What's new
Van's Air Force

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

Wing Attach Bolt Torque Values

Amazon-1

Active Member
I put my wings on the RV-14A project today. I searched the VAF forums and didn't find a satisfactory answer to my question (below).

The plans call out the larger bolts as NAS1309-58 close tolerance bolts. They use an AN365-918A elastic stop nut. I determined that this means that it is an AN9 bolt.
In the chart in section 5.20 it calls for 66.6 – 83.3 ft-lbs of torque. It does not specify why this is a range, it also does not specify if this includes the drag torque or not. I do not have an easy way to determine drag torque (I don’t own a dial torque wrench).

Similarly, the smaller four bolts are NAS1306-58 bolts. They use an AN365-624A elastic stop nut. I determined that this means they are AN6 bolts. The chart in section 5.20 calls for 13.3-15.8 ft-lbs of torque. Again, why is this a range and does it include drag torque?

I am interested in what other 14 builders have done and also anyone's opinion on the topic.

Having the wings properly attached is a high priority for me. I want to get his right.

Thanks!

Bruce
 
Most people just aim for the middle of the range. If you are sure your torque wrench is well calibrated you could shoot for the smaller number, since these bolts are 100% in shear. Make sure you follow instructions, e.g., most airframe bolts are torqued ‘clean and dry’. If you used a little grease to help get the bolts in, make sure you use a solvent to clean the threads, let dry. The specified numbers do not include drag from the nut locking feature. To measure this: Put the bolt head in a padded vice, bolt horizontal. Screw the nut on until the locking feature is fully engaged, a couple of threads showing thru the nut. Put a wrench on the nut, wrench at 2 o’clock. Tie a string to the end of the wrench, and tie some weights to the end of the string. Now, push down on the wrench. At 3 o’clock let go. If it stops, add more weights and repeat. When you have enough weight that it continues on its own past 3 o’clock (just a bit), you’re there. Add up the weights, add in half the weight of the wrench, in pounds. Measure the distance along the wrench between the string and the bolt, call it X inches. Drag torque is the total weight times X, in inch-pounds.
 
I answered a similar question a while back. This will be an unpopular post if taken the wrong way but hear the whole thing out. Vans does a great job with their designs and support. Like any OEM, their recommendations are appropriate only for their designs and applications. Their released approach shouldn't be taken as gospel and applied to other applications without proper consideration.

My first engineering job out of college was in the Space Shuttle Program. I learned those old North American Aviation/Rockwell International/NASA specs before I learned any FAA related specs. Their QA/QC programs were a bit more rigid and rightly so. The application the OP is asking about is in shear load. Shear and tension applications have different target values; typically ~50% and 90% of the elastic range of the fasteners, respectively. Besides the values, installation considerations are most always similar.

There's a reason there's a range for torque values and you apply them per installation considerations.

- The prevailing torgue (e.g. self locking nut drag) should be measured and added to any final torque values. Vans specifies a range that considers this; perfectly acceptable as they are the design center/OEM/etc.

- Torque to the mid value above the running torque when you torque from the nut side. (preferred method)

- Torque to the high side (again above the running torque) when you have to torque from the bolt side as friction in the stack will add to the perceived (not actual/applied) torque

- Torque to the minimum value when utilizing a castellated nut then align the next hole (only for a non-specifically engineered application). Sub "lite" washers if final torque exceeds the high limit of the range.

This is overly simplified this and there are exceptions. Any aircraft structural engineers that read this, please don't panic. Sticking with "good shop practice" here.

Now if the Vans values that I've witnessed are improperly applied (I have RV4 plans but that section seems "universal" in V world), the consequences could be catastrophic. A critical tension application (e.g. crankcase through bolt, cylinder head bolt) torqued to the lower shear values (not using specified thread/washer/nut base lube/new cad coatings would allow for a cyclic load and pending fatigue failure of the fasteners.

In summary, stick with the released engineering. The acceptable hierarchy is:

Drawings
Specs (from the OEM)
Aerospace standards
Good shop practice

The aforementioned three application points (above) will keep you closer to the designers intent. When in any doubt, check with the OEM. Build safe.
 
