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

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).

Scott, the actual torque value aside - in the launch of this thread, I went to the basement shop and tried some different snug and turn values. It seemed that the 1/2 flat turn after firm was pretty good. A full flat seemed to be too much. Can you confirm that?
 
Thanks Andy

I went back and took a hard look at my lines. None of my flares were cracked, of course they haven't been in service yet either. The wall thicknesses were all a little thin, so I'll be getting some more "practice" with my Rigid flaring tool.:eek: Thanks for starting this thread. It may help others avoid potentially screwing up their lines, due to the combination of tight clutch and soft material.
 
Groundhog day on VansAirforce again

All of the information in this thread (and a whole lot more) was previously supplied in the following 2011 thread on the subject of cracks in flared tubes started by Walt Aronow. With 92 posts and a huge amount of serious information this 2011 thread is still the definitive thread on the subject.

Go here.

http://www.vansairforce.com/community/showthread.php?t=78240&highlight=flare+torque

This much shorter thread will also prove very enlightening:

http://www.vansairforce.com/community/showthread.php?p=589037#post589037
 
Links were not definative

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.

None of the linked information covers what Andy is describing here. It really looks like the wrong flare angle is being used. After tightening the B nut, is should not have "gained a shoulder from the nut/sleeve on both alloys".

NOTE: he said that the same thing happened with both alloys. It is not a material issue.

The material is yielding or reforming when tightening the B nut, not by squeezing the flair to a thinner cross section, but by shearing the flair and reforming the flair as Scott and others have pointed out.

Look again at the before and after photos of the flair. Has anyone seen something similar? This looks like a much more serious problem than just what material to use for fuel lines.

The first picture showed that the failure occurred at the sheared section.

What else can do this other than tightening a 45 degree flair onto a 37 degree AN fitting? Could it be that the fitting is the wrong angle and a 37 degree flair is being tightened onto a 45 degree fitting with a 37 degree B nut, or something similar?
 
B

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


Here is another photo that does not look right. The text states that the fit was tight, but a close look shows that the flare does not match the fitting. It may be that they are just weperated, but if the fit was tight, this could be a 45 degree flare on a 37 degree fitting.
 
B

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


Here is another photo that does not look right. The text states that the fit was tight, but a close look shows that the flare does not match the fitting. It may be that they are just weperated, but if the fit was tight, this could be a 45 degree flare on a 37 degree fitting.

It's most definitely a 37 degree flare on both, pictures can add "new" angles.
I believe the shoulder is the result of the nut being over tightened. I believe the shoulder is about the same dimension of the excess between the sleeve's inner diameter and the tube's outer diameter, over tightening squishes the flare leaving the shoulder.
Feel free to make a flare in 3003 or 5052, connect it to a fitting and tighten it a 2 full flats from finger tight and see what you get.
 
It's most definitely a 37 degree flare on both, pictures can add "new" angles.
I believe the shoulder is the result of the nut being over tightened. I believe the shoulder is about the same dimension of the excess between the sleeve's inner diameter and the tube's outer diameter, over tightening squishes the flare leaving the shoulder.
Feel free to make a flare in 3003 or 5052, connect it to a fitting and tighten it a 2 full flats from finger tight and see what you get.

I agree completely, Andy. That shoulder is indeed the ID of the sleeve. I passed by and purchased some crow foot flare nut wrenches today and tested just one joint to see what torque corresponded to a 1/2 flat past firm. 100 in-lb was way too much. I'll have to make a lot of short pieces and test them to various torques and note the angle turned.

If some engineer gets bored maybe they can tell us what the contact stress is vs torque on (a well lubricated) joint. Past the yield point of the tube material, the torque does not increase until the sleeve has met the high strength of the flare nipple resulting in a paper thin tube flare as it has extruded out of the contact area.

I, for one, still do not know what should be used as a torque/tightening specification for these joints. The Vans chart might not be the correct torque, but I did use the 1/2 flat past tight technique.
 
