What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

RV-8 fuel starvation accident

Maybe they should ask.....

Who removed the fuel pressure sender and line that was connected to that restricted fitting?

Was the plane taken to the shop with that restricted fitting capped off?

I suspect there were many oversights here, by more than one person.
Given that’s the only practical place to tap off for the fuel pressure, ima guessing there was no fuel pressure gauge. Which would be highly unusual for a new G3x setup.
It appears to have had a fuel pressure sender before.
So many questions
 
One thing that might be relevant is that the Fuel Pressure Sensor is an option which is not included in the G3X Lycoming Sensor kit.
 
I suspect there were many oversights here, by more than one person.
Given that’s the only practical place to tap off for the fuel pressure, ima guessing there was no fuel pressure gauge. Which would be highly unusual for a new G3x setup.
It appears to have had a fuel pressure sender before.
So many questions

There's only one question, but you have to keep asking it until you get to the root cause(s):

Why?

Why was the line connected to the smaller fitting rather than the obvious larger one? Answer that, then ask again...why? Keep asking until the *real* root causes are identified.
 
There's only one question, but you have to keep asking it until you get to the root cause(s):

Why?

Why was the line connected to the smaller fitting rather than the obvious larger one? Answer that, then ask again...why? Keep asking until the *real* root causes are identified.

I guess the root cause is somebody with no idea monkeyed with the fuel system.
how far further up the chain do we go?

My point is more that this is like any other incident in that the many holes of the swiss cheese lined up. There were likely many places to break the chain of events.
-Was the pilot staring at a big red X on his nice new display where the FP was supposed to be? or was nothing configured and there was no FP gauge at all?
-Is a FP gauge mandatory?
-did the installer omit it because it wasn't in the sensor pack box - its omitted so you can specify the PSI range, not because it is an "option" - Ive done this 3 times - its intended to be there and aways has been for me. Anyone who'd done this before (or read the install manual) would know this.
-I noticed all the log entries referred to it as a G3 system. once is a typo. Every instance is something different.

Anyway, its very unfortunate. I hope the owner makes a full recovery.
 
Restrictor fitting 'generally' have that reduced orifice in the NPT side, like the first pic here of one we mass produce (like several other companies, with a minor difference). The second pic is of an example of some I custom make (just happens to be a 45* in the pic, but have done straight and 90*s).

With the commercially available one with the restrictor in the NPT side, if its already installed, a mechanic may not know its restricted without removing it and inspecting it. I one I make custom for some builders has a THREADED restrictor in the FLARE side. Pretty easily seen.

My reasoning was the ID of the port hole in the flare side more closely matched the thread OD of the restrictor, so threading a commercially available restrictor was less expensive than making a custom one. 2) a threaded restrictor isnt coming out, versus a pressed in version that 'might' be subjected to excessive heat. A typically available restrictor fitting 'generally' has that restrictor pressed into the port on the NPT side. Normally brass pressed into steel, so 2 different expansion rates. Not knowing FOR SURE what the manufacturers were doing dimensionally, Steve and I decided a threaded orifice insert was the way to go. Notice the screwdriver slot. We didnt want an insert coming loose and getting sucked into a cylinder. That would be an obvious bad day.
3) The custom ones I make go a step further---being in the flared side ( especially on a 45* or 90* fitting) there is no way its coming loose, or getting sucked into a cylinder. AND----its captured not only by the threads, but by the hose end.
And for those that had some cokeing issues, with the flare side restrictor, it seems thats not an issue.

Tom
 

Attachments

  • 2503R-0402 pic 2.jpg
    2503R-0402 pic 2.jpg
    126.6 KB · Views: 109
  • 2503R-0402.jpg
    2503R-0402.jpg
    394.1 KB · Views: 88
... A #4 line could feed enough fuel flow for taxi and run up (with no restrictor fitting), but unable to flow 18+ gph at takeoff power...

Minor correction to the point above - Every bendix FI system I have flown behind or seen has a -4 line from the servo to the flow divider. We know that the servo is "dry" flow, so 100% of the fuel required for the engine runs through a significant length of -4 line. This includes engine sizes up to and including the 540. And I can tell you with absolute certainty that the 540 in my Rocket was fed with -4 line and burns a LOT more than 18 GPH in TO and initial climb.

Not relevant to this accident, but thought it noteworthy.

