Wing attach bolts in tension

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,506
Location
Saline Michigan
Going to have to disagree with you on this point. The wheel is held from shifting by the taper that is machined into the rim and on the surface of the lug nuts. Think of it like a taper pin. On an OEM application, the load is carried by the center of the rim against the hub. This was the problem with "Unilug" custom wheels a few years back. Designed to "fit many", these transferred the twisting (friction) and load carrying to the actual lug studs instead of the hubs. Dennis

Let's check that against field experience. Most, but not all wheels I have removed from cars over the last 48 years slip off the hub easily when lugnuts are removed. Some are firmly held by corrosion products at the wheel contact with the hub. Most recently, I pulled four alloy wheels from my Focus ST for replacement with other wheels and tire, and two have light witness marks in the pilot bore. The other two have corrosion products in the bore that interfered with the pilot on the hub. So maybe we have some support by pilot, but it does not look like much...

Let's review the loads on the wheels. There is the weight of the car downward trying to slip the wheel upward along the hub. There is braking or acceleration trying to slip the wheel fore and aft along the hub. There is torque applied either from power or braking that attempts to slip the wheel in rotation on the hub. And there is cornering which is compressing the joint between wheel and hub for part of the rotation and increasing tension on the joint during the other half of rotation. Let's see how the piloted hub works (or doesn't work) to restrain the wheel against movement on the hub.

I ran a simple calculation using specified wheel lug torque (100 ft-lbs), lug diameter (14 mm), and contact ring radius (2.625") for my Focus ST. I assumed coefficient of friction of 0.15 on the contact ring, which is low end, but gives us a minimum to work to. Using standard calcs (from Shigley), preload from tightening each lug nut is 10,700 pounds. Five lug nuts makes the total preload 53,500 pounds. Why do we suppose Ford put that much preload in?
  • Even a low estimate of coefficient of friction of 0.15 means the wheel will slip across the hub when 8000 pounds is applied across the hub. The biggest loads I can get from vertical force plus max braking is on the order of 1200 pounds - a factor of safety of 6.7! This system is NOT slipping radially, so the hub pilot does not really provide any support against these loads;
  • Checking wheel torque against slip of the wheel on the hub - tractive limits set torque at about 11300 in-lb while torque to slip the wheel on the hub is about 21000 in-lb, with a factor of safety of 1.86. Hmm, resisting torque has way less factor of safety than radial slip. Being as the pilot is about 1" radius, if there was 11300 lb of friction between pilot and bore, the pilot could resist these torques, but then you would need extraction devices to remove the wheel for service - not present on any car I know of. So, the pilot does not provide rotational support to prevent fretting and stud/wheel fatigue - it comes from friction that the stud system provides;
  • I also checked the cornering moment from sideways traction at the tire patch and found that the joint stays loaded up with large margins, so fatigue of the studs is unlikely;
In short, nothing we do while driving the car will make the wheel slip on a hub that has been preloaded with properly torqued lug nuts. Crash forces might exceed these limits, but not driving. This is all due not to the pilot but due to friction provided by the preload from properly tightened lug nuts.

Pilot fit of hub to wheel centers the wheel for tightening the lug nuts, but once that is done, the design of the lug system and contact between hub and wheel is sized to have excess torque and radial friction capacity to prevent movement of wheel on the hub. Experience with these systems and analytical approaches bring us to the same conclusion - the wheel is held in place by friction from large preload, while the pilot serves to get things decently centered before we tighten the lug nuts.

Now if the wheels were to be held in place by the pilot, but friction was inadequate to hold the wheel in place, pilot clearance seen in the field would allow the wheel to orbit or nutate on the hub. The lug system would see substantial variation in load as the vehicle is driven, and I would expect fatigue and failures of the stud and/or the wheel. This is exactly the failure mode seen when lug nuts are inadequately tightened - there is fretting (wear from small but real slipping back and forth of surfaces) at the contact between wheel and hub, studs break off and/or lug nuts get pulled through the wheels, then wheels depart the vehicle.

