Help Troubleshooting Fluctuating Fuel Pressure

[PT20J]

Well, the air in the line theory didn’t pan out. I removed the line from the transducer and it was dry. So, I ran the boost pump with it attached but the fitting loose so that it would fill the line with fuel. No change in the fluctuations. 

I noticed quite a bit of flow when I primed the line. So, I need to check that the fitting at the servo is restricted. I’m pretty sure we transferred all the old fittings when we swapped the engine, but I need to check it. 

Skip

[PT20J]

3 hours ago, Ragsf15e said:

Do you think it shows up on our electric senders but not on the old analog ones just because they’re more sensitive?

I must say, it’s annoying to see it jumping around so much during takeoff. It’s hard to get a last quick look for gages steady and green when that one is jumping all over the place.

I doubt it is the transducer as they all should be designed for fast response so that the smoothing can be tailored to the application in whatever processes the signal. Therefore, I would expect the damping to be in the gauge itself. Mechanical gauges have some natural damping due to the mass and spring mechanism. But who knows what JPI or Garmin do? It would most likely be done with some sort of software smoothing algorithm. 

Skip

And, yes, it’s annoying. 

[47U]

20 hours ago, Ragsf15e said:

Unfortunately I think it’s pulling air in somewhere and it seems to be hard to find.

In the IPB, I counted 12 threaded fittings between the selector valve and the engine driven fuel pump inlet.  Lot’s of opportunity to suck air.  I think you said your FP fluctuations started in the middle of a flight?  How old is the hose from the firewall to the engine driven fuel pump?  Stratoflex PN 111F417-6S-0184 (18.5” long).  With age and vibration, a hose under suction can fail internally and never leak externally.   

From the firewall backwards to the tank selector valve the fuel system appears to be all hard lines.  I would think less prone to leaking, unless there’s hidden chaffing or corrosion on a section of tube.

[Ragsf15e]

3 hours ago, 47U said:

In the IPB, I counted 12 threaded fittings between the selector valve and the engine driven fuel pump inlet.  Lot’s of opportunity to suck air.  I think you said your FP fluctuations started in the middle of a flight?  How old is the hose from the firewall to the engine driven fuel pump?  Stratoflex PN 111F417-6S-0184 (18.5” long).  With age and vibration, a hose under suction can fail internally and never leak externally.   

From the firewall backwards to the tank selector valve the fuel system appears to be all hard lines.  I would think less prone to leaking, unless there’s hidden chaffing or corrosion on a section of tube.

Yes, I agree with you, lots of fittings to check.  Yesterday after flying I took off the panel below the selector and electric boost pump.  I checked all those fittings and they are secure.  You’re correct, those are hard lines.  There’s one fitting from the boost pump hard line to the bulkhead/firewall pass through that’s nigh impossible to even touch.  It’s in a little cube built into the firewall?!  
Originally I thought it was a problem before the electrical boost pump since turning on the boost doest help, but I think I should check the line you mentioned.  It’s completely wrapped in firesleeve, but I’ll put it on my list.  The Tempest guys said the inlet to the mechanical pump is notorious for this as well, so I ordered a new O ring.  When we change the oil (so the filter is off), we’ll check that connection and O ring.

[Ragsf15e]

6 hours ago, PT20J said:

Well, the air in the line theory didn’t pan out. I removed the line from the transducer and it was dry. So, I ran the boost pump with it attached but the fitting loose so that it would fill the line with fuel. No change in the fluctuations. 

I noticed quite a bit of flow when I primed the line. So, I need to check that the fitting at the servo is restricted. I’m pretty sure we transferred all the old fittings when we swapped the engine, but I need to check it. 

Skip

One thought I had about air in the lines… do you think mounting the sensor line “downhill” would help?  Where my transducer is, I could easily adjust it to encourage any bubble to flow back up to the servo.

[carusoam]

5 hours ago, Ragsf15e said:

One thought I had about air in the lines… do you think mounting the sensor line “downhill” would help?  Where my transducer is, I could easily adjust it to encourage any bubble to flow back up to the servo.

