Fuel flow stuck at 9.9 gph

[PT20J]

OK, here's a weird one. I wonder if anyone else has ever seen this. 

M20J with and IO-360-A3B6. Floscan 201B fuel flow transducer installed between the engine-driven pump and servo. It's mounted wire side up with no angle fittings at input or output. Engine monitor is a Garmin G3X EIS.

Climbed to 6500' and set power WOT/2500. Leaned, and cylinder 3 peaked at 9.9 gph. Continued leaning and the EGTs continued to peak and then decrease but the fuel flow stayed at 9.9. Eventually, with continued leaning the fuel flow began moving at which point it was quite a bit lean of peak. I've seen it do this a few times and the 9.9 gph point is quite repeatable, but I fly a lot of short trips at 2500' and I never have seen this behavior during those flights. It didn't do this until recently. I'm thinking something wrong with the transducer, but I'm puzzled that it works fine at low altitudes which makes me think vapor -- but then why always at 9.9 and why didn't it so this until recently? Turning on the boost pump raised the fuel pressure a pound or so, but didn't change the fuel flow. I have been chasing a fuel pressure variation which I believe to be due to fuel vapor forming in the engine-driven pump  because all the fuel lines and the flow transducer are well insulated and directing a heat gun at them doesn't affect static fuel pressure but it doesn't take much heat directed at the engine-driven pump to start to raise the static fuel pressure. 

[GeeBee]

28 minutes ago, PT20J said:

Turning on the boost pump raised the fuel pressure a pound or so, but didn't change the fuel flow.

That tells you everything. If it were vapor the increased fuel pressure would have resolved it. No increase in FF with the pump on and verified working by the pressure gauge pretty much indicts the transducer.

[PT20J]

6 minutes ago, GeeBee said:

That tells you everything. If it were vapor the increased fuel pressure would have resolved it. No increase in FF with the pump on and verified working by the pressure gauge pretty much indicts the transducer.

Ah, but then why does the boost pump not resolve the fluctuating fuel pressure, and why does the fuel flow not behave this way at lower altitudes?

 

[GeeBee]

17 minutes ago, PT20J said:

Ah, but then why does the boost pump not resolve the fluctuating fuel pressure, and why does the fuel flow not behave this way at lower altitudes?

 

OK, you indicted fluctuating fuel pressure but I thought it was only with the engine driven pump not with the boost and engine pump. If that is the case, I would look for restriction to the engine pump creating a vapor issue as it sucks hard. Have you done a fuel pump test or replaced the engine driven pump? I believe the diaphragm on that pump has a flapper built into it. If the flapper is bad that can create fluctuations. 

 

https://vansairforce.net/threads/dying-engine-driven-fuel-pump.160388/

 

[PT20J]

The fuel pump came new with the engine about 500 hrs ago. The hoses are all new. Gascolator and servo inlet screens are always clean. 

I know that it's common on injected Continentals to use low boost pump to suppress vapor lock. I'm not entirely clear that this works with a Lycoming because the fuel systems are different. The Continental engine fuel pump includes a vapor separator and a return line to the tank so that running the boost pump can circulate the fuel through the engine pump and the vapor separator. The RSA servo is closed with no return line so the boost pump on a Lycoming can only increase the pressure to whatever the pump is set for. 

I know that my fuel lines fill with vapor when the engine is shut down because if I remove the pressure transducer the line is always dry. If I fill it with fuel and replace the transducer it will be dry again later. Also, after shut down, the fuel pressure rises off scale from the residual heat in the engine compartment. Lycoming switched to AvStar injection systems and that's what came with my rebuilt. The RSA idle cut off valve was never a perfect seal but AvStar seems to have "improved" on it and so when it is shut off there is nowhere for the pressure to bleed off. If I leave the mixture in ICO it will take over a day after shutdown for the pressure to bleed off so I don't think there are any air leaks.

The pressure rise after shutdown used to show up on my OEM instrument but I never noticed the fluctuations. They were probably there but the gauge was probably damped more so they weren't noticeable.

All these interesting anomalies show up on the instrumentation but the engine seems to run just fine. 