One historical point that occurred in the 1980's IIRC. A car had been repeatedly brought in to the dealer complaining of large vibration on one corner, with the dealer unable to solve the problem. Eventually a wheel and tire departed from that corner of the car, and a tragic accident occurred. All five studs had failed in fatigue. During the subsequent teardown of the live rear axle, the axle shaft was found to have red paint on it and the hub face was found to be crowned so that wheel contact from tightening the lug nuts occurred near the pilot instead of out near the periphery of the hub, as specified in the part drawing. Red paint denoted scrap parts in the factory that built the axle. Other evidence included much fretting of the wheel, brake drum, and axle hub, all near the ID of the wheel, indicating much movement and slipping about under load. The wheel was firmly held on near the pilot, but this provided little capacity against small but real slipping of the wheel on the hub, and cyclic loading of the studs. It also did not support the wheel in a uniform manner, negating any dynamic balance of the wheel, and the resulting vibration. In this case we see friction resisting rotational movement was low, the pilot did not provide support, and the studs failed.

The attempts at universal fit aftermarket wheels with no piloting except the studs and lugnuts can result in substantial and non-repeating imbalance. Yes, the pilot fit between wheel and hub is useful and probably needed. To conclude that the pilot holds the wheels in place only covers radial loads (the easiest part) but provides absolutely zero support from slipping rotationally under accel and braking loads. Indeed supporting torque is the defining issue and one only reasonably achieved with preload and friction...

Billski
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,506
Location
Saline Michigan
I don't have pictures, but I've installed many bolts in the Bonanza/Baron a/c. The wing attach "bathtubs" have machined rings in the faces of the two bathtubs at each attach point (2 each, main and rear spars). A soft aluminum washer goes between the fittings and the bolt through all pieces. The bolt has a double washer set, a large washer around a smaller diameter washer, under it's head. The outer washer has holes drilled in it. As you apply torque to the bolt head you apply rotation to the outer washer with a pin inserted into one of the holes. When the inner washer swells to the point it takes a certain force to move the outer washer, the bolt is torqued. The machined rings in the bathtub fittings bite into the soft washer between them for vertical position. The soft washers are only used once. The bolts are a loose fit in the bathtub fittings.
Those bolts are a loose fit to the holes and are not bearing load in shear. This sure sounds like bolts preload the fittings, and the fittings carry the shear.
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,506
Location
Saline Michigan
(New thread so as not to hijack the one about U bolts.)

The Cassutt spar sits on wood shims on the top longerons with a wood shim and an aluminum strap across the top. Four 5/16 (IIRC) bolts, two per side, tie the strap to bushings welded vertically into the top longeron. There is additional bracing below the longerons, around the bushings.

Really simple to build, really easy to shim the spar for wing alignment, really easy to remove the wing for trailering, really strong.

What other aircraft use bolts in tension to attach the wing?


BJC
Let's confirm some facts here... IIRC, the Cassutt has a one piece wing with the main and drag spars bolted to the fuselage. Is my recollection correct?
 

BJC

Well-Known Member
Supporting Member
Joined
Oct 7, 2013
Messages
15,942
Location
97FL, Florida, USA
One piece wing, main spar carries through the fuselage, attached as previously described. The drag spars are pinned on each side with a single bolt to the fuselage and a carry-through tube. That bolt, IIRC, is vertical.


BJC
 

viva_peru

Member
Joined
Apr 12, 2012
Messages
19
Location
Detroit, Michigan
Interesting discussion. Looking at the first photo posted of the Cassutt it appears to me that the bolts clamp down on what appears to be a wooden spar. In order to develop significant preloads on the bolts the spar would need to have some kind of metal insert which I suspect it does not. Given this, while in flight, I suspect that the loads on the bolts will vary with G. In the C-130 discussion it is interesting that the mentioned using longer bolt as they generally result in a more stable and consistent bolt preload.
 

ScaleBirdsScott

Well-Known Member
Joined
Feb 10, 2015
Messages
1,555
Location
Uncasville, CT
This discussion got me thinking since I'm trying to sort out a few things myself, and I find it interesting that indeed the P-51 fuselage sits on top of the wing and is held down by bolts in tension; bolts clamping the longerons to riser brackets bolted to the spars. And they're "only" 2x AN12 bolts (main spar attach) and 2x AN10 bolts (aft spar) plus what looks like a handful of AN5 bolts for the forwardmost spar under the firewall. Oh and what looks like some clevis pinned turnbuckles for the very leading edge to the engine mount? All told, not a whole lot going on. Very simple, for its size. the clevis bolts are a little curveball maybe.

Of course in another drawing it seems to contradict this, showing only a pair of forward spar AN5 bolts at the firewall. But then it goes to show the real secret sauce of a metric ton of fasteners between the lower longeron and the stiffener angles along the top of the wing. So that's doing a lot of work.
 
