FI Fuel Return Circuit

[pschwenn]

As with many/most fuel injected engines, Rotax 912i, 914 & 915T require a return circuit of excess gasoline - not delivered by the injectors. The excess fuel returns to the fuel tank. This circuit requires a backflow cutoff - a one way valve - to prevent fuel coming from the tank into the return circuit. Rotax recommends such a circuit but does not detail one or provide one.
When I look at schematics of examples of such circuits I can't fully understand them because it's not just a single path: source, fuel line, one-valve, fuel line, tank. There's more going on, some kind of parallelism in lines of the circuit.
Who knows of a explanatory example? And of a particular type/capacity/... of valve to use?

[Related: I'm going to route the return thru the overflow space (small metal box at the top of the firewall) that contains the tank neck and then directly thru that neck not too far below the gas cap, on the theory that feeding it back lower in the tank risks putting it below the surface of the fuel and thus providing a back pressure. On the other hand, routing as I am means that return fuel will be running down the non-liquid gas above the fuel. Since the tank has a breather this means that highly aromatic gasoline may be constantly mixing with some air and thus either degrade (e.g. lose oxygen or additives) or increase the fire risk at the breather.]

Thank you,

Peter Schwenn

[GordonTurner]

I’m running mine to a steel bulkhead fitting around the middle of the firewall (vertically and horizontally, but not intended to be exactly centered), through the one way check-valve which will be just inside of the firewall, then through stainless braided Teflon line to the top port on the back of the tank intended for the sight gauge.

I wouldn’t worry about back pressure. The fuel pump will never know the difference.

[Kai]

Peter,

I know nothing about the efi system Rotax supplies with their engines, but I have seen their diagrams. However, I have been flying my Sonex behind aftermarket efi systems since 2011- and my present 120hp EP915ECI is no different: they all require a fuel return system to the tank.

Like you’re planning, I have my return system dumping the fuel in the tank neck coming out in the fuel spill tray. I arranged this with an apprx 4” long brass pipe routing the fuel out of the bottom of the neck, and spilling it into the fuel. The neck is the highest point in the whole system, and in the AN6 return line between the fuel pressure regulator and the neck, there is nothing- and certainly no non return valve.

So far this has worked out according to expectations: no snags whatsoever.

However, you should be aware of another point: the return fuel gets heated. Warm enough, and especially running on bioethanol mogas, the stuff starts producing vapour. These vapour bubbles must be removed from the fuel before it gets back into the pump, and dumping it into a larger, colder quantity certainly helps, going into your main tank, where your fuel supply gradually gets hotter. You can under no circumstance allow the temperature to get up to 40 centigrades, so get back on the ground with at least 15l in the mains.

[GordonTurner]

That’s a very interesting observation Kai. Do you have a fuel temperature probe? It seems like it should be easy enough to install one.

[GordonTurner]

This duscussion raises another question in my pea brain. If the return flow is flowing through vapor at the top of the tank or cascading down the plastic interior of the tank is there any potential being built up that could under the wrong conditions cause a static discharge spark?

[Kai]

That’s a very interesting observation Kai. Do you have a fuel temperature probe? It seems like it should be easy enough to install one.

No- but I had one!

During the testing phase of the system we experimented with as little fuel as possible in the main tank- say something like one gallon: barely enough to avoid pump cavitation. We found that by just running the fuel through the system against the pressure regulator with the engine at standstill, the fuel temperarure increased from 6 centigrades to 23 centigrades in 20 minutes. With additional heat from a running engine this would of course be worse.

Now, with the system up and running, when flying I occasionally put my hand against the bottom of the tank: believe me, it’s warm- though not as warm as 40 centigrades. So, as stated previously, my personal limit for being back on the ground is 15l fuel remaining in the tank- I need it’s cooling!

Finally, bear in mind that your tank is atmospheric- unpressurized! The fuel in there is producing vapour just on its own, splashing around in there during flight. The effective way to avoid fuel gas in the tank would be to pressurize it- just like it’s done in your car. As far as I can understand, that is a possibility we do not have!

The small amount of additional fuel gas vapour coming out of the return line orifice would in my opinion have little bearing on the overall vapour situation in the tank: since 2011 and more than 300 hrs behind efi engines, I don’t seem to recall any tank explosions……..

[pschwenn]

I might add a simple cooling means to the return line, such as an externally finned section, careful routing away from hot things, maybe via a cool inlet - whatever presents itself.

It occurs to me - warm fuel, like warm intake air, may not be best for power. But its a tiny effect with fuel compared to intake air, so safety seems to be the important thing.

[Skippydiesel]

Peter,

I know nothing about the efi system Rotax supplies with their engines, but I have seen their diagrams. However, I have been flying my Sonex behind aftermarket efi systems since 2011- and my present 120hp EP915ECI is no different: they all require a fuel return system to the tank.

Like you’re planning, I have my return system dumping the fuel in the tank neck coming out in the fuel spill tray. I arranged this with an apprx 4” long brass pipe routing the fuel out of the bottom of the neck, and spilling it into the fuel. The neck is the highest point in the whole system, and in the AN6 return line between the fuel pressure regulator and the neck, there is nothing- and certainly no non return valve.

So far this has worked out according to expectations: no snags whatsoever.

However, you should be aware of another point: the return fuel gets heated. Warm enough, and especially running on bioethanol mogas, the stuff starts producing vapour. These vapour bubbles must be removed from the fuel before it gets back into the pump, and dumping it into a larger, colder quantity certainly helps, going into your main tank, where your fuel supply gradually gets hotter. You can under no circumstance allow the temperature to get up to 40 centigrades, so get back on the ground with at least 15l in the mains.

Hi Kai,

As a long time diesel freak, I know that high pressure common rail fuel injection systems do heat the fuel (pressure = heat) much more so than the old injector systems. Most if not all such automotive applications have a cooling system on their return fuel line. It usually consists of a convoluted serpentine pipe surrounded by fins. I am not familiar with common rail petrol systems but would assume that the same principals apply. You could probably source a fuel cooler from the local car wreckers, failing that an aftermarket automatic transmission oil cooler wold probably do the job for you.

[Kai]

Well, yes,

But I was reluctant to suggest this- there is this deltaT thing!

In moderately warm summers a fuel cooler could perhaps be useful, but if you have something really hot like the more than 100F’s experienced a lot of places, the fuel cooler could turn into a fuel heater.

You would have to run a lot of tests before you have enough data to thrust its reliability and usefulness.

[Murray Parr]

I have Rotax 912ULS with Rotec TBI and is primarily gravity fed with a boost pump for back up. Gravity fuel flow tests are fine even when almost empty so the boost pump might not have been necessary. There is a check valve just before a Tee which goes through the firewall and straight down the the gascolator (all downhill to there then all uphill to TBI. I don't have a lot of gravity head pressure so the boost pump branches off the gravity line then back to the Tee that goes through the firewall. The check valve mentioned earlier prevents boosted pressure flowing directly back to the tank. When fuel pressure is boosted, excess fuel passes through the TBI and through a restricted orifice and back to the gravity feed line through a check valve then to a Tee from there it can circulate back to the boost pump and/or return to the tank. Might not be perfect at removing air bubbles from lines as some bubles can circulate back to the pump.

As can be seen in the PDF attachment, the lower downhill line with the check valve is the gravity system and the boost pump loop can be seen above that. The return line from the TBI can be seen passing through the firewall, it has the red coloured elbow then its check valve. you can just make out where it TEE's back into the system just after the fuel flow meter and just before the TEE that feeds either gravity or boost pump.