Fuel System Questions

[Nev]

Fuel Servo Filter

When I did my fuel system inflight testing I decided to set a fuel pressure warning at 18 psi. My engine driven pump maintains around 22-25psi depending on the flow rate and if it ever dropped below that I figured I'd want to know about it.

Last week while cruising at 9500ft, LOP and admiring the view I got a "Fuel pressure low" warning in my headset.
I saw the fuel pressure briefly at 16 psi, then it returned to normal. I had immediately turned the fuel pump on and set full rich mixture (probably didn't help). Over the next 7 minutes I observed the fuel pressure briefly decrease again, then return to normal. Overall it did this 4 times. Following that I didn't see any fluctuations on subsequent flights. I didn't notice any other symptoms and the engine appeared to be running normally in all other respects. I kept an eye on it over the next few hours but everything seemed normal.

This week I downloaded the data for the flight in question, which confirmed a reduction in fuel flow each time, concurrent with increasing EGT's (slightly delayed), and reducing fuel pressure. Due to the CSU I hadn't noticed any change in engine sound. Each episode had lasted around 5-7 seconds. I decided it warranted looking into further and I contacted Mark who in turn spoke to Bob who suggested I checked the gascolator. I did, and as previously, it was clean.

I had read a number of times about the fuel servo "last chance" filter, but I hadn't ever seen any schedule to clean it. I finally found a Lycoming reference today that does recommend cleaning it at the annual/100 hour check. So we removed and checked it. Incidentally, it's located where the fuel "in" hose attaches on a Lycoming Bendix fuel servo and is very quick and simple to remove.

It contained quite an amount of swarf and microscopic debris and although we won't know for sure if it caused the pressure fluctuations (though it's probably likely), I'm very glad that we found and removed the contamination.

Prior to final assembly I had cleaned all my fuel lines with the air compressor, and flushed the tanks. However it looks like this swarf may have originated in the electric fuel pump as it was running in. Not certain about this, but if so then the usual fuel filter wouldn't have caught it. Same for the gascolator as they are both prior to the electric pump. It does make me wonder if the fuel flow transducer has been affected in any way, though so far it has been accurate when comparing "fuel in" on refueling.

Had I not have had the fuel pressure warning set I would have missed the whole thing and been none the wiser, which made me wonder how many times previously in small aircraft this has happened. What it amounts to is 4 small pressure fluctuations in 65 hours of flying, that I would just as easily have missed completely if not for the first fuel pressure warning.

I'm going to make cleaning the fuel servo filter part of the 50 hour check (and hopefully observe a reduction in contamination) - hopefully it's just a "one off" occurrence. Otherwise I'll be revisiting filter options.

Having now seen this, if I was building again I would clean this filter at around 20 hours during the test flying. If running an engine monitoring system I would also definitely set a fuel low pressure warning again.

[svyolo]

I didn't know that filter existed. I always kind of questioned aircraft fuel system design where the gascolator/filter is upstream of fuel pump(s). I have added a second, upstream filter on a couple of vehicles and boats over the years, but as far as I recall all of them had the factory filter downstream of the fuel pump.

[Battson]

Normally that filter is clean and we don't bother checking it every 100 hrs, just every few years - and it's always clean. It has been dirty exactly once - after the first 10 hours of flying on a new build. There's been lots of discussion about this filter and the need to check it after first flights.

[rodsmith]

That is great info, thanks for sharing it.

[Nev]

I always kind of questioned aircraft fuel system design where the gascolator/filter is upstream of fuel pump(s).

You've raised an interesting point because I'm not sure if gascolators and fuel pumps were ever designed to go together in this way. My limited understanding is that gascolators were originally designed as a combined filter/water trap in carbureted aircraft. Bob probably had this in mind, and those of us that have added fuel injection systems have then added fuel pumps downstream of the filter/gascolator. With a carbureted gravity fed system there's usually nothing after the gascolator to produce swarf.

[svyolo]

Having a "last chance" filter in the Bendix type system gives makes me more confident in that system, but I never heard of it before.

[Ray Strickland]

Nev,
I am wondering where the fuel pressure sensor is in your system. There is a port in the side of the fuel servo screen housing. If that is where your sensor is plumbed, then any flow restriction in the servo screen could cause a low pressure reading. Glad you found the debris prior to an engine stoppage. Hopefully, all of the debris has been cleaned out such that the problem will not occur again.

[Nev]

I am wondering where the fuel pressure sensor is in your system.

Ray, I need to check but I'm fairly sure that my fuel pressure sensor is plumbed in the vicinity of the engine driven pump. For that reason I'm not 100% certain that the low pressure reading was directly correlated with what we found. The data download showed the exact sequence of events was a decrease in fuel flow, followed by decrease in fuel pressure. Then increased EGT's which appear to have been accompanying the flow "bump" as everything returned to normal and caught up again. I think having the pressure sender read from the downstream side of the fuel servo would make a lot of sense.