Last edited:

Pops

Well-Known Member
Supporting Member
Joined
Jan 1, 2013
Messages
11,521
Location
USA.
In the early 1980's a friend bought a T-6 in a shipping crate from the Spanish AF. We went to D.C. and put it together and flew it home. Don't remember the torque valve of all the bolts, just that there were a lot of bolts holding the wings on.
 

Hawk81A

Well-Known Member
Joined
Sep 3, 2021
Messages
342
Let's check that against field experience. Most, but not all wheels I have removed from cars over the last 48 years slip off the hub easily when lugnuts are removed. Some are firmly held by corrosion products at the wheel contact with the hub. Most recently, I pulled four alloy wheels from my Focus ST for replacement with other wheels and tire, and two have light witness marks in the pilot bore. The other two have corrosion products in the bore that interfered with the pilot on the hub. So maybe we have some support by pilot, but it does not look like much...

Let's review the loads on the wheels. There is the weight of the car downward trying to slip the wheel upward along the hub. There is braking or acceleration trying to slip the wheel fore and aft along the hub. There is torque applied either from power or braking that attempts to slip the wheel in rotation on the hub. And there is cornering which is compressing the joint between wheel and hub for part of the rotation and increasing tension on the joint during the other half of rotation. Let's see how the piloted hub works (or doesn't work) to restrain the wheel against movement on the hub.

I ran a simple calculation using specified wheel lug torque (100 ft-lbs), lug diameter (14 mm), and contact ring radius (2.625") for my Focus ST. I assumed coefficient of friction of 0.15 on the contact ring, which is low end, but gives us a minimum to work to. Using standard calcs (from Shigley), preload from tightening each lug nut is 10,700 pounds. Five lug nuts makes the total preload 53,500 pounds. Why do we suppose Ford put that much preload in?
  • Even a low estimate of coefficient of friction of 0.15 means the wheel will slip across the hub when 8000 pounds is applied across the hub. The biggest loads I can get from vertical force plus max braking is on the order of 1200 pounds - a factor of safety of 6.7! This system is NOT slipping radially, so the hub pilot does not really provide any support against these loads;
  • Checking wheel torque against slip of the wheel on the hub - tractive limits set torque at about 11300 in-lb while torque to slip the wheel on the hub is about 21000 in-lb, with a factor of safety of 1.86. Hmm, resisting torque has way less factor of safety than radial slip. Being as the pilot is about 1" radius, if there was 11300 lb of friction between pilot and bore, the pilot could resist these torques, but then you would need extraction devices to remove the wheel for service - not present on any car I know of. So, the pilot does not provide rotational support to prevent fretting and stud/wheel fatigue - it comes from friction that the stud system provides;
  • I also checked the cornering moment from sideways traction at the tire patch and found that the joint stays loaded up with large margins, so fatigue of the studs is unlikely;
In short, nothing we do while driving the car will make the wheel slip on a hub that has been preloaded with properly torqued lug nuts. Crash forces might exceed these limits, but not driving. This is all due not to the pilot but due to friction provided by the preload from properly tightened lug nuts.

Pilot fit of hub to wheel centers the wheel for tightening the lug nuts, but once that is done, the design of the lug system and contact between hub and wheel is sized to have excess torque and radial friction capacity to prevent movement of wheel on the hub. Experience with these systems and analytical approaches bring us to the same conclusion - the wheel is held in place by friction from large preload, while the pilot serves to get things decently centered before we tighten the lug nuts.

Now if the wheels were to be held in place by the pilot, but friction was inadequate to hold the wheel in place, pilot clearance seen in the field would allow the wheel to orbit or nutate on the hub. The lug system would see substantial variation in load as the vehicle is driven, and I would expect fatigue and failures of the stud and/or the wheel. This is exactly the failure mode seen when lug nuts are inadequately tightened - there is fretting (wear from small but real slipping back and forth of surfaces) at the contact between wheel and hub, studs break off and/or lug nuts get pulled through the wheels, then wheels depart the vehicle.