Back in the day…

The air bubble was part of the installation process…

With the line being run, so the bubble didn’t go away…

The snubber would get selected to be for air…

Sure! Some fuel vapor will be at equilibrium with the temp of the air…

Contact JPI and see what they say…

Or ask @oregon87 because EI has really good customer service (FuelP sensor installation question… air bubble, snubber, or no air bubble…. )  and they are smarter than the other guys, and better looking too… 

Best regards,

-a-

[PT20J]

10 hours ago, Ragsf15e said:

One thought I had about air in the lines… do you think mounting the sensor line “downhill” would help?  Where my transducer is, I could easily adjust it to encourage any bubble to flow back up to the servo.

I don’t know. You could give it a try and see what happens. Whatever is going on doesn’t seem to affect the servo operation since the fuel flow is steady and the engine runs fine. 

[PT20J]

Today I did a test similar to @Ragsf15e. I installed a tee in the fuel pressure line from the servo so that it feeds an analog pressure gauge and the Kavlico transducer used by the G3X EIS. In this configuration, both gauges were rock steady up to 2400 rpm (highest tested). Removing the extra plumbing and reconnecting just the transducer, I saw variations up to +/- 2 psi in the G3X fuel pressure display. 

The fuel pressure takeoff from the RSA servo appears to be the proper Mooney flow restricted part. I disconnected the line from the transducer end put it in a bottle and ran the boost  pump for 30 seconds yielding 5.5 oz in the bottle which corresponds to 5 gph -- much less than cruise fuel flow. So, I conclude that the line has the appropriate restrictor.

I'm not sure what is happening, but the I think it is an instrumentation issue. The engine runs fine at all power settings and the fuel flow is steady. Since adding the analog gauge steadied the indications on the G3X, I think the extra plumbing was probably acting as a snubber.

Skip

[Ragsf15e]

1 hour ago, PT20J said:

Today I did a test similar to @Ragsf15e. I installed a tee in the fuel pressure line from the servo so that it feeds an analog pressure gauge and the Kavlico transducer used by the G3X EIS. In this configuration, both gauges were rock steady up to 2400 rpm (highest tested). Removing the extra plumbing and reconnecting just the transducer, I saw variations up to +/- 2 psi in the G3X fuel pressure display. 

The fuel pressure takeoff from the RSA servo appears to be the proper Mooney flow restricted part. I disconnected the line from the transducer end put it in a bottle and ran the boost  pump for 30 seconds yielding 5.5 oz in the bottle which corresponds to 5 gph -- much less than cruise fuel flow. So, I conclude that the line has the appropriate restrictor.

I'm not sure what is happening, but the I think it is an instrumentation issue. The engine runs fine at all power settings and the fuel flow is steady. Since adding the analog gauge steadied the indications on the G3X, I think the extra plumbing was probably acting as a snubber.

Skip

Possibly I used a snubber with the wrong viscosity, but after I came to the exact same result, I added the snubber and it did absolutely no good.

[Ragsf15e]

On 4/22/2022 at 8:14 PM, PT20J said:

Today I did a test similar to @Ragsf15e. I installed a tee in the fuel pressure line from the servo so that it feeds an analog pressure gauge and the Kavlico transducer used by the G3X EIS. In this configuration, both gauges were rock steady up to 2400 rpm (highest tested). Removing the extra plumbing and reconnecting just the transducer, I saw variations up to +/- 2 psi in the G3X fuel pressure display. 

The fuel pressure takeoff from the RSA servo appears to be the proper Mooney flow restricted part. I disconnected the line from the transducer end put it in a bottle and ran the boost  pump for 30 seconds yielding 5.5 oz in the bottle which corresponds to 5 gph -- much less than cruise fuel flow. So, I conclude that the line has the appropriate restrictor.

I'm not sure what is happening, but the I think it is an instrumentation issue. The engine runs fine at all power settings and the fuel flow is steady. Since adding the analog gauge steadied the indications on the G3X, I think the extra plumbing was probably acting as a snubber.

Skip

I have no idea why, but my fuel pressure problem has self corrected for now.  Yeah, I realize it will be back to fluctuating next time out since I posted this, but it is what it is...  So after T-ing in another gage and checking that my aircraft gage worked (above posts), I added a snubber immediately before the JPI pressure sensor.  The snubber was for light oil.  The other option was one for air.  100LL has a viscosity between those and the snubber did not help the fluctuations at all.  Further, the snubber is an npt fitting and so required an adapter to the 37degree jic on either side.  I couldn't tell exactly where, but one of them was slightly leaking.  So two weeks ago, I took all of that back apart, removed the snubber, and put it all back together just as it has been for the last 10 years.  Last week I flew it and the fuel pressure was stable.  I didn't say anything because I didn't want to jinx myself.  Flew it again today.  Stable still.  What the heck?!