[Ray Strickland]

In that case, I would look upstream of the electric pump.
Are your fuel tank vents connected?
Ray

[Nev]

They're vented through the fuel caps. I do have a suspicion that the low fuel pressure I saw may have been caused AFTER the fuel flow started recovering back to normal. In other words if a restriction caused the flow to reduce, then cleared a few seconds later, the flow increasing back to normal might have reduced the pressure and set the alarm off after the fact. There is a 5 second delay between the fuel flow dropping and the fuel pressure dropping. I'll add a screenshot.

[Nev]

This is zoomed right in on one small occurrence that wasn't significant enough to trigger the warning.
But you can see the (blue) fuel flow decreases for a couple of seconds, then starts to rise. At this point the fuel pressure starts decreasing. Fuel flow continues to increase above normal, then settles back to the set flow rate, effectively compensating for the previous reduced flow. At that point pressure returns to normal. It was the same pattern 4 times, each lasting only a few seconds. The delay makes me think that perhaps the fuel pressure decrease was triggered by the increase in fuel flow. Thoughts ?

22E79728-95E4-4535-BFC1-1947EC461403.jpg

[Ray Strickland]

Nev, a real head scratcher…three thoughts, FWIW:
1. There seems to be an intermittent flow obstruction somewhere in the system, but upstream of both fuel pumps.
2. The delay in the pressure response may well be due to the flow restrictor fitting that leads to the pressure transducer.
3. The overshoot in recovery of fuel flow followed by return to steady-state is likely the pump(s) “catching up” with the lowered downstream pressure after the restriction clears.

[Battson]

Yeah, a blocked filter has not caused this. Not downstream of the pumps anyway. This could be a blockage upstream of the pump / sensors.

[svyolo]

This is zoomed right in on one small occurrence that wasn't significant enough to trigger the warning.
But you can see the (blue) fuel flow decreases for a couple of seconds, then starts to rise. At this point the fuel pressure starts decreasing. Fuel flow continues to increase above normal, then settles back to the set flow rate, effectively compensating for the previous reduced flow. At that point pressure returns to normal. It was the same pattern 4 times, each lasting only a few seconds. The delay makes me think that perhaps the fuel pressure decrease was triggered by the increase in fuel flow. Thoughts ?

22E79728-95E4-4535-BFC1-1947EC461403.jpg

Nev;
Fuel pressure sensors and FF sensors are completely different, which I know, you know. Another variable is the type of "digital filter" used in your engine monitoring system for the individual sensors. I am not sure you can say the fuel flow even happened before the fuel pressure event when they are only a couple of seconds apart. It might be true, but might also not be true. It might be displayed, and recorded that way. If they were farther apart, like 20-30 seconds, I might agree.

I am not sure, but hopefully cleaning the filters fixed it.

[AZBearhawk272]

Nev,

Interesting data, This stuff is my day job. I read and re read your initial post a few times and think I understand your system. That said a few observations:

I think the data is good and in proper phase and timing.
I am making a few assumptions on the location of your sensors based on system and event descriptions.

1. Pressure is the result of a restriction to flow. ( always, if you not flowing you make no pressure )
2. The fuel flow decreased first followed by a decrease in pressure. The fuel flow did not go down to zero.
3. The rate of change of fuel flow is in concert proportionally with the change in pressure. The slope of the lines is consistent and indicative of a " Tight Downstream Condition."
4. Both the fuel flow and the pressure exhibit fluxuations just prior to the " Big Dip".
5. After the "Big Dip" Flow and pressure are more stable, little to no oscillations.
6. The time interval between the preceding flow oscillations is about the same period as between the "Big Flow" dip and the pressure drop. This makes sense.

So what and what.

a. Cavitation of the mechanically driven engine pump? Restricted inlet to the pump? Restricted Air inlet to the Fuel Tank? Plugged finger strainer at the tank? Thermal Vapor lock / bubbles?..........
b. Intermittent sealing of the engine driven diaphragm pump outlet check valve? Internal bypass relief?
c. Air leaking, sucking into the fuel delivery line prior to the pump?
d. How full were your tanks?
e. Sketchy diaphragm pump? Is the pump a diaphragm or a gear pump?
f. What was your power setting and rich or lean of peak?

Just food for thought, I am north of the equator........ : )

Kevin D
#272
KCHD

[Nev]

Kev,

Thanks for your help with this. I'll check the tank cap vents just to be sure, and on the upstream end that leaves just the finger strainers. I cleaned the tanks and used compressed air to clean the lines prior to use. We also disconnected the fuel lines and checked the fuel flow at the servo a few days ago. It had a strong flow on each tank individually. This made me think that if the issue was tank vents or the strainers, that both sides would have to have been blocked simultaneously, because the fuel selector was on BOTH when the issues occurred.