I agree completely, Andy. That shoulder is indeed the ID of the sleeve. I passed by and purchased some crow foot flare nut wrenches today and tested just one joint to see what torque corresponded to a 1/2 flat past firm. 100 in-lb was way too much. I'll have to make a lot of short pieces and test them to various torques and note the angle turned.

If some engineer gets bored maybe they can tell us what the contact stress is vs torque on (a well lubricated) joint. Past the yield point of the tube material, the torque does not increase until the sleeve has met the high strength of the flare nipple resulting in a paper thin tube flare as it has extruded out of the contact area.

I, for one, still do not know what should be used as a torque/tightening specification for these joints. The Vans chart might not be the correct torque, but I did use the 1/2 flat past tight technique.

Bill,
I also found the 1/2 flat past finger tight to "seem" right too. It felt right with the wrench and it also did not squish a shoulder under a proper flare in 3003 or 5052. 1 flat past finger tight felt too tight and resulted in a shoulder under the flare on the 3003 and 5052.
 
I, for one, still do not know what should be used as a torque/tightening specification for these joints. The Vans chart might not be the correct torque, but I did use the 1/2 flat past tight technique.

Question for those smarter than me: Is the chart in AC43.13-1b for AN818 nut torque on aluminum flared hydraulic tubing acceptable? Is this not the definitive torque spec for our use? The paragraphs preceeding this chart talk about 3003 and 5052 tubing as well.
tahhz4.jpg
 
Looking at the pictures in post number 40, it appears from the longitudinal line squeezed into the tubing and right to the edge of the flare that the OP is forming the flare in the the die block. The instructions that accompanied my flare tool cautioned against forming the flare in the die block and specified it is formed by the the rotation of the cone only, not rotation of the cone against the material squeezing the flare against the die block. Forming the flare against the die block thins the material and creates a stress riser at the base of the flare. CrabAndy, please reread the instructions that came with your flare tool to see if they direct the flare be formed above the die block, not in it.
 
Bill,
I also found the 1/2 flat past finger tight to "seem" right too. It felt right with the wrench and it also did not squish a shoulder under a proper flare in 3003 or 5052. 1 flat past finger tight felt too tight and resulted in a shoulder under the flare on the 3003 and 5052.

Until there is a really dang good analysis, this is what I have used, and will stand by it.

In the "27 years of th RVator" Vans lists 40-65 inch pounds for 1/4" and 75-125 inch pounds for 3/8". That coms right out of The Standard Aircraft Handbook.

The 1/2 flat method using 3003 tubing is right at 50 in-lb. See calc discussion below.

Question for those smarter than me: Is the chart in AC43.13-1b for AN818 nut torque on aluminum flared hydraulic tubing acceptable? Is this not the definitive torque spec for our use? The paragraphs preceeding this chart talk about 3003 and 5052 tubing as well.
tahhz4.jpg

First, let me say, I am an engineer and did some rough calc and tested on the bench. But - YOU MUST SATISFY YOURSELF UNTIL VANS TELLS US DIFFERENT - DON'T TAKE MY COMMENTS AS GOSPEL.

Now, with that disclaimer, the -6 sleeve has about a .040" contact width. 3003 has approximately a 5000 psi yield strength, and a 9/16-18 thread with a .20 friction factor will produce a contact stress of 5000 psi with ~315 pounds of force. Translated into theoretical torque, this is about 30 in-lbs. Taken to the real world, and compared to a dry tighten, then turn 1/2 flat, the nut began to turn at about 45 in-lb. So - draw your own conclusions, but I think that yielding is beginning there, and the torque value will increase with more "flats" because the contact area is made by chamfers in the sleeve. Further displacement yields a larger contact area and takes higher torque, BUT - the contact stresses are not higher. The the material is yielding. This is less than a precise analysis, thus the difference with the real results.

So, why are all the charts off?? Is it because they don't consider 3003 as the material? They do say "ALLOY". If yield stress is 10ksi, then the torque would be twice as much, but 1/2 flat is still good.