Concerning the accident, it seems someone went considerably out of their way to do it wrong. Did the airplane fly to the shop with the carb fed by a -4 line? And was part of the avionics install to swap out the existing unrestricted fitting and install the restrictor version? Its somewhat understandable that the technician might simply follow the BOM and instructions and install the restrictor fitting as part of the avionics installation (it is in the right spot after all), and then completely miss the fact that the other end of the hose was connected to the carb inlet. This idea that its "understandable" to me only comes after watching greenhorn aircraft manufacturing technicians assemble an airplane on an assembly line with ZERO concept of what does what or why. They just follow the assembly pictures and if the planning happens to have omitted a critical O ring in a fuel fitting, it does not get one. If the planning misses a value on a torque wrench setting and tells the tech to apply 1000 inch pounds on a #10 fastener, they will snap that head off every time without a second thought.

"Critical thinking" is not a common thing anymore, it seems.
 
Last edited:
Digging in to the NTSB docket, there are 5 items to VIEW at the lower right. The third is the Maintenance Record 5/24/21. Shows who did the work. Also indicates “Aircraft was Groundrun by the owner to confirm all systems operational”.

Something interesting here....

You can see the older work done by Precision Aircraft (Frank Hinds) for the annual and other work - note the dates and Tach time.

Then Poor Boy did the panel install and lists the tach time as 590 hours on May 24, and indicates the owner did the runup to check function.

The NTSB report says the accident flight was the FIRST flight since the panel install, and the time on the aircraft was 600 hours at the time of the accident on May 27. Where did the missing 10 hours come from? You know good and well that the airplane did not FLY with the restrictor fitting in place. Was that maybe a case of "590 hours is about 600 and good enough for this report"? Seems pretty sloppy for an NTSB report...

Now - another thing to consider here - PBA is known to work with pilots/owners to assist in installation and especially troubleshooting of misbehaving systems, I've done it myself with them, and the mx records for this aircraft also show that they changed the oil/filter "with assistance from the owner" so we know at least SOME of that was occurring on this airplane during this visit. It's entirely possible, and entirely unknown at this point, whether the pilot/owner may have been working on the fuel system himself while the shop did the electronics install. I'm sure those questions will get asked by someone at some point. Regardless - a full power runup WITH REFERENCE TO THE NEWLY INSTALLED FUEL FLOW would have brought this to the attention of the pilot.
 
Last edited:
The best takeaway here may not be to definitively find the cause of this accident, as much to avoid it happening again. Maybe call it a lesson learned.

Dan made the comment that setting yourself up to fail by using an oversized hose on the restricted fitting is a trap, and a valuable caution for any builder. We're builders here.

Tom's approach of building a good fitting that makes the restricter visible is great stuff, but once again, that requires the mech to take a minute to LOOK at the fitting and understand what's going on.

Anytime you're doing something (like setting your wallet down somewhere unusual or a glass on the edge of the workbench) and your spidey senses saw quietly "that's not the best idea here ... ", you need to listen! Make the decision to act right then and fix the trap, or take the extra time to do the job right.

We all have swiss cheese, but also a lot of control over all those holes

CHeers
 
Not sure I agree with this, quite a few accidents on record that were likely due to carb/vapor lock issues, there is definitely enough fuel in the bowl for a quick run up and get you off the ground, but not for long.

vapor lock typically reduces flow, not eliminate it. probably Enough flow to support idle, taxi and quick runup, but not enough to keep from going too lean at WOT take off.
 
Minor correction to the point above - Every bendix FI system I have flown behind or seen has a -4 line from the servo to the flow divider.

unfortunately that is apples to oranges. The bendix is pushing 12-14 PSI at high flow rates through that -4 hose, therefore can move more fuel in that same hose ID, compared to a carb that is being fed at 1 PSI (minimum allowable is spec'd at 1/2 PSI). The higher the pressure, the greater the flow rate for any giving ID. Simplified explanation as other variables are also involved. I put a restrictor on my bypass fuel line. A #60 drill orifice (around 1/6 the size of a -4 hose) nets around 9 GPH at 25PSI; It would be a small fraction of that at 1 PSI. Pressure makes a BIG difference.

Larry
 
Last edited:
unfortunately that is apples to oranges. The bendix is pushing 12-14 PSI at high flow rates through that -4 hose, therefore can move more fuel in that same hose ID, compared to a carb that is being fed at 1 PSI (minimum allowable is spec'd at 1/2 PSI). The higher the pressure, the greater the flow rate for any giving ID. Simplified explanation as other variables are also involved. I put a restrictor on my bypass fuel line. A #60 drill orifice (around 1/6 the size of a -4 hose) nets around 9 GPH at 25PSI; It would be a small fraction of that at 1 PSI. Pressure makes a BIG difference.