- The prevailing torgue (e.g. self locking nut drag) should be measured and added to any final torque values. Vans specifies a range that considers this; perfectly acceptable as they are the design center/OEM/etc.

My apologies, but I am a bit unclear on this statement. Do you mean here that Van's numbers should be used unmodified, as they already consider/include the prevailing torque?

Also, you use the terms "prevailing torque" and "running torque." Are these identical in meaning?

Sorry, chemical engineer here, not a structural guy.
 
Geez Louise.

Considering this nut and bolt combination is used on every RV-14 (and possibly on other models) and it is pretty darn important, I would think it would be in Van's interest to publish an "all in" torque value (including prevailing torque).

Can the drag torque of an individual nut of a given size and type vary all that much from nut to nut?

I get that it makes sense to use published figures from the hundreds of other bolts on the airplane, but these nine bolts are more critical than most.
 
My apologies, but I am a bit unclear on this statement. Do you mean here that Van's numbers should be used unmodified, as they already consider/include the prevailing torque?

Also, you use the terms "prevailing torque" and "running torque." Are these identical in meaning?

Sorry, chemical engineer here, not a structural guy.

Yes/no. Example = Section 5.20 of Vans manual; Where a value for an application (AN4 with an elastic stop nut) is listed by the OEM, that bit of engineering has been considered. No need to measure the drag of the nut locking feature.

"Prevailing" is more accurate but really splitting hairs here. Didn't want to upset any structural guys reading this. If you where torqueing a fastener in a blind hole and/or a close tolerance fit, prevailing is the all-encompassing term for of the summation of the perceived (not actual) torque. This term is never really wrong but once again, splitting hairs.

Your avatar is of the Space Shuttle Main engines. That's my old system. Lots of critical torques there that the bolt strain (ultimate goal of fastener torque) is ultrasonically measured. A bit of an art for a 60+ bolt joint in joints that aren't that rigid for thrust/weight considerations. Adjustment of one bolt affects many other locations. Could take many work shifts to get everything in spec. I'll leave it to you as a ChemE to explain supersonic combustion chemistry to VAF world and help everyone conceptualize a launch vehicle getting lighter at ten tons/second.
 
I have measured many of the nylon locking nuts (for grins) and they span quite a range of torque values. You are going to have to quote chapter and verse where Vans includes the nut torque in the specification.

No application that needs the tension of the fastener as its designed function will assemble dry. Final stretch is just too variable, 20-30%. That is why you will always see "lubricated" as part of that torque spec.
 
Yes/no. Example = Section 5.20 of Vans manual; Where a value for an application (AN4 with an elastic stop nut) is listed by the OEM, that bit of engineering has been considered. No need to measure the drag of the nut locking feature.

Section 5.20 specifically says prevailing torque must be added:

When tightening fasteners with self-locking nuts the chart values must be modified. Due to the friction of the locking device noticeable torque is required just to turn the nut onto the threads and does nothing to actually tighten the parts together and stretch the bolt (clamp load). This is called friction drag (or prevailing) torque. The friction drag torque must be determined and then added to the standard torque from the table. Run the nut down to where it nearly contacts the washer or bearing surface and check the friction drag torque required to turn the nut. (At least one thread should protrude from the nut). Add the friction drag torque to the standard torque. This sum is referred to as the final (or total) torque, which should register on the indicator or setting for a snap-over type torque wrench.

As an example illustrating the importance of determining the friction drag torque consider a new AN3 bolt and MS21042-3 all-metal lock nut. Our tests showed an average friction drag torque of 14 in-lbs (your results may vary). The standard torque for this nut/bolt combination from the table below is 28 in-lbs. This results in a final torque setting on our wrench of 14 plus 28 or 42 in-lbs. Though we exceeded the 28 in-lb value listed in the table by using a final torque of 42 in-lbs we are still well within the capability of the nut. (Incidentally this nut must meet a maximum torque test value of 60 in-lbs per the military standard spec sheet.) Now what if we completely ignore the friction drag torque and set our wrench to just 28 in-lbs? Recall that it requires about 14 in-lbs (friction drag torque) just to turn the nut. We subtract 14 from 28 and arrive at only 14 in-lbs of torque(torque being the measurement of friction, not tension) applied to induce preload (clamp load) in the bolt. Not a satisfactory result.