For my airplane, and my life, I will be using the 1/2 flat or 40 -45 in-lb method for 3003 tubing and 3/8" diameter.

Critical, definitive discussion is invited. A second, third, critical look is always good for a peer review.

Let it begin.
 
Last edited:
Until there is a really dang good analysis, this is what I have used, and will stand by it.

The 1/2 flat method using 3003 tubing is right at 50 in-lb. See calc discussion below.



First, let me say, I am an engineer and did some rough calc and tested on the bench. But - YOU MUST SATISFY YOURSELF UNTIL VANS TELLS US DIFFERENT - DON'T TAKE MY COMMENTS AS GOSPEL.

Now, with that disclaimer, the -6 sleeve has about a .040" contact width. 3003 has approximately a 5000 psi yield strength, and a 9/16-18 thread with a .20 friction factor will produce a contact stress of 5000 psi with ~315 pounds of force. Translated into theoretical torque, this is about 30 in-lbs. Taken to the real world, and compared to a dry tighten, then turn 1/2 flat, the nut began to turn at about 45 in-lb. So - draw your own conclusions, but I think that yielding is beginning there, and the torque value will increase with more "flats" because the contact area is made by chamfers in the sleeve. Further displacement yields a larger contact area and takes higher torque, BUT - the contact stresses are not higher. The tune material is yielding.

So, why are all the charts off?? Is it because they don't consider 3003 as the material? They do say "ALLOY". If yield stress is 10ksi, then the torque would be twice as much, but 1/2 flat is still good.

For my airplane, and my life, I will be using the 1/2 flat or 40 -45 in-lb method for 3003 tubing and 3/8" diameter.

Critical, definitive discussion is invited. A second, third, critical look is always good for a peer review.

Let it begin.

Bill - all engineering and clinical analysis aside, there is something to say for experience and mechanical inclination. I follow your rules. It feels right, has never produced a problem, and I also will continue to do this. I was trained by an "old timer", a test pilot for North America during the war, and a mechanic for his career which spanned 60 plus years. He never had an "engineered" answer. It was always "this is how you do it", and "I don't care what the book says, what book?"
No disrespect to the fine engineers out there and their contributions, and it never hurts to verify what you think you know with good data.
 
No disrespect to the fine engineers out there and their contributions, and it never hurts to verify what you think you know with good data.

Jon, that is exactly the point. Also, these charts are published and it is good to go by the industry standards, but with 3003 we are NOT ON A CHART and there are a lot of novice builders that want to do a good job and use a "chart", me included. All this discussion had my engineers' spidy senses tingling. Novice Builders NEED definitive information. And . . . the tightening of the 3003 tubing is not well documented, and with clear, unambiguous, guidelines.

I merely wanted to state that there is a reason based in theory NOT to use the chart for 3003. If we look at the "failure" data for fuel systems on RV's it makes one wonder if this might be a contributor.
 
The nut threads are 9/16-18. (.0555" Pitch) Divided by 6 flats would be .0093" nut advancement when turning the nut one flat. Using the 3003 soft tubing allows the nut to move forward and the tube material is yielding, thinning the flair. The charts don't really consider this either. Like Andy discovered, half a flat would be about right. And don't compress the flare in the flaring tool.
 
Jon, that is exactly the point. Also, these charts are published and it is good to go by the industry standards, but with 3003 we are NOT ON A CHART and there are a lot of novice builders that want to do a good job and use a "chart", me included.

Ahh, but 3003 alloy is indeed referenced in section 9-30 of AC 43.13-1b, along with 5052 alloy and even 1100 . The very same section in which the torque chart is found. And no language stating not to use the torque chart on certain alloys.