Larry

Yes, pressure makes a big difference. A broken -4 line on a 5000 PSI hydraulic system will drain a reservoir in a heartbeat. That said, if the accident airplane didnt fly to the shop (and hundreds of hours prior) with a -4 fuel line installed then someone did a LOT of work to do it wrong.

That means they removed the existing -6 line and sourced a cap. Then installed the -4 fitting, fabricated a new supply hose, and replaced the -6 fitting on the carb with a -4. They then re routed the new hose (presumably with new Adel clamps), and did a leak/Ops check.

Like I said, a lot of work in the wrong direction.
 
That said, if the accident airplane didnt fly to the shop (and hundreds of hours prior) with a -4 fuel line installed then someone did a LOT of work to do it wrong.

That means they removed the existing -6 line and sourced a cap. Then installed the -4 fitting, fabricated a new supply hose, and replaced the -6 fitting on the carb with a -4. They then re routed the new hose (presumably with new Adel clamps), and did a leak/Ops check.

Like I said, a lot of work in the wrong direction.

Actually, much easier than that, other than finding a cap laying around an AN816-6D in one end of the cube and an AN816-4-4D in the other end is all, just 2 fittings needed, the -4 restricted fitting was already there I suspect.

I think from now on I'll use Tom's fittings (restrictor clearly visible without removing) and paint them red.

Had the restrictor been clearly visible in the fitting I doubt this would have happened, and there is no doubt the engine would run just fine with a -4 line from the pump to the carb, so the size of the line is really not an issue.

And as previously mentioned the 'owner', which may not be the builder, could have done this work.
 
Last edited:
Just to satisfy my curiosity, I used a 24 inch length of -4 line to siphon out a 5 gallon bucket of water. Hardly scientific, but it took less than 8 minutes to empty the bucket. That translates to more than 37 GPH - with no pump.
 
Actually, much easier than that, other than finding a cap laying around an AN816-4 in one end of the cube and an AN814-4-4D in the other end is all, just 2 fittings needed, the -4 restricted fitting was already there I suspect.

I think from now on I'll use Tom's fittings (restrictor clearly visible without removing) and paint them red.

Had the restrictor been clearly visible in the fitting I doubt this would have happened, and there is no doubt the engine would run just fine with a -4 line from the pump to the carb, so the size of the line is really not an issue.

And as previously mentioned the 'owner', which may not be the builder, could have done this work.

I disagree Walt. An an 814 – 4 –4 D fitting is not all that common. Just by the reasoning that it is unusual to put that small of a fitting connection into that big of a threaded female port. That size of a port is usually that size because it is meant to flow a lot more than any of the smaller fittings. So I think this is being overlooked too lightly in this discussion. That someone had to remove the fitting that was originally in the carburetor and install something that for a lot of shops might even be a special order part.

As for the seatbelt not tight enough comments… That is entirely possible but depending on the impact loads the human body is amazingly flexible and able to displace a large distance forward of where you think it would relative to how tight the are. I’ve seen this many times over the years. This is why I personally don’t wear seatbelts adjusted to where they’re comfortable, I adjust them to where they’re actually slightly uncomfortable, and even then that may not be enough to keep parts of your body from contacting structure within the cockpit.
 
I disagree Walt. An an 814 – 4 –4 D fitting is not all that common. Just by the reasoning that it is unusual to put that small of a fitting connection into that big of a threaded female port. That size of a port is usually that size because it is meant to flow a lot more than any of the smaller fittings. So I think this is being overlooked too lightly in this discussion. That someone had to remove the fitting that was originally in the carburetor and install something that for a lot of shops might even be a special order part.
.
Typo, I meant 816-6D and 816-4-4D (not 814), that would give you a 1/4 in and 3/8 out of the cube, both standard fittings used in the cube.
 
Last edited:
Typo, I meant 816-6D and 816-4-4D (not 814), that would give you a 1/4 in and 3/8 out of the cube, both standard fittings used in the cube.

