supply line, it will prefer to pull from the one with air in it instead of the one with fuel. This is bad(tm). Bearhawk systems with fuel pumps have Y's supplying the pump. IMHO, these systems need to have 150% flow if disconnected just below any Y (or T). If all Y/T junctions can get 150% flow, then the final single line can have a pump sucking on it, for the consideration of preferring to flow fuel instead of suck air. This final comment doesn't consider cavitation concerns.

The systems that are flowing 30 gph before the fuel pump have not demonstrated 150% flow rate where it is needed. I also believe that they almost certainly have 150% where it is needed and most of the restriction is after the selector valve. After the selector valve, I think that the 125% requirement (23 gph * 1.25 = 28.75 gph) is appropriate.

The EFII System32 pump flows 40 gph. That system needs to flow 40*1.5 = 60 gph through the selector valve. I don't think that 125% should apply to that part of that type of system. I also feel that that flow should be from a single tank when "low on fuel". When low, feed lines will un-port.

Data:
2 cases of 30 gph feeding from both tanks through the selector, pump and flow sensor.
1 case of 48 gph feeding from one tank or 78 from both. Also feeding through the whole system.

Conjecture: Most of the resistance to flow in the 30 gph cases is downstream of the selector valve. I expect that these systems will still flow nearly 30 gph from a single tank. If not, they may not be safe, but I think they will.

We have the one example demonstrating that 48 gph is possible from a single tank with 3/8 lines.

From this and my calculations of cross section area and circumference, I conclude that a system that is identical to Kevin's except for being built with 1/2 inch lines will flow at least 80 gph through the selector valve from a single tank. Jarred and Johnathan's should flow something greater than 60 gph through the selector valve. Since there is much less variation before the selector valve, they may well flow through the selector valve just about as much as Kevin's.

Can we address the engineering/math here rather than jumping to conclusions and projection our own decisions and preferences? It is only a question of how much flow might be expected through the selector valve.
 
Last edited:
This is exactly the point. It's the restrictions which create most of the losses. Orifice calculations would be more material overall.

We have so much practical experience - and the first thing you learn about calculations and models of fluids is: you must validate calculations / models with practice tests. The practical test results are king.
 
Last edited:
Practical, real world data are what I'm looking for. The orifice of a 1/2" fitting vs. a 3/8" fitting scales with the tube diameter. I know the fuel flow sensor and pumps won't, but they are not part of the sub-system I'm asking about. As for validating calculations... well... no $hIt... That doesn't make the calculations stupid. Do the calculations to make an informed design. Then run the test.
 
You must log in or register to reply here.
Back
Top