What's new
Van's Air Force

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

Cracked Fuel Line Flare

crabandy

Well Known Member
During preflight this morning I caught a faint fuel smell when I opened the canopy, I re-adjusted the seat backs and thought I could still smell a bit of gas. There is a bit of a gap around the fuel selector cover and with a flashlight I thought I could see a bit of moisture. After I removed all the covers and started pumping fuel through the lines I found fuel seeping out the outlet connection of the fuel valve going forward to the fuel filter and fact pump. After removing the fuel line there was no visible crack, after a little bit of prodding I found it.

9A1EA385-D738-400D-9E9F-845AADFADA4F_zpszac9cgbq.jpg

B5335717-39B3-4559-8D12-0FE4879F0B89_zpsdzuxbfwd.jpg


I literally just finished my first condition inspection, just over an hour flight time since unbuttoned everything up. During the condition inspection there was no sign of any fuel from the previous 96 hours, all connections were tight and I pumped the remaining fuel out of the tanks with the boost pump while everything was opened up looking for leaks.

I remade my original lines prior to final assembly after reading this post several years ago:
http://www.vansairforce.com/community/showthread.php?t=78240

I made the flares with the Rigid RFT37 and this "how to:"
http://www.gen-aircraft-hardware.com/images/pdf/GAHco_Flaring_Stainless_Tubing.pdf

I thought I did a decent job, perhaps not. It is the line I bend slightly to remove my Earls 85 Micron inline fuel filter, is it possible that slight bending of the fuel line without loosening the flare nut on the fuel selector would cause this? I tightened via the Flats From Wrench Resistance charts, did I over tighten it?

Thanks,
Andy
 
Last edited:
Bad Flare.

It looks like it is possible the flare was overly thin in the beginning. I just went to the shop and torqued, over torqued actually, a connection on spare material I have, then removed the nut inspected and repeated. I don't think there is any way that torque alone did this, it was the result of the flaring process. Make a few and measure the flare thickness.

To be sure, make a new a flare and measure the diameter and thickness of the flange before and after, then deliberately over torque and remeasure. Then repeat.

YMMV
 
B5335717-39B3-4559-8D12-0FE4879F0B89_zpsdzuxbfwd.jpg



I made the flares with the Rigid RFT37 and this "how to:"
http://www.gen-aircraft-hardware.com/images/pdf/GAHco_Flaring_Stainless_Tubing.pdf

I tightened via the Flats From Wrench Resistance charts, did I over tighten it?

Thanks,
Andy

From the photo, it looks like the material is very thin where it cracked, compared to the outer bell of the flare.

The tutorial you linked if for stainless------------lot harder than alum---------did you possibly overdo the flaring operation following the stainless instructions exactly?
 
Last edited:
To my eye there appears to be a change in the flare angle, starting at the crack. Are you sure you used an aviation flare, not an automotive one? They're different angles.
 
I thought I did a decent job, perhaps not. It is the line I bend slightly to remove my Earls 85 Micron inline fuel filter, is it possible that slight bending of the fuel line without loosening the flare nut on the fuel selector would cause this? I tightened via the Flats From Wrench Resistance charts, did I over tighten it?

Thanks,
Andy

I don't think so Andy. This is a common set up and I have been removing my filter having to slightly bend the line out of the way to get the filter out for a long time.
I think the other folks have some good ideas to check.
 
Looks like it's time to play with the flaring tool some more, very possible I just over-flared. If I recall correctly I set the tubing flush with the top of the flare tool, lubricated and rotated 1/2 turn forward and 1/4 turn back until the clutch broke free. I did have to file down the flare to get the Bnut to slide over the flare on some of my connections.

BillL,
I found the max diameter I'm shooting for (basically the Bnut slides over it), what about the thickness of the material?

Thanks,
Andy
 
My .02 worth. 3003 does get pretty thin after flaring, both at the mouth of the flare, and the contact point where the male seals. This is especially true when you tighten the flaring tool to make the flare. You think that tighter is better, but actually the flaring cone is thinning out the material. Combine that with the sleeve pressed against the back side of the flare AND at the inner point of the flare. So when you tighten the nut, the sleeve compresses an already thinned and soft tube. Add some repeated disassemblies, and maybe some minor misalignment, and oops, we have an issue.
We use 304 stainless, and 5052-O aluminum, for the extra strength.
If you are going to use 3003, try NOT cranking down on the flaring cone when the flare is made. Inspect this closely, paying attention to the inner area of the tube as well as the wall thickness at the mouth. If its .020 or less, dont use it--it will eventually fail.