I've added some thoughts/data to your other points:

a. Cavitation of the mechanically driven engine pump? Restricted inlet to the pump? If the electric pump has produced swarf then it could be lodged in the engine driven pump.
Restricted Air inlet to the Fuel Tank? Plugged finger strainer at the tank? I'll check these.
Thermal Vapor lock / bubbles?.......... This is definitely one on my mind after reading the Cessna bulletin. It seems very coincidental to be seeing a similar issue at similar altitudes and similar fuel system.
b. Intermittent sealing of the engine driven diaphragm pump outlet check valve? Internal bypass relief? I need to learn more about this. Perhaps this would be solved at the same time as (a) checking for restriction ?
c. Air leaking, sucking into the fuel delivery line prior to the pump?
d. How full were your tanks? 43 Gallons remaining.
e. Sketchy diaphragm pump? Is the pump a diaphragm or a gear pump? Diaphragm I believe, I'll check.
f. What was your power setting and rich or lean of peak? 50% Power / 20" MAP and very lean of peak.

[Battson]

Yours is a diaphragm pump

[Nev]

Regarding the potential for vapor-lock, there's quite alot of info floating around applicable to the early Cessnas, although not so much explaining the reason behind the issues, I did find one explanation today on another website relayed after the pilot experienced an engine failure and landed successfully. It sounds plausible and matches one of the Cessna bulletins. This is likely to be quite dependent on individual fuel pipe bends (particularly among our home builds), and particularly local climate. The bulletin stated it is more likely in conditions of high temp, high humidity, and low pressure. The low pressure might also be being amplified by an out-of-balance condition.

"....... there is a procedure that says to switch from BOTH tanks to Left or Right once attaining cruise altitude above 5,000........Apparently, in that era 172, fuel in the lines running down behind the doorframe can boil and lead to vapor lock. The solution is to increase the flow through one of the lines by switching to a single tank. Any bubbles that form in the unused line will rise up and exit into the tank. If one were still to experience a vapor in the line being used, switching to the unused line should immediately restore power."

[Gerhard Rieger]

Hi Nev, I looked at all the issues there was with different fuel systems, I have a IO 540, I came to the conclusion that for me I have to use a Header tank. It all starts with my fuel caps with vent pipes above the fuel tanks, this will assure proper pressure in the tank so that it will drain well. Next up is the vent pipe between the Aux and main tanks. From there the standard 3/8 fuel lines as per plans down to the fuel selector Left and Right only. This way you always know how much fuel is in each tank. From the fuel selector it goes through the gascolator to the header tank. The header tank hold about 2;2 USG and it have two vent pipes, one on the top left rear and one top right front. This way if you are climbing any vapour will go out the top right to the right main tank, descending the rear left vent will allow it to flow to the left main tank, it will also leave you with a vent high when banked left or right. When there is a long distance for fuel to flow from the tanks through fuel selectors, gascolators and bends and fitting it leaves you with resistance and possible air locks. But having a huge vent line at the end just before the engin eliminates those issues. When in level flight both these vents of 1/2 inch will assure easy flow to keep the header tank filled at all times. The aux fuel pump is next leaving the aux tank un-pressurised. This setup leaves me with a very short distance of less than 1m of 3/8 fuel line from the header tank through the pump to the engin driven pump, un-vented. I know it will be very difficult to add this to your aircraft now. Hope you gets yours sorted soon.

[Chris Werner]

Nev I have one of these Cessnas with the issue similar to what you are seeing. There was a Cessna service kit to fix the issue and eliminate the placard for single tank above 5000 ft. The service kit vents the fuel feed line at the rear of the tank back to the vent at the front. I'll attach a link to the instructions so you can see the drawings.

https://support.cessna.com/custsupt/contacts/pubs/ourpdf.pdf?as_id=33762

[Battson]

It's not an issue with the engine pump Nev , that would show as the pressure dropping first, not the flow. The Bendix Injection system works on wasted pressure.

Edit: I assume the date is accurate and there is no sensor lag. I assume time is minutes : seconds

[Nev]

It's not an issue with the engine pump Nev , that would show as the pressure dropping first, not the flow. The Bendix Injection system works on wasted pressure.

Edit: I assume the date is accurate and there is no sensor lag. I assume time is minutes : seconds

That's good to know thanks. Today I checked the location of the fuel pressure sensor. It comes from a T on the downstream side of the engine driven pump, where the fuel hose continues to the fuel servo.

I also checked the fuel cap vents again today - they're also clear. I think the possibility of the tank finger strainers being blocked is not likely, because they would both have to have blocked simultaneously, and then both unblocked. We did a flow check afterwards and found the fuel to be flowing well from both tanks when selected individually, and it the BOTH position.

Currently it's pointing to a more spurious cause of the pressure/flow decrease, with the crud in the fuel servo filter being perhaps coincidental.

I guess I'm trying to keep a very open mind to the similarities between the Bearhawk fuel system, and the early Cessna fuel system that resulted in a number of vapor-lock issues.

I should also reiterate that this didn't cause any change in the engine sound and I would have gone completely unnoticed if it hadn't been for me setting the low fuel pressure limit at 20 psi and triggering a warning. In all subsequent flights the data download shows no reoccurrence of the issue.