Both of these are about a half hour section in cruise.  Before it was 22-30psi constantly.  Now it's 27psi almost the entire flight.  Green line is fuel pressure.

Before:

 

Now:

[carusoam]

Looks like…

1) Fuel is in the snubber

2) Air is in the snubber

 

During the re-assembly exercise… did you make sure to leave air in the system?

Or did you make sure to fill it with fuel?

Something is currently making it work pretty well…

Since the snubber’s purpose is to mechanically average what you are seeing…

One is 26psi, and the other is 27psi… pretty close.

Fly on!

Best regards,

-a-

[PT20J]

I think it has a lot to do with damping. My Mooney OEM fuel pressure gauge never did this. I even had to change the failed Kulite transducer for an Omega  transducer and it didn't to this. It is only when I switched to the Garmin G3X with EIS that I noticed it. 

The Mooney flow restrictor fitting on the servo measured at a max flow of 5 gph which isn't a great deal of restriction. I saw a post on Beechtalk saying that the restrictor should go on the source end of the line feeding the transducer. His point was that it is analogous to an RC low pass electrical filter where the flow restrictor is the resistance and the fuel in the line acts as the capacitance. Kind of makes sense. I could try adding another restrictor there, but it would take some thinking (and swearing) to figure out how to mount it there.

Skip

[carusoam]

1 hour ago, PT20J said:

I think it has a lot to do with damping. My Mooney OEM fuel pressure gauge never did this. I even had to change the failed Kulite transducer for an Omega  transducer and it didn't to this. It is only when I switched to the Garmin G3X with EIS that I noticed it. 

The Mooney flow restrictor fitting on the servo measured at a max flow of 5 gph which isn't a great deal of restriction. I saw a post on Beechtalk saying that the restrictor should go on the source end of the line feeding the transducer. His point was that it is analogous to an RC low pass electrical filter where the flow restrictor is the resistance and the fuel in the line acts as the capacitance. Kind of makes sense. I could try adding another restrictor there, but it would take some thinking (and swearing) to figure out how to mount it there.

Skip

With the old FuelP gauge in the cabin…

The preference to put the restrictor/snubber near the source…

It nicely restricts the open flow in the event the line gets disconnected in the cabin….

Yikes! Lots of blue fluid on your shoes….  

There is such a low level of fluid movement…in this close ended tube… it’s amazing that it really causes a real problem… but it does.

Probably a different problem for the fuel pressure diaphragm instruments… compared to an electronic piezo type sensor….

Makes me wonder how many diaphragms are in the fuel system to add to the wavy data… fuel pumps, injection system, ?,

best regards,

-a-

[Ragsf15e]

2 hours ago, PT20J said:

I think it has a lot to do with damping. My Mooney OEM fuel pressure gauge never did this. I even had to change the failed Kulite transducer for an Omega  transducer and it didn't to this. It is only when I switched to the Garmin G3X with EIS that I noticed it. 

The Mooney flow restrictor fitting on the servo measured at a max flow of 5 gph which isn't a great deal of restriction. I saw a post on Beechtalk saying that the restrictor should go on the source end of the line feeding the transducer. His point was that it is analogous to an RC low pass electrical filter where the flow restrictor is the resistance and the fuel in the line acts as the capacitance. Kind of makes sense. I could try adding another restrictor there, but it would take some thinking (and swearing) to figure out how to mount it there.

Skip

Yes, where mine is mounted, low on the copilot side of tge firewall, I can see it, but I have to remove the heater ducting and it’s associated adel clamps to work on it.  Much swearing ensues.  

[Ragsf15e]

16 hours ago, PT20J said:

I think it has a lot to do with damping. My Mooney OEM fuel pressure gauge never did this. I even had to change the failed Kulite transducer for an Omega  transducer and it didn't to this. It is only when I switched to the Garmin G3X with EIS that I noticed it. 

The Mooney flow restrictor fitting on the servo measured at a max flow of 5 gph which isn't a great deal of restriction. I saw a post on Beechtalk saying that the restrictor should go on the source end of the line feeding the transducer. His point was that it is analogous to an RC low pass electrical filter where the flow restrictor is the resistance and the fuel in the line acts as the capacitance. Kind of makes sense. I could try adding another restrictor there, but it would take some thinking (and swearing) to figure out how to mount it there.