Again---my .02 worth.
Tom
 

I think a key to the problem lays in the diameter change of the tube right below where the flair angle starts.
There should not be a step in the tube diameter at this location.
I suspect this was caused by one of two things.

1. The B nut was tightened excessively which caused the sleave to compress a portion of the flair, leaving the step that is visible at the base of the flair.

2. The flairing tool you used is not forming the flair properly.
 
Looks like it's time to play with the flaring tool some more, very possible I just over-flared. If I recall correctly I set the tubing flush with the top of the flare tool, lubricated and rotated 1/2 turn forward and 1/4 turn back until the clutch broke free. I did have to file down the flare to get the Bnut to slide over the flare on some of my connections.

BillL,
I found the max diameter I'm shooting for (basically the Bnut slides over it), what about the thickness of the material?

Thanks,
Andy

I think the issue is the clutch release. I used a 45 deg version of this tool from the rental store. I lubricated it do the tip rotated free and the threads were clean but oil coated. I flared some 3/4" hard copper for my air compressor. Both on this tool and the my 37 deg Parker Roloflare, when operating correctly it is a light touch to realize when the flare is complete, only a slight increase in torque, this 3003 aluminum is quite soft and weak.

My flare flange is the same diameter as the sleeve, and thickness is only slightly less than the tubing itself. Do a couple with the "feel" in mind and you will quickly recognize when the flare is complete. Over torque one and look at it too.

Edit: I measured a few flare flange thicknesses - all were virtually the same as the tubing thickness - 0.35"
 
Last edited:
It may be that the tube was a little short, and the B-nut put tension on the tube as it was tightened. It seems to explain the stepped flare (different angle below the crack) that we see in the picture.

As one is tightening them, B-nuts should go from loose to very snug in a very small number of flats of the nut. If they seem to start building up torque slowly, there may be aluminum deformation going on. The flare of the tube should rest against the male fitting and be co-axially aligned prior to tightening the B-nut.
 
To my eye there appears to be a change in the flare angle, starting at the crack. Are you sure you used an aviation flare, not an automotive one? They're different angles.

I looked again, and I take my post back.
I agree with others here, it was over-flared (should fit thru the B nut threads) and over-torqued.
 
Tubing

The mutilation on the OUTSIDE of the tube is completely unacceptable. If any of this existed before the material was flared, the crack may have started in one of these deep scratches.
Regarding Tom Swearengen's comments on tubing, I only use 5052-O. With all due respect I would not use 3003 on a tractor.
 
I only use 5052-O. With all due respect I would not use 3003 on a tractor.

Everyone is entitled to their own opinion, but lets not loose site of the fact that what ever material is used to do a specific job, it still requires proper processes and techniques be used for it to perform well.

Someone could pend a fortune using the materials they used on the space shuttle, but if not used properly, there is no assurance that it will perform well.

Actual service history says a lot.... there are literally thousands of RV's safely flying with lines made from 3003 tubing....
 
The mutilation on the OUTSIDE of the tube is completely unacceptable. If any of this existed before the material was flared, the crack may have started in one of these deep scratches.
Regarding Tom Swearengen's comments on tubing, I only use 5052-O. With all due respect I would not use 3003 on a tractor.

Same here. Only 5052-0 for me (both fuel and hydraulic lines). The 3003 is too soft. This has been discussed in detail on VansAirforce before.

My best guess is Vans supplies the 3003 because it is cheap cheap cheap and because it comes in rolls so is easy to freight. The 5052 is far superior but it comes in long straight lengths so there are logistical problem in transporting it.

Of course I agree with Scott in that it is useless buying superior tubing if you have totally inadequate fabrication skills.
 