Edit: The AMT Handbook on Fluid Lines and Fittings also uses what appears to be this same table, and gives additional instructions on torquing flared fittings. Notably, it also specifies that 3/8" tubing is to be double-flared, which is not what the instructions from Vans call for.

https://www.faa.gov/regulations_pol...craft/amt_handbook/media/FAA-8083-30_Ch07.pdf
 
Last edited:
With all of the discussion in this thread regarding proper tightening of flared lines on RV’s, I decided to spend a few minutes analyzing the torque process specified in the kit documentation (specifically, Section 5.14 on page 05-14 of Construction Manual Section 5)

In that section it quotes the values from the Standard Aircraft Handbook. For 1/4 tubing it is
40-65 inch-pounds, for 3/8 tubing it is 75-125 inch-pounds. Note, that these torque values (as in most all cases of torque values on threaded fasteners, unless specifically specified otherwise) are for new dry threads. Any oil or hyd. fluid on the threads will have an influence on the torque reading (would require a lower value to be used to obtain the same true value).
I did a few test samples with ¼ inch and 3/8 inch 3003 tubing with properly formed flairs.
¼ tubing -
I found that 1.5 flats of B nut rotation produced a torque of 40 inch pounds. This produced a barely perceptible imprint of the sleeve on the back face of the flare.
2 flats of B nut rotation produced a torque of 65 inch pounds. This produced a visual, but totally acceptable imprint of the sleeve on the aft face of the flare.

3/8 tubing –
I found that 2 flats of B nut rotation produced a torque of 75 inch pounds. This produced a very slight imprint of the sleeve in the aft face of the flare.
3 flats of B nut rotation produced a torque of 125 inch pounds. This produced a slightly deeper imprint of the sleeve in the aft face of the flare, but no excessive thinning of the flare was noted.

Disclaimer….
A major issue when recommending a “number of flats” tightening process is that it is dependent on how tight it already is when the final torquing is done. The problem is if you say “start with it finger tight”, that level of tightness is different from one person to another. This will result in different final torque values depending on who is doing it.
These number of flat values I have listed, are with the B nut initially tightened only until all looseness / play was removed. No actual “tightening” was done. Even this can have some level of variability because sometimes a slight misalignment makes it hard to detect that point when initially connecting the line to the fitting.
It is for these reasons that use of an inch pound torque wrench with a crows foot is highly recommended when ever possible (using the standard compensation correction formula, or keeping the crows foot oriented 90 degrees to the tool handle). The majority of people (myself included) are not good at judging torque values when using just a plain wrench, and when they try to, they are over torquing the majority of the time.

Based on these test results, I am totally comfortable with the published torque values for the 3003 tubing, but to take possible variability’s into account, it is probably a good idea to shoot for the middle of the torque value range.
 
As Ken says, the flats method on the Vans web site starts from "in contact", not "finger tight", which is quite a difference.

Perhaps we getting a little bit into the RTFM area?
 
With all of the discussion in this thread regarding proper tightening of flared lines on RV’s, I decided to spend a few minutes analyzing the torque process specified in the kit documentation (specifically, Section 5.14 on page 05-14 of Construction Manual Section 5)

<snip>

Based on these test results, I am totally comfortable with the published torque values for the 3003 tubing, but to take possible variability’s into account, it is probably a good idea to shoot for the middle of the torque value range.

Community: I stand corrected on the 3003 chart reference, I should have done my homework.

Scott, I did the same later this morning too. Although it "seems" like 125 in-lb is a bit high for 3003 3/8 tubing it does not appear to cause an issue, if it is strictly adhered to in reassembly procedures. 75 in-lb is probably more forgiving.

Flats - Yes, my flats won't match Scotts, but I tested a few and concluded that firm was about 20 in-lb. So this is where all of these results begin. The starting point makes a big difference.

OK, here are the pictures. Forgive the photos, I don't own a good camera and had to use my iPhone with an inspection glass.

They are labeled.
50%2520in%2520lb%2520trq.JPG

75%2520in%2520lb%2520trq.JPG

125%2520in%2520lb%2520trq.JPG

125%2520in%2520lb%2520cutaway.jpg


Edit: 3-19-2015 - -With Scott's comment (below) I retested. Ensured 90 deg, and did multiple samples. All the photos above should be considered lubricated, not dry. This means more of the torque gets converted into clamping force. The 75 in-lb view was repeated 4 times and the impressions look like the photo above. One of the samples was generously lubed with BoeLube liquid and showed no difference, thus the change from dry to lubricated test condition. I compared two torque wrenches and there was minor difference. YMMV, but using the values in Vans chart in Section 5 will result in a good joint with 3003 tubing.
 