Skip

I’d heard the same thing about adding the snubber closer to the source, but there was no way I was going to try to fit one in immediately after the lycoming restrictor coming out of the servo.  It was bad enough getting to where my transducer was accessible.  

[PT20J]

4 hours ago, Ragsf15e said:

I’d heard the same thing about adding the snubber closer to the source, but there was no way I was going to try to fit one in immediately after the lycoming restrictor coming out of the servo.  It was bad enough getting to where my transducer was accessible.  

It looks like the easiest (which is not to say easy) way to install a snubber at the servo end would be to remove the servo. I'm just going to live with it for now since everything is working fine. BTW, after bleeding all the air out of the fuel lines as a test a while back it seemed to stop the fluctuations for a couple of flights on mine also, but it eventually came back. On the G3X, the fuel pressure is not on the main screen, so I have to tap the engine monitor page to see it so it's not as easy to note what it's doing as with  JPI.

Skip

[PT20J]

Even with the Mooney restrictor, a LOT of fuel flows out of the hose with the boost pump on and the transducer removed. As previously noted, I measured it at 5 gph. So, the restrictor doesn’t restrict a lot.

[EricJ]

On 4/25/2022 at 5:00 PM, PT20J said:

I saw a post on Beechtalk saying that the restrictor should go on the source end of the line feeding the transducer. His point was that it is analogous to an RC low pass electrical filter where the flow restrictor is the resistance and the fuel in the line acts as the capacitance. Kind of makes sense. I could try adding another restrictor there, but it would take some thinking (and swearing) to figure out how to mount it there.

Skip

Saying the order matters suggests that at least one of them isn't linear.   Even in a basic series RC circuit the order of the components doesn't matter since they're linear devices.   That's not to say there isn't something non-linear in this system, like potentially the transducer, but if the snubber and the hose and the mass in the hose are all linear, the order shouldn't matter, and I suspect that's the case.   Putting the snubber closer to the pressure source can limit leaks if the hose fails, though.

[N201MKTurbo]

6 hours ago, EricJ said:

Saying the order matters suggests that at least one of them isn't linear.   Even in a basic series RC circuit the order of the components doesn't matter since they're linear devices.   That's not to say there isn't something non-linear in this system, like potentially the transducer, but if the snubber and the hose and the mass in the hose are all linear, the order shouldn't matter, and I suspect that's the case.   Putting the snubber closer to the pressure source can limit leaks if the hose fails, though.

I have to call a bit of BS on this. In an RC circuit:

IN ---RRRR--- OUT

                        |

                       C

                        |

                    GND

You cannot swap the components. This would be the circuit the poster was talking about. In a series RC circuit, you would be correct, but that is not analogous to the meter plumbing.

[EricJ]

6 minutes ago, N201MKTurbo said:

I have to call a bit of BS on this. In an RC circuit:

IN ---RRRR--- OUT

                        |

                       C

                        |

                    GND

You cannot swap the components. This would be the circuit the poster was talking about. In a series RC circuit, you would be correct, but that is not analogous to the meter plumbing.

Is it?   I was interpreting that the hose has two ends, the source at the servo and the transducer/instrument at the other, and the placement of the snubber was either on one end or the other, i.e., at the instrument/sensor/transducer as usual, or at the source end.   Your circuit shown is the normal low-pass filter configuration, but it would require tapping the sensor directly at the snubber and what happens to the other end of the hose?   Cap it off?

If the idea is to tap the sensor at the middle between the R and the C, then swapping them converts it from a low-pass filter (as you've shown), to a high-pass filter, which wouldn't be very desirable.

So my understanding of the circuit model was that the mass of the fluid (the capacitor analog) would swap in series with the snubber (the resistor analog).   In that case I can't see a difference in the RC model since the order of an R and a C in series (or any two linear elements or processes) doesn't matter.  

[Ragsf15e]

1 hour ago, EricJ said:

Is it?   I was interpreting that the hose has two ends, the source at the servo and the transducer/instrument at the other, and the placement of the snubber was either on one end or the other, i.e., at the instrument/sensor/transducer as usual, or at the source end.   Your circuit shown is the normal low-pass filter configuration, but it would require tapping the sensor directly at the snubber and what happens to the other end of the hose?   Cap it off?