Last edited:
I'm fairly sure now that I over-flared the tubing by using the clutch mechanism as my stopping point. I found this reference online,

AB9F1402-DA4C-4D51-A70C-73423938B7F5_zpsjfm1hess.png


My flair is over the max diameter, sticking past the sleeve

A4E41681-0067-4883-AFA7-6A973987C7AD_zps63uqhyas.jpg

12D6357B-A74C-4E28-B3C7-8F7CF541F497_zpsfmvpwgy6.jpg


As much as I hate to say it, looks like I'm going to be redoing all my hard fuel lines :eek:.
 
I routinely use the Rigid RFT37 at work with 3/8" .028" Stainless so I'm familiar with it. The few times I have done copper line with it, It always over flares. I believe the clutch mechanism is simply too tight for soft material. While the clutch mechanism is a nice feature that allows you to burnish the inside face for the flare, I don't find that necessary when doing soft tube like copper or aluminum. I would suggest trying a different flare tool, I get good results with the Parker Rolo Flare. This tool does not have a clutch, so you will need to practice on a few pieces to get the feel of how tight to thread the flaring cone down.
 
...Actual service history says a lot.... there are literally thousands of RV's safely flying with lines made from 3003 tubing....

There are probably plenty still flying around with the plastic brake pressure lines too... But that doesn't mean we ignore a more suitable material.
 
I'm fairly sure now that I over-flared the tubing by using the clutch mechanism as my stopping point. I found this reference online,

My flair is over the max diameter, sticking past the sleeve

As much as I hate to say it, looks like I'm going to be redoing all my hard fuel lines :eek:.

Good work getting to the root cause! Stuff happens, on the plus side, your due diligence found a solution quickly.
 
There are probably plenty still flying around with the plastic brake pressure lines too...

Your right, there are. Probably thousands.

Nothing wrong with people wanting to use something different ( and others making a business out of fulfilling that desire), but I get tired of the suttle implications that others who do not choose to do the same, are fools.
 
I'm at the point of making fuel lines. Out of curiosity I just made some test flairs with my Parker flaring tool then tightened them to a mating 3/8 fitting. When tightening the nut I could feel when the nut becomes firm against the flair and just a little more tightness felt fine. Then I started progressively tightening more and more checking the flair and tightening again. I noticed that once over tightened then the torque goes down and feels like it's not tight enough but it is stretching the flair and thinning it. Feels like stripped bolt sort of. Disassembling the fitting revealed a flair that looked exactly like yours. What I learned is that over tightening is really easy to do. 1/4 inch tubing is way easier to over tighten.
 
Again--my .02 worth. YES, I do agree with Scott that there are many RV's flying with 3003. Yes, it is relatively inexpensive, and for the job it does, it does it well. Yes, like everything else, to get the best performance from it, it needs to be fabricated properly. I'm not flaming Andy, far from it.
Yes there are some 'minor' differences in fabricating the different tubing. For me, I had to learn to adjust the amount of tube that extends out of the flare bar to get the proper end result. Yes---slightly different (not by much) for 3003, 5052-O, 6061T6, and 304 stainless. I DO use a RFT37 that I really like. HOWEVER---I did have to learn the little tricks to make good flares on the different tubes I was fabricating.

Not everyone one on VAF is going to agree or disagree with 3003 tubing, Like plastic tube for brakes versus teflon hose. We all have our opinions. NO --we arent building space shuttles---well sort of--but we are traveling in these wonderful machines. WE all have had little issues with things about these planes. BUT---we can all take our experiences and knowledge and help a fellow RV brother to solve an problem that could have been catastrophic.
ANDY---good catch on finding the issue---we will all help you to fix the problems.

Tom
 
...Nothing wrong with people wanting to use something different ( and others making a business out of fulfilling that desire), but I get tired of the suttle implications that others who do not choose to do the same, are fools.

3003, while perfectly adequate for this application if properly fabricated, requires a skillset that is largely absent from the average homebuilder. Both my Rocket and RV-8 were done with 3003 and were both "unairworthy" due to poorly executed flares. Add that to the long and growing list of examples on this forum that show the same thing, and eventually one must come to the conclusion that a more forgiving material should be employed. This is no way a slam on the OP - this issue is clearly widespread. And the fact that "thousands of RV's are using it" is in no way a guarantee that they are all airworthy. In my experience, 5052 is far more forgiving than 3003 both to fabricate and maintain. Even stainless is easier to get right than 3003.