The torque values given for "Standard Nuts" are straight out of 43.13, which also says:

d. Add the friction drag torque to the desired torque. This is referred to as “final torque,” which should register on the indicator or setting for a snap-over type torque wrench.

(I don't know where the MS21042 torques Van's uses as listed in section 5.20 come from)
 
My first engineering job out of college was in the Space Shuttle Program. I learned those old North American Aviation/Rockwell International/NASA specs before I learned any FAA related specs. Their QA/QC programs were a bit more rigid and rightly so. The application the OP is asking about is in shear load. Shear and tension applications have different target values; typically ~50% and 90% of the elastic range of the fasteners, respectively.

This is generally incorrect on our aircraft, and not to be used as guidance. The range of permitted torque on the fastener is usually limited by the materials it joins, not the bolt itself.

Let's take an extreme example: a 300 ksi bolt made of unobtainium joining two sheets of 2 ksi plastic. Torqueing the bolt to 150 ksi will just crack the plastic. This is not an academic problem: e.g. there have been SBs for cracks in the horizontal stab spar on the 12 in recent memory--I'll bet my $2 these weren't caused by faulty design of the joint, but rather by habitual overtorqueing of the fasteners. Overtorqueing the main spar bolts will similarly reduce the number of cycles to failure.

There is no hard *lower* limit on the torque of a fastener in shear. As long as all the materials are in proper contact, the only practical consideration is that neither the nut nor any of the materials joined start moving under vibration. Using the "upper end" of the range specified in the manual for a given fastener size has the advantage that it will absorb the plastic insert friction and set the bolt above the practical lower limit below which things start to move. This is good, time-tested practical guidance. Use it!
 
This is generally incorrect on our aircraft, and not to be used as guidance. The range of permitted torque on the fastener is usually limited by the materials it joins, not the bolt itself.

Let's take an extreme example: a 300 ksi bolt made of unobtainium joining two sheets of 2 ksi plastic. Torqueing the bolt to 150 ksi will just crack the plastic. This is not an academic problem: e.g. there have been SBs for cracks in the horizontal stab spar on the 12 in recent memory--I'll bet my $2 these weren't caused by faulty design of the joint, but rather by habitual overtorqueing of the fasteners. Overtorqueing the main spar bolts will similarly reduce the number of cycles to failure.

There is no hard *lower* limit on the torque of a fastener in shear. As long as all the materials are in proper contact, the only practical consideration is that neither the nut nor any of the materials joined start moving under vibration. Using the "upper end" of the range specified in the manual for a given fastener size has the advantage that it will absorb the plastic insert friction and set the bolt above the practical lower limit below which things start to move. This is good, time-tested practical guidance. Use it!

I'm not debating you, Sir but parts of your argument don't make sense in some regards. I believe there's more agreement than initially considered.
The 50%/90% elastic range torque is standard. If the underlying materials require something different as per your steel through plastic example, that should be considered in the engineering and the correct fastener, torque, whatever, applied.

While your "no minimum torque for shear" is perfect in theory for a joint with perfect geometries, from a practical case it is flat wrong. There is a lower limit for shear torque. Without such, the fastener will move in the stack creating wear and possible fatigue failure. This was established in the 50s and 60s when this and many specifications and codes were written in blood. The North American Aviation specs were the majority contributor to the FAA Advisory Circulars and Mil-Specs (Citation needed. This is at least tribal knowledge to the ancient NASA world). The minority of the applications in Vans (or others) designs are not critical in shear; hence, they're a standard drilled hole not reamed to close tolerances where the joint geometry provides the quality. A good case for this reasoning is the fact that torque values don't change for critical versus non-critically loaded joints; example = engine mount/fuselage holes, wing attach joints. In RV-4 world with the "single piece" spar, these critical tolerances are so tight they uniformly remove the cad plating from the bolts when installed; hats off to the fabricators. None of this discussion includes "supercritical' shear apps where interference fits are required (rivets, hi-loks, shrink fits, etc.)

For a tension application, not torqueing to 90% of the material elastic limit is inviting cyclic fatigue and failure.