I don't know which came first here, the chicken or the egg. Did someone put a reducing AN816 fitting into the Red Cube first, and then someone (else) came along and said "Aha, -4 line, must go to this -4 fitting on the pump!"? Or did someone not understand the O-ring boss fitting on the pump, decide for some reason to use the smaller one and cap off the larger one, and then had to use the reducing fittings on the cube?

Or was it all just a bunch of not understanding *anything* and using whatever would screw into something else?
 
Typo, I meant 816-6D and 816-4-4D (not 814), that would give you a 1/4 in and 3/8 out of the cube, both standard fittings used in the cube.

I forgot what the 4–4 D fitting was. My point is that somebody had to install a fitting that adapted from a large female port on the carb to a -4 hose end. This is not something that anyone who had their head in the game should think about doing. If they hadn’t, then they likely would have just installed the hose that was there originally.
 
I forgot what the 4–4 D fitting was. My point is that somebody had to install a fitting that adapted from a large female port on the carb to a -4 hose end. This is not something that anyone who had their head in the game should think about doing. If they hadn’t, then they likely would have just installed the hose that was there originally.

They wouldn't have installed the original hose if they were installing the red cube. That would require making up two new, shorter hoses, and suggests that all the reduced-diameter hardware would have been installed at that time.
 
Steve----that was sort of my point in showing the AE466 hose and the others on the pic. DEFINITATELY a different style hose, although still teflon. Couldnt see enough of it to see the routing, but I surmise that someone just happen to have had that hose and figured why not use it. WE see it alot, where guys borrow a hose from a hangar mate for an install. Most of the time, its the proper size, but lengths are off.

Yes in this case 2 hoses would be needed. Overall, its appears the installation was not thought through, and was more of a convienence. The fact is the supply from the pump through the transducer to the carb, contributed to a mishap. Who, or what group of people installed it wouldnt be my first concern. Thats for the investigators to decide. The knowledge gained would be my interest. There are literally thousands of pics of various installs online. I'll bet that there are that many in Texas. Point is, even if you are a A&P or an IA, there is knowledge to be gained here. There are alot of resources out there if you arent sure about something. Even the best ask questions.

My opinion is mine, and not meant to ruffle any feathers or start a flame war. This is serious business. At least 2 people, possibly 3 or more, had their hand in this. The installer, the A&P that 'inspected' the install and signed it off, and the owner. The ramifications are something to think about.

Tom
 
I forgot what the 4–4 D fitting was. My point is that somebody had to install a fitting that adapted from a large female port on the carb to a -4 hose end. This is not something that anyone who had their head in the game should think about doing. If they hadn’t, then they likely would have just installed the hose that was there originally.

It sure looks to me like the outlet side of the Red Cube has a -6 line on it, and if so, the carburetor fitting likely wasn’t changed. I suspect that the aircraft originally was equipped with a fuel pressure instrument of some sort or at least the stock Van’s FWF kit provisions for it, but no fuel flow. During the avionics upgrade, it was too convenient to repurpose the original fuel pressure instrument line to feed the red cube by simply installing the AN822-4-4D on the inlet of the Red Cube, then using the more typical AN816-6D fitting on the Red Cube outlet along with, I suspect the original -6 carburetor supply line between the red cube and the carburetor. This essentially meant that no new lines were installed and just the -6 cap and the fittings for the red cube needed to be added to the fuel system for the Red Cube installation.

If this work was truly done by a professional technician, I am really appalled!

Skylor
 
Last edited:
It sure looks to me like the outlet side of the Red Cube has a -6 line on it, and if so, the carburetor fitting likely wasn’t changed. I suspect that the aircraft originally was equipped with a fuel pressure instrument of some sort or at least the stock Van’s FWF kit provisions for it, but no fuel flow. During the avionics upgrade, it was too convenient to repurpose the original fuel pressure instrument line to feed the red cube by simply installing the AN822-4-4D on the inlet of the Red Cube, then using the more typical AN816-6D fitting on the Red Cube outlet along with, I suspect the original -6 carburetor supply line between the red cube and the carburetor. This essentially meant that no new lines were installed and just the -6 cap and the fittings for the red cube needed to be added to the fuel system for the Red Cube installation.

This seems likely - and possibly the whole assemblage of fittings was reoriented 90º clockwise to give a more convenient run from the -4 (restrictor) fitting to the carb.

I don't see any provisions for a FWF fuel pressure sensor in the NTSB photos, so I wonder what the thought process there was. Maybe the old one wasn't replaced.
 