So only speaking for myself, I don't consider the use of 3003 "foolish", but you better know what you are doing. And for what it's worth, I DO know how to fabricate tube assemblies... And I reserve 3003 for mock up work.
 
Andy, good catch and great decision-making. As painful as it might be now, re-making your fuel lines will give you peace of mind when travelling with your girls.

If you'd like, you're welcome to my Parker Rolo-Flare for as long as you need, and I also have a rather large roll of 3/8" 3003 tubing that I have no plans for. That is, if you're not planning to use 5052. I might even be able to deliver them to OWI.

I agree with others that 3003 is adequate, however 5052 does have much better fatigue properties.

You're lucky that you don't have a nosewheel model - re-making the lines from the tank to the fuel selector is no treat. Ask me how I know :p
 
Last edited:
RFT-37

The clutch break on the RFT-37 results in thin flares wider than spec. I called and asked about adjusting it and was told it was factory set and not adjustable.
I place the tubing in the jig per instructions but stop at about two turns past initial contact. Then a half turn back and forth to burnish and remove it. Make a few practice flares and document the number of turns. It's consistent after that. YMMV
 
No root cause yet

If you look closely at the photos, the material is not too thin. What appears to be thin is just the sharp edge of the fracture hiding the rest of that section. But the other side of the fracture shows the actual thickness of the material.

However, the flair has been sheared. The flair starts out fine as a cone, but the turns and forms a cylinder for a while, then continues as a cone. This cylinder appears to be a shear failure of the material. The actual fracture occurs where the material has yielded the most.

Assuming that the flair was overtightened is jumping to conclusions. What could cause this sort of shear? Was the flair made to the wrong angle? what tool was used to make he flair?
 
I always double flare my lines, it's a mandrel that rolls the tubing inward slightly before actual flare, it increases the thickness of flare. have never had flare problems after 40 airplanes
 
What do you use to do this?

I always double flare my lines, it's a mandrel that rolls the tubing inward slightly before actual flare, it increases the thickness of flare. have never had flare problems after 40 airplanes

I have the Rigid 37 deg tool, which I have used to do my hard lines. Luckily, I haven't done that many (only the ones in the wings). However, I'm now reconsidering them and may go back and redo them with 5052. If the Parker RoloFlare doesn't have a clutch (i.e. you have to learn the right point to stop), couldn't the Rigid tool be used in the same manner? Then I would still have the "clutch feature" available for doing stainless steel tube later in the build..
 
Double flair is (was?) standard practice in automotive brake lines---been doing them for decades---havent seen it in aircraft, but no reason not to that I can think of.

The little round gadgets are the key, they are a gauge to show how much the tube extends, and also they are used to start the inverted part of the flare.

This one is automotive, makes 45* flares-------I suspect you could use the double flare insert with an aircraft 37* tool??

http://www.globalindustrial.com/p/p...JvGtWuSVCh1V0Viyyqi-fsSnkga_2TOkMQaAss98P8HAQ

GEO_157.jpg
 
Last edited:
I have the Rigid 37 deg tool, which I have used to do my hard lines. Luckily, I haven't done that many (only the ones in the wings). However, I'm now reconsidering them and may go back and redo them with 5052. If the Parker RoloFlare doesn't have a clutch (i.e. you have to learn the right point to stop), couldn't the Rigid tool be used in the same manner? Then I would still have the "clutch feature" available for doing stainless steel tube later in the build..

That's exactly how it has to be used. The clutch break is way to tight for 3031. It only takes a few flares to find the sweet spot in # of turns past initial contact. I'm glad this thread started. I plan to do everything with 5052. The vent line in the tank will stay 3031.
 
technique

I have the Rol-Air 37 degree flaring tool. If you pick it up, you would assume by looking at it that you form the flares IN the die block. But the instructions are very specific that the tube must extend a specified length above the die block and the flare is actually formed ABOVE the die block by the rotation of the cone only. I don't know this is what happened with the OP's example, but forming the flare within the die block will cause thinning of the flare.
 
Thanks for everyone's input and replies, I really do appreciate them even if I'm being flamed ! At least I'm not being flamed in the cockpit....Awareness/education is a good thing.