The main point is to use the released engineering properly and only for it's intended use. Vans doesn't differentiate between shear and tension applications from anything I've ever seen (my docs and build). Applying inappropriately will get someone hurt. As mentioned, build safe.
 
Last edited:
lock nut friction

How much friction would the lock nut have? I can't imagine on a huge bolt like a -9 it would make much difference when the range is 66.6 – 83.3 ft-lbs of torque.
 
Bruce you said you do not own a dial torque wrench so I assume you do have the other kind where you can set the force with a twist. You can figure out the drag torque by starting with a very low setting if you can go that low, and keep increasing torque until the bolt moves instead of getting the "click".
 
Bruce you said you do not own a dial torque wrench so I assume you do have the other kind where you can set the force with a twist. You can figure out the drag torque by starting with a very low setting if you can go that low, and keep increasing torque until the bolt moves instead of getting the "click".

Be careful. You want the sliding friction, not the static (which is higher).
 
...
Lots of critical torques there that the bolt strain (ultimate goal of fastener torque) is ultrasonically measured. A bit of an art for a 60+ bolt joint in joints that aren't that rigid for thrust/weight considerations. Adjustment of one bolt affects many other locations. Could take many work shifts to get everything in spec.
...

I used to work in a power plant. During outages, when we had the entire turbine/generator train torn down for overhaul, the critical torques were set during reassembly by applying a measured torque to the case nuts and threaded tie rods, which had been heated to a particular temperature to allow a certain elongation. The lengths were measured ultrasonically, then the torque was applied. The rods were allowed to cool, and once everything was at equilibrium, the rod lengths were again measured ultrasonically to confirm that the proper strain had been applied. Since the HP turbine shells were operating at 1005F on the inlet and much lower on the outlet, the torques had to be set differently around the turbine shell to allow for differential expansion. Of course, the entire power train, which was around 150 feet long, if I recall correctly, had to be done properly to allow proper alignment. The boiler feed pumps had even closer tolerances, if I recall correctly.

Being a ChE, I was relegated to the part of the plant that didn't make any money--environmental controls. It was always an interesting dynamic between the many MEs, EEs, and the 2 or 3 ChEs they grudgingly kept around.
 
I used to work in a power plant. During outages, when we had the entire turbine/generator train torn down for overhaul, the critical torques were set during reassembly by applying a measured torque to the case nuts and threaded tie rods, which had been heated to a particular temperature to allow a certain elongation. The lengths were measured ultrasonically, then the torque was applied. The rods were allowed to cool, and once everything was at equilibrium, the rod lengths were again measured ultrasonically to confirm that the proper strain had been applied. Since the HP turbine shells were operating at 1005F on the inlet and much lower on the outlet, the torques had to be set differently around the turbine shell to allow for differential expansion. Of course, the entire power train, which was around 150 feet long, if I recall correctly, had to be done properly to allow proper alignment. The boiler feed pumps had even closer tolerances, if I recall correctly.

Being a ChE, I was relegated to the part of the plant that didn't make any money--environmental controls. It was always an interesting dynamic between the many MEs, EEs, and the 2 or 3 ChEs they grudgingly kept around.

That’s all very familiar to me. I went into the power generation business when I left aerospace. The main advantage (regarding this topic at least) is nothing has to be made light enough to leave the ground. The steam chest and gas turbine cases are rigid enough where one fastener’s strain has less of an effect on another. That said, there is quite a bit art involved with the science to get those cases back together safely during an outage. From the length of the generation train you mentioned, that was a fairly large coal plant or almost small nuke size. Anyway, guessing around 800MW or so. Imagine a case rupture that let lose a million horsepower of energy. The importance of proper joint fastener strain can’t be overstated. A more modern application would probably involve hydraulically stretching the studs to get them where they need to be. Heat is still used. If you want to learn a new, vulgar vocabulary, listen to the millrights as they break those bolts loose.

Thanks for looking after the environment while you were there.
 
Last edited:
This is all fun, but the OP asked a question. The standard handbook has the following table. The NAS bolts are specified as 160ksi strength (different location) and the table shows a higher torque for 160 ksi. I assume that use of 88 ft-lb would be within range for the last two columns.

Why would the 14 not use this torque?