Last edited:
Good stuff Bill.

I didn't have the available time to shoot and post photos, but if I had it would be evident that there is a difference in the amount of deformation/reforming that occurred on the 3/8 tubes that I tested vs yours.
In fact, my 75 inch pound test sample looks very close to the same as your 50 lb test photo.


As mentioned before, I was careful to use a crows foot oriented 90 degrees to the handle on the torque wrench. If you did that also, then there must be some other variation causing the difference (torque wrench calibration, slight oil film on fitting threads, etc.







Community: I stand corrected on the 3003 chart reference, I should have done my homework.

Scott, I did the same later this morning too. Although it "seems" like 125 in-lb is a bit high for 3003 3/8 tubing it does not appear to cause an issue, if it is strictly adhered to in reassembly procedures. 75 in-lb is probably more forgiving.

Flats - Yes, my flats won't match Scotts, but I tested a few and concluded that firm was about 20 in-lb. So this is where all of these results begin. The starting point makes a big difference.

OK, here are the pictures. Forgive the photos, I don't own a good camera and had to use my iPhone with an inspection glass.

They are labeled.
50%2520in%2520lb%2520trq.JPG

75%2520in%2520lb%2520trq.JPG

125%2520in%2520lb%2520trq.JPG

125%2520in%2520lb%2520cutaway.jpg


7aomtqbaupFYsuUpi4MS7BxGHcIHIVK9UcZwMxGyfVo

YLM0lmPrXVPvtYqR3TMjrBxGHcIHIVK9UcZwMxGyfVo

67wejTuIJgYZHc5SrlwjPBxGHcIHIVK9UcZwMxGyfVo

ACgiyeiTQrFAbVoV2P6sRBxGHcIHIVK9UcZwMxGyfVo
 
RFT-37 degree

The ones we sell and are to be utilized for stainless steel anealed material:eek:
 
Wrap Up

I respect Scott's work, and had to go back to determine what I had done to yield different results. I edited my post above with the results so they would all be in the same place. Jumping to the final result - Apparently, my pictures correspond to lubricated joints. My technique and torque wrench accuracy were confirmed in range, but not precisely calibrated so that was not a source of difference. My "dry" pictures correspond to later tests with generous lubrication.

Van's posted range (Section 5) for 3/8" joints using 3003 tubing is 75-125 in-lb. This testing confirms that it is not a cause for concern, and using torque in this range will produce good joints, as shown above.
 
I really appreciate everyone's replies, I thought I was doing everything properly by researching post on VAF and reading manuals etc. I've had several tech inspections and several sets of eyes on parts of my project during the build, but I never had a second set of eyes on my flares.
I really encourage anyone else to have a EAA tech or other knowledgeable person check your flares.
 
Over tightening

Andy,
I initially followed the torque values Vans listed in the QA section of his website and screwed up my tank vent lines. http://www.vansaircraft.com/pdf/Torque_Spec_Aluminum_Fittings.pdf
I then got some scrap and played around with it, and could not satisfy myself with the incredible amount of torque that I was using to secure the B nut. I finally took everything over to an AP friend and he said that was way to much torque! After researching the whole page in the QA section it finally dawned on me that these torque values were for some performance lines built by Aeroquip. The company listed at the bottom. A search from the good folks at Google turned up the answer in the FAAs own AC 43-13.
2ur7div.jpg

After playing around with more scrap I finally have the "feel" for working with the flairing tool and correctly torquing these aluminum fittings. It looks to me from the pics that you had good flairs and may have just over tightend the nut. Good catch. Better on the ground than in the air with those cute kids of yours. :D
 
Back
Top