I almost didn't catch this, the fuel smell was barely detectable and I had just inspected the fuel system during the condition inspection. I've flown a lot of certified airplanes where a little bit of fuel smell in the cockpit was "normal."

I think Van's puts out a quality product with quality service (more like exceptional), their aircraft and success speak for themselves.

In the spirit of my "Education," I decided to run up to AirpartsInc since it's sort of in the neighborhood and pick up some 5052-O. I'm going to play with it, some 3003 and my tools.

Thanks again,

Andy
 
Last edited:
Andy, good catch and great decision-making. As painful as it might be now, re-making your fuel lines will give you peace of mind when travelling with your girls.

If you'd like, you're welcome to my Parker Rolo-Flare for as long as you need, and I also have a rather large roll of 3/8" 3003 tubing that I have no plans for. That is, if you're not planning to use 5052. I might even be able to deliver them to OWI.

I agree with others that 3003 is adequate, however 5052 does have much better fatigue properties.

You're lucky that you don't have a nosewheel model - re-making the lines from the tank to the fuel selector is no treat. Ask me how I know :p
I appreciate it Kurt very generous, I'm going to play with my Rigid and go from there. I do have a good size roll of the 3003 as well from my second go with the fuel lines.

Yes, the slacker in me says to just fix the one line......It'll be fine....the others probably won't leak.....just keep a close eye on it and you can catch it before it becomes a problem........It sure is nice flying weather right now.....
The other part of me recalls the thread where the pilot jumped out of his burning plane without a chute.
 
If you look closely at the photos, the material is not too thin. What appears to be thin is just the sharp edge of the fracture hiding the rest of that section. But the other side of the fracture shows the actual thickness of the material.

However, the flair has been sheared. The flair starts out fine as a cone, but the turns and forms a cylinder for a while, then continues as a cone. This cylinder appears to be a shear failure of the material. The actual fracture occurs where the material has yielded the most.

Assuming that the flair was overtightened is jumping to conclusions. What could cause this sort of shear? Was the flair made to the wrong angle? what tool was used to make he flair?

Brice, I received no comments or replies to my earlier post in this thread regarding a potential cause. I still suspect the line may have short, such that the B-nut was putting tension on the flare. The photo evidence supports this.
 
Thanks for everyone's input and replies, I really do appreciate them even if I'm being flamed ! At least I'm not being flamed in the cockpit....Awareness/education is a good thing...

Andy

There's a difference between constructive criticism and "flaming" for entertainment purposes. Though this forum has plenty of both going on at times, I can appreciate your willingness to put your ego on the line for the sake of "our" education. I've had to eat crow in the past and I'm sure I will again.

To your particular example, there are still quite a few tooling/emery cloth marks in a very critical, highly stressed area. Personally, I strive for a polished appearance. That said, that unusual "stretched" area of the flare itself is something I've never seen before. I think it's plausable the line was too short, but it would be nice to verify.
 
I predict the verdict is to use 5052-O after trials

In the spirit of my "Education," I decided to run up to AirpartsInc since it's sort of in the neighborhood and pick up some 5052-O. I'm going to play with it, some 3003 and my tools.

Thanks again,

Andy

after you experiment with flares of 3003 and 5052 I predict the verdict is 5052.
 
after you experiment with flares of 3003 and 5052 I predict the verdict is 5052.

I have used both, first started with 3003 that came with the kit then changed to 5052 for a stronger tubing.

I did not notice much difference in my flares but the 5052 certainly is a stiffer and stronger tube.
 
So I flared the 3003 and the 5052 with my Rigid RFT37 tool, I inserted the tubing flush with the face of the tool and used a couple drops of lube. I cranked the flare until the clutch released, here's the results between the two.

C2C849A2-3D88-43DF-9668-492FD7374278_zpsyghk1gmn.jpg


Here's the results of 3 turns of the flare tool for the 3003 and 4 turns for the 5052, pretty consistent.
08B08F31-A55B-4642-A96D-08DFE4E5202F_zps6og1rdhf.jpg


I used this chart as a reference for flare diameter,

FCB56AED-9FDC-4AAF-B4E0-7AD43B1B904D_zpsjynuipqa.png


When the 3003 is over flared per the flare diameter it loses a lot of thickness, I wasn't able to over-flare the 5052 because of the clutch. When the 3003 was flared to the correct diameter it retained the same thickness as the 5052.