Screen Shot 2020-09-28 at 5.32.56 PM.png
 
This is all fun, but the OP asked a question. The standard handbook has the following table. The NAS bolts are specified as 160ksi strength (different location) and the table shows a higher torque for 160 ksi. I assume that use of 88 ft-lb would be within range for the last two columns.

Why would the 14 not use this torque?

View attachment 2913

Because the spar is not made of 160 ksi steel.

The permissible range for 2024 is shown on page 05-20 in the plans: 66.6-83.3 ft*lbs for the AN9s, and 13.3-15.8 for the AN6s.

So, if the OP sets his (trusted) torque wrench to "83" and "15", after accounting for the plastic insert friction the bolts will end up clamping comfortably below what the aluminum will tolerate, and comfortably above the loosening threshold.
 
Last edited:
Because the spar is not made of 160 ksi steel.

The permissible range for 2024 is shown on page 05-20 in the plans: 66.6-83.3 ft*lbs for the AN9s, and 13.3-15.8 for the AN6s.

So, if the OP sets his (trusted) torque wrench to "83" and "15", after accounting for the plastic insert friction the bolts will end up clamping comfortably below what the aluminum will tolerate, and comfortably above the loosening threshold.

This is an interesting topic, and maybe it’s my overly ‘tho must follow the procedure to the letter’ nuclear/submarine background, but I don’t understand why there is a range to the torque value. Is there an answer on this? The range appears to be slightly larger than +/-10%. So if I factor in torque wrench error of 4% (max calibrated error) at 66.6, we are now at 63.9 as a worst case (BTW, how many torque wrenches out there can be used to a decimal place, I guess maybe some digitals!?). Why isn’t this torque simply specified as 75 ft-lb in this case? And then it’s a simple matter of adding the nut friction.
 
Because the spar is not made of 160 ksi steel.

The permissible range for 2024 is shown on page 05-20 in the plans: 66.6-83.3 ft*lbs for the AN9s, and 13.3-15.8 for the AN6s.

So, if the OP sets his (trusted) torque wrench to "83" and "15", after accounting for the plastic insert friction the bolts will end up clamping comfortably below what the aluminum will tolerate, and comfortably above the loosening threshold.

That is a baseless argument. Look at the chart. It uses the strength of the fastener not the base material being compressed.

Try again.
 
Effect of thread lube on final torque value?

Just curious how lube on the thread will affect the final torque value percentage wise? Can't find anything online. Is it 10%, 30% or more?
 
Just curious how lube on the thread will affect the final torque value percentage wise? Can't find anything online. Is it 10%, 30% or more?

About 33% increase in clamping force for bare steel (e.g. see here), but the cadmium coating on AN bolts is a lubricant, so less (e.g. here). One of the main functions of the ubiquitous washer under the nut is to provide consistent friction loss when torqueing.
 
I went way back and looked up my copy of MSFC-STD-486B, dated 11/92. This was one of the documents we used for analyzing payloads for the space shuttle. It gives torque values for different fasteners. It does NOT include the effects of the structure that the fastener is in. That's another document that I mention below. So this is the torque that the analyst starts with in their evaluation of a bolted joint.

I needed to know the strength of the NAS1309 bolt material so I looked that up. It's 160 ksi, which is stronger than an AN9 bolt (which is 1225 ksi). For a bolt of this strength,

The dry tension torque is 117 foot pounds.

The lubed tension torque is 100 foot pounds.

The tolerance is +0% and -15%.

For shear applications, which this is, no more than 60% of these torques is acceptable. This gives 70.2 foot pounds dry and 60 foot pounds lubed maximum for shear.

There is a requirement that a countersunk washer must be used under the bolt head (due to the radius there) and a regular washer under the nut.

In addition to these torques, the free-running locking torque must be measured and added.

This is NASA, not Van's and not the FAA. For the various factors that adjust and possibly limit these torques, we can look at NSTS 08307. My copy is from dated 10/89. Some of the factors that go into the final determination include

Joint separation or gapping,
Adequate fatigue and fracture life,
Yielding and failure,
Temperature changes,
Bending of the bolt.