I During the avionics upgrade, it was too convenient to repurpose the original fuel pressure instrument line to feed the red cube by simply installing the AN822-4-4D on the inlet of the Red Cube

Sounds about right. So asking again "why?", we'd have to ask whomever did this "Why didn't you use the larger fitting (the size that matched the output port on the red cube) on the inlet to the red cube, and the obviously larger output from the fuel pump, and simply cap the *smaller* one? Why didn't you think that going from a smaller diameter line (pump to red cube) to a larger diameter line (red cube to carb) was unusual and probably incorrect?

We may never know the answer to any of these, unless the owner were to chime in here with what they know, which is unlikely. It's good for learning, though...not just about which fitting on the pump to use, but about how to *think* about what you're doing and whether it makes obvious sense or not; how not to cut corners and just use whatever parts you have lying around without thinking about it carefully; etc.
 
Pressure Fitting

I don't see any provisions for a FWF fuel pressure sensor in the NTSB photos, so I wonder what the thought process there was. Maybe the old one wasn't replaced.

How can you tell? The stock -8 FWF kit has the restrictor fitting threaded into the fuel pump outlet just as shown, and the -4AN line runs to a manifold on the firewall for the pressuresensor. See the attached excerpt from the Van's FWF drawing.
 

Attachments

  • CarbFuelDrawing.jpg
    CarbFuelDrawing.jpg
    224.5 KB · Views: 133
How can you tell? The stock -8 FWF kit has the restrictor fitting threaded into the fuel pump outlet just as shown, and the -4AN line runs to a manifold on the firewall for the pressuresensor. See the attached excerpt from the Van's FWF drawing.

I agree that the collection of fittings is the same, but it looks to me that it has all been reclocked so the -6 (capped) fitting faces forward instead of down.
 
Last edited:
It looks like they re-clocked so the 45* fuel pressure fitting pointed direct to the cube to make it easy to run the line.

Fuel pressure is way more important than that red cube. So did they not have a fuel pressure indicator after the completed install??
 
One additional tidbit of information. The GCR-30P installation instructions do mention the AN816-4-4D as one of the possible fittings for the red cube. This could potentially let someone who is not knowledgeable to think it was acceptable in an installation which it isn't. They do say to use the same size hose as the originally installed hose but it could be a little misleading. I don't see anywhere where it is mentioned to test the fuel flow after installation either.
 
Ray---thats true, but mainly for servo to flow divider installs. The biggest clue would have been the carburetor is the end of the fuel system in this example. In a mechanical injection system, like a Bendix, Avstar, AFP, etc, the flow divider is the last 'hose supplied' accessory on the fuel system. (Unless we build you some hoses to replace the stainless injector lines.) Most all servo outlets are -4 ( Don can correct me on some of the big HP race motor systems he does, but generally speaking a -4 hose from servo to flow divider).

I dont think we'll ever know exactly what was changed during the install, or why. Not at least until the NTSB and FAA get done with their stuff. But I think that information should be released----after all the wrangling gets done.

Tom
 
Just a thought in the discussion of whether enough fuel could flow through that restrictor to get through a run up and takeoff... Don't we always turn on boost pumps before takeoff? What kinda pressure are you pushing with them? 30psi? More? Would it be enough?

6psi will supply a 500hp V8 through a -4 no problem...
 
Just a thought in the discussion of whether enough fuel could flow through that restrictor to get through a run up and takeoff... Don't we always turn on boost pumps before takeoff? What kinda pressure are you pushing with them? 30psi? More? Would it be enough?

6psi will supply a 500hp V8 through a -4 no problem...

Mute point...it was a carburettor not injected so the float bowl will fill and the restrictor will easily fill the bowl for low power operations and a 1700 rpm run up is still low power...so it will show up when leaned out after takeoff when the float bowl is emptying.
 
An .040 reducer is OK for oil, and could be reduced in size, but it is really too big for gas. An .015 restrictor would provide fuel pressure (and the loss of it) and provide a greatly reduced hazard during a hose failure..... Instead of one size fits all.

If a .015 fuel reducer was installed in this situation, this accident would have never happened.
 
If a .015 fuel reducer was installed in this situation, this accident would have never happened.

That's a very astute observation. Similar to ground leaning - lean it enough so that any application of power will bog down and make it obvious.
 
That's a very astute observation. Similar to ground leaning - lean it enough so that any application of power will bog down and make it obvious.

Another one of those posts that makes me wish we had a 'like' button
 
Back
Top