Both the 3003 and the 5052 felt way to tight wit 2 flats past finger tight, I need to compare torque vs flats from wrench resistance (FFWR). At 2 FFWR the flare gained a shoulder from the nut/sleeve on both alloys.
After flaring but before tightening,
95E417BC-7D56-40C1-9B67-DD65BE09F72F_zps7xt3jrtt.jpg


After tightening the same flare 2 FFWR,
3FCB121A-1944-4F27-97E4-E0BDBD289AF4_zpsnjqexxw0.jpg


I'm guessing that I first over flared the original 3003 tubing (it was .510 diameter) and then over torqued the tubing causing the eventual crack, all in all it was pretty good to me considering what I did to it. It was about .017-.020 thick.
 
Last edited:
Because it was easily accessible ($11 for 6 feet) it seems to have a larger tolerance to operator error I chose to start remaking my fuel lines with the 5052. That being said I have no fear of using the 3003 now that I think I know a little more about flaring.

The flares of the 5052 look smoother/better than the 3003.
D8DA1FE7-378D-4FDC-BDE7-E4A462F188D8_zpsvrds51zy.jpg


My first redo was a nice fit, hopefully the rest will be just as easy.

F241AF8E-86D0-4AFE-8933-8EA5A3A15157_zpsiciqimmi.jpg


2467E040-1CC0-49CB-A65D-5AA3008A9A56_zpsiwweofda.jpg


E5673989-FF06-4CBB-9B7F-62ED7ABFE986_zpsvrhpdmnh.jpg
 
Andy, I see there is circular line on the inside of both your flares. I would polish up the cone on you flaring tool. There might be some aluminum transferred to the forming cone. Also on the outside of the flares the sharp corner in pretty severe. I would consider breaking that edge on the flailing tool.
 
Andy, I see there is circular line on the inside of both your flares. I would polish up the cone on you flaring tool. There might be some aluminum transferred to the forming cone. Also on the outside of the flares the sharp corner in pretty severe. I would consider breaking that edge on the flailing tool.

Now that we are to the final improvements in techniques, yes, I take my ends (took) and buffed on the fine scotcbrite wheel to eliminate tool marks. I even took some 1000 grit and polished the clamping die on my Rolo Flare (parker) .

I really don't know if it makes a big difference.

Andy, kudos for this whole process. It is clear that many others have the same flaring tool and could (or have) unwittingly done the same thing. There are just so many details to learning in a build. All of us refine techniques as we go.

The last pictures of flares look very very good. Thankfully, Vans designs are quite forgiving in many respects. It was not luck on Vans part.
 
Nice work. This is also a testimate to how reliable these fittings are, even when improperly done. They usually leak well before manifesting into a more serious issue.
Great catch and good thread.
 
Looks mucho better!

...and I'm another who knows the pain of the fuel line replacement process on a completed airplane. I've run through a bunch of new 5052 behind the firewall, and an equal amount of stainless forward. My scrap box of 3003 and oil soaked firesleeved hose assemblies gets bigger all the time.
 
Last edited:
Torque values

Just to let you know you are not alone in redoing fuel lines. I had mine all done in 3003, was not particularly happy with them and subsequently redid them in 5052 before first flight. Second time around is much easier as the patterns are made.

The spec on torque/flats is here: http://www.vansaircraft.com/pdf/Torque_Spec_Aluminum_Fittings.pdf

I am no expert so this is a serious question. I am not questioning the post or the link.
Why are the values in AC 43-13 Table 9 lower. The Vans doc mentions "high performace". Which values do we use?
 
Just to let you know you are not alone in redoing fuel lines. I had mine all done in 3003, was not particularly happy with them and subsequently redid them in 5052 before first flight. Second time around is much easier as the patterns are made.

The spec on torque/flats is here: http://www.vansaircraft.com/pdf/Torque_Spec_Aluminum_Fittings.pdf

Keep in mind that this chart is for hose end fittings.
The seating material in the fitting that produces the seal is much harder than the flared end of a line made from tubing (whether that be 3003 or 5052).
 
Back
Top