The fastener preload affects joint sliding, or the ability of the joint to carry some load through friction. Considering this requires that the minimum fastener preload be considered. So does separation. Other considerations need the maximum preload, so those, too, need to be found.

The joint characteristics affect separation and some of the other factors.

Okay, that's a brief summary of fastener analysis. Don't forget that appropriate factors of safety go into each one of these items.

It gets complicated and the ratio of dry vs. lubed torque varies with the size and strength of the fastener - it's not merely a simple ratio.

Dave
Old retired aerospace stress analyst.
 
Last edited:
Thanks Dave, that's a great summary and sounds very much in line with what Van says, if a bit stricter given the higher stakes.

It does bring to mind the obvious follow-up question: what materials was the shuttle structure built of?

Also, in your work were the cumulative cyclic loads greater or less than for our little aircraft? I'd imagine the shaking is far more violent, but only designed for an hour total of use as opposed to 10000 hours on aircraft?
 
Adjusted Torque settings

Okay, thanks. So, in my case I have an RV-7 which calls for NAS 1304 and NAS 1307 close tolerance, cadmium plated bolts and Nyloc nuts. So taking into account all the previous comments and condensing all the rocket science into laymen's terms, I have done the following calculation to reach the desired torque settings on my wrench. See table. Please let me know if my reasoning is flawed.
 

Attachments

  • Screenshot 2020-09-29 at 14.15.05.png
    Screenshot 2020-09-29 at 14.15.05.png
    12.2 KB · Views: 198
I did very little work on the shuttle itself, just one of the struts that held the external tank on. But I saw the last one being built and it was mostly aluminum, as I recollect.

The payloads were of all sort of materials, but mostly aluminum. 6061-T6 was common but we did use other alloys as needed. For us, weight often wasn't as critical as stiffness. If you'll recall, the ratio of Young's modulus to density is nearly the same for aluminum, magnesium, titanium and steel. Tungsten, copper and beryllium were different. So might as well make it of aluminum if it can be made to work, and 6061 has a high resistance to crack propagation.

And your expectation of the vibration environment is generally correct. But the duration was only for several minutes during the powered launch phase. Usually thermal stresses were low partly because we had to minimize thermally-induced deflections.

We were often after pointing accuracy, the ability of something like a telescope to remain pointing in a known direction not only after launch but during the on-orbit thermal environment.

Dave
 
From the length of the generation train you mentioned, that was a fairly large coal plant or almost small nuke size. Anyway, guessing around 800MW or so.

850 net MWe, around 930 gross after we uprated the HP turbine to a newer design with more stages.
 
Okay, thanks. So, in my case I have an RV-7 which calls for NAS 1304 and NAS 1307 close tolerance, cadmium plated bolts and Nyloc nuts. So taking into account all the previous comments and condensing all the rocket science into laymen's terms, I have done the following calculation to reach the desired torque settings on my wrench. See table. Please let me know if my reasoning is flawed.

I think if you look for fine print, the top line is for dry, cad-plated bolts. No need to correct for Cad in the second line.
 
Okay, thanks. So, in my case I have an RV-7 which calls for NAS 1304 and NAS 1307 close tolerance, cadmium plated bolts and Nyloc nuts. So taking into account all the previous comments and condensing all the rocket science into laymen's terms, I have done the following calculation to reach the desired torque settings on my wrench. See table. Please let me know if my reasoning is flawed.

Van's already did the hard work for you. Go lighter on the smaller ones.
 

Attachments

  • Capture.PNG
    Capture.PNG
    220.6 KB · Views: 933
I went way back and looked up my copy of MSFC-STD-486B, dated 11/92. This'd was one of the documents we used for ..... .

Spec MAO101-301 governed torquing on the orbiter and main engine proper, IIRC. Old and not too fond memories. Competing government design centers (JSC and Marshall) and contractors. Hard to be productive.

Anyway. An interesting thread from a social standpoint. If you talk/post long enough, you can convince yourself and maybe others that you were correct. As mentioned, the hierarchy is OEM drawings, their specs, industry standards, etc. If the drawing doesn’t specify otherwise, stick with the released engineering, the aforementioned section 5.2 that managed to get circled back to. No one else knows the design considerations utilized. Build safe. Follow the OEM
 
Back
Top