If the idea is to tap the sensor at the middle between the R and the C, then swapping them converts it from a low-pass filter (as you've shown), to a high-pass filter, which wouldn't be very desirable.

So my understanding of the circuit model was that the mass of the fluid (the capacitor analog) would swap in series with the snubber (the resistor analog).   In that case I can't see a difference in the RC model since the order of an R and a C in series (or any two linear elements or processes) doesn't matter.  

And that’s why I nearly failed EE in college and had to settle for being a simple Mechanical Engineer…

[PT20J]

Well, I didn't think deeply about it, and analogies between fluids and electricity invariably break down at some point, but I thought of the pump output being a voltage source, the restrictor being a resistor and the fluid/air in the line being analogous to a capacitor to ground. But I see Eric's point if the fluid/air is actually analogous to a series capacitor then the order doesn't matter. I'm going to have another glass of wine and think about it.

[EricJ]

1 hour ago, PT20J said:

Well, I didn't think deeply about it, and analogies between fluids and electricity invariably break down at some point, but I thought of the pump output being a voltage source, the restrictor being a resistor and the fluid/air in the line being analogous to a capacitor to ground. But I see Eric's point if the fluid/air is actually analogous to a series capacitor then the order doesn't matter. I'm going to have another glass of wine and think about it.

If it's air its compressible, if it's hydraulic it's not, so that should make a difference in the analog model.   A cap is a normal model for mass, and apparently also for air with a compressor, but I think the details vary between them...I've never thought of modelling a pneumatic system with basic elements but apparently that's reasonably straightforward.

 

[EricJ]

Sadly, I woke up thinking about this.   Engineering life, I guess.

I think @N201MKTurbo's model is probably right, with the C that matters being the transducer/sensor/instrument, and the R just whatever goes before it.   Adding the snubber just increases R which decreases the bandwidth of the low-pass filter.

So, presumably, it shouldn't matter which end of the line the snubber gets added, because it's just increasing R.

I don't know how much of the line from the servo to the transducer/sensor winds up with fuel in it or air in it.   Putting the snubber at the servo might mean it's resisting fuel flow rather than air flow, which might mean R is different than if it is resisting air at the other end of the line.    The installation instructions for my JPI don't say anything about this, so it's no help.

 

 

[PT20J]

The whole thing is a bit baffling. Here is everything I know about it. 

1. The line to the transducer is a 29" long 3/16" diameter OD (-3) hose. 

2. The hose is connected to a Mooney flow restrictor attached to a port on the servo just past the inlet filter screen and before the idle mixture/cutoff valve. So, when the mixture is in ICO, the servo is completely isolated from the fuel lines.

3. The fuel injector is an Avstar. Compared to a Bendix RSA, it has a very tight mixture valve. With the mixture in ICO, it will hold some pressure in the lines for a day. Because of this, I'm certain that there are no air leaks in the fuel lines.

4. I can bleed all the air out of the line all the way to the transducer and after running up to operating temperature, the line to the transducer will be mostly void of fuel. According to Precision Airmotive, this is normal because the heat boils the fuel in the lines if it isn't flowing and the line to the transducer is a dead end with no flow. After shutting down a hot engine by placing the mixture control in ICO, this effect can be seen as a rise in fuel pressure. In fact it will peg the gauge. When I had the analog gauge installed, it would go up to about 75 psi. I think it might go even higher with the cowling installed which would trap more heat.

5. I never saw fluctuations in the fuel pressure with the factory gauge and transducer, but it shows up on the G3X EIS as a slow variation with an amplitude of about +/- 2 psi and a period of maybe a second. The only change made was the G3X and it's transducer.

6. I installed a precision analog gauge in the system using some old hoses I had around. The hose from the servo went to a T and then a 17" hose went from the T to the transducer and another 29" hose went from the T to the analog gauge. With all this extra plumbing, there was no variation in the G3X or the analog gauge. The G3X gauge was more heavily damped (probably by software filtering) than the analog gauge. At lower rpm I could see rapid variation of the analog gauge needle (caused by the mechanical fuel pump pulsating) which was not evident in the G3X.

7. I measured the Mooney flow restrictor at about 5 gph flow with the boost pump and about 25 psi pressure.

Skip