[GeeBee]

I've dead sticked several RSA systems (broken diaphragm) but never had vapor lock with them. Being a closed system they should actually be less prone as the supply operates at a higher pressure which is one of the reasons Continental has a vapor separator. That all said vapor lock can only come from three sources. Rapid expansion of the liquid (with a resulting pressure drop), excessive fuel temperature or out of spec vapor point fuel. I think the last one we can eliminate so the other two you have to look at.

I know there is a lot of problem on auto to marine fuel injection because the fuel pumps on marine have to suck as opposed to push fuel (to meet CG regulations). With a push pump, the pressure can be increased, increasing the temp at which vapor lock occurs. I've seen several instance when the mechanical sucking pump, a similar design to that on the Lycoming (basically a GM fuel pump)  will blow a diaphragm and with the sucking along with the fuel being thrashed around by the diaphragm vapor lock will occur. On these engines there is actually a "cool fuel" mod to insulate the fuel system from heat since the vapor lock so easy. The other alternative is to install a pushing electric pump with an oil pressure cutoff such that it will not operate without engine oil pressure.

I'm not familiar with the J fuel system but if the electric pump pushes fuel before it reaches the mechanical the use of the electric pump should stop vapor locking by increasing fuel system pressure unless the mechanical is thrashing or there is a heat source somewhere.

[PeteMc]

I just had my FF fail, bad transducer.  Sucked because the plane sat on the ground forever in the interior shop, but the manufacturer said it was 1 month out of warrantee regardless of the minimal hours it actually was in use. 

[EricJ]

50 minutes ago, GeeBee said:

I've dead sticked several RSA systems (broken diaphragm) but never had vapor lock with them. Being a closed system they should actually be less prone as the supply operates at a higher pressure which is one of the reasons Continental has a vapor separator. That all said vapor lock can only come from three sources. Rapid expansion of the liquid (with a resulting pressure drop), excessive fuel temperature or out of spec vapor point fuel. I think the last one we can eliminate so the other two you have to look at.

I know there is a lot of problem on auto to marine fuel injection because the fuel pumps on marine have to suck as opposed to push fuel (to meet CG regulations). With a push pump, the pressure can be increased, increasing the temp at which vapor lock occurs. I've seen several instance when the mechanical sucking pump, a similar design to that on the Lycoming (basically a GM fuel pump)  will blow a diaphragm and with the sucking along with the fuel being thrashed around by the diaphragm vapor lock will occur. On these engines there is actually a "cool fuel" mod to insulate the fuel system from heat since the vapor lock so easy. The other alternative is to install a pushing electric pump with an oil pressure cutoff such that it will not operate without engine oil pressure.

I'm not familiar with the J fuel system but if the electric pump pushes fuel before it reaches the mechanical the use of the electric pump should stop vapor locking by increasing fuel system pressure unless the mechanical is thrashing or there is a heat source somewhere.

The engine-driven mechanical fuel pump draws fuel up from the tanks, i.e., sucks, so the pressure on the input side of the pump is lower than ambient.   A heat source on an exposed portion of the low-pressure side makes it susceptible to boiling at the reduced pressure, so the mechanical pump is then subject to vapor lock.

The boost pump pressurizes the input side of the mechanical pump so that it is less likely to vapor lock.   It moves the low pressure (suction) side of the fuel line to behind the firewall (less heat) and below the fuel tanks (more input pressure), so it is unlikely to suffer from vapor lock.

[GeeBee]

1 hour ago, EricJ said:

The engine-driven mechanical fuel pump draws fuel up from the tanks, i.e., sucks, so the pressure on the input side of the pump is lower than ambient.   A heat source on an exposed portion of the low-pressure side makes it susceptible to boiling at the reduced pressure, so the mechanical pump is then subject to vapor lock.

The boost pump pressurizes the input side of the mechanical pump so that it is less likely to vapor lock.   It moves the low pressure (suction) side of the fuel line to behind the firewall (less heat) and below the fuel tanks (more input pressure), so it is unlikely to suffer from vapor lock.

So that would argue for the mechanical pump thrashing the fuel with a bad diaphragm or check valve because the electric pump should offset the effect of the inlet pressure drop.