One historical point that occurred in the 1980's IIRC. A car had been repeatedly brought in to the dealer complaining of large vibration on one corner, with the dealer unable to solve the problem. Eventually a wheel and tire departed from that corner of the car, and a tragic accident occurred. All five studs had failed in fatigue. During the subsequent teardown of the live rear axle, the axle shaft was found to have red paint on it and the hub face was found to be crowned so that wheel contact from tightening the lug nuts occurred near the pilot instead of out near the periphery of the hub, as specified in the part drawing. Red paint denoted scrap parts in the factory that built the axle. Other evidence included much fretting of the wheel, brake drum, and axle hub, all near the ID of the wheel, indicating much movement and slipping about under load. The wheel was firmly held on near the pilot, but this provided little capacity against small but real slipping of the wheel on the hub, and cyclic loading of the studs. It also did not support the wheel in a uniform manner, negating any dynamic balance of the wheel, and the resulting vibration. In this case we see friction resisting rotational movement was low, the pilot did not provide support, and the studs failed.

The attempts at universal fit aftermarket wheels with no piloting except the studs and lugnuts can result in substantial and non-repeating imbalance. Yes, the pilot fit between wheel and hub is useful and probably needed. To conclude that the pilot holds the wheels in place only covers radial loads (the easiest part) but provides absolutely zero support from slipping rotationally under accel and braking loads. Indeed supporting torque is the defining issue and one only reasonably achieved with preload and friction...

Billski
While I respect your engineering background, I still disagree. Further disagreement would not only be counterproductive, it would lead to a thread hijack. Dennis
 

TFF

Well-Known Member
Joined
Apr 28, 2010
Messages
17,984
Location
Memphis, TN
There was a need to mount the wing like Cassutts do. To be able to throw the plane into a trailer and drive to races. Wittman V-Witt which was essentially a Bonzo used flying and landing wires and jammed the spar root into a box fitting. Solutions for a budget racing. Best solutions? Best built in a garage in the 50s solution, probably.
 

BBerson

Light Plane Philosopher
Supporting Member
Joined
Dec 16, 2007
Messages
16,100
Location
Port Townsend WA
I think Hawk 81 was talking about the taper on one side of each lug nut that fits the taper in wheel holes.. The holes in the wheel are sloppy fit. If you install the lug nuts upside down the wheel might not center and fail.
 

Puggo1

Active Member
Joined
May 19, 2020
Messages
31
hi,
Interesting thread.
Some aerobatic monoplane aircraft with demountable low wings bolt their wing onto the fuselage from underneath. The single wing spar is lifted into a inverted "U" shape within the steel fuselage and then a plate is bolted upwards and across the spar. This retains the wing from falling out of the "U". The bolts are only in tension when pulling negative G's.
I'm stunned to see the aircraft with wings totally reliant on bolts in tension (i.e. parallel to the wing spar). It would be an interesting calculation to determine the fatigue life of these bolts. Under normal flying conditions, the wing joint should not open up. This would require the tightening torque on the bolt to apply a load exceeding say 6G's on the wing for a safety factor of 1.5 over a utility loading of 4G's.
Secondly, the quality inspections required on the bolt material, thread cutting/size depth, etc. would need to be very thorough.
cheers
Puggo
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,506
Location
Saline Michigan
Let's go through a little fastener theory:

In joints where threaded fasteners attach members together, the whole scheme lays on the threaded fastener being a spring that clamps members together, and the clamped joint then becoming essentially one piece under operating loads.

If the preload is significantly higher than the loads trying to separate the members, and the joint does not open and close or shift, the bolt sees little cyclic load and does not fatigue.

If instead, we only lightly preload the fastener or have the joint undersized relative to the loads, the joint will open/close with load cycles and/or shift, and fastener or members will fatigue and fail. This is why use of torque wrenches is important.

I know that I offered to run numbers on the C-47 and AT-6. These wing joints are complicated assemblies and I still do not have enough detail to go forward on them, but I think we can rest assured that the bolts are torqued and thus under significant tension, and that the joints remain clamped and do not shift. We have plenty of field demonstration that these entire airframes are, barring corrosion and fabric covereds, pretty much infinite life assemblies. I shall dispense with further analysis of these designs.

Back to the Cassutt. This does seem to be a proven design. Many flying, many at much higher airspeeds than originally intended, few failures. It may be an entirely adequate set of compromises flying in tight formation.

But to answer the questions about its mountings, I will assess the relative stiffness of the members being compressed (the spar) and the springiness of the bolts and plates. This does get fussy as the plates are probably softer springs than the bolts... Someone please supply some dimensions for:
  • Width and depth of the spar, assumed to be a single piece of Sitka Spruce;
  • Material, thickness and distance between bolt holes on the plates;
  • Dimensions of the longeron cross sections, assumed to be 4130;
  • Bolt dimensions and assembly torque, assumed to be AN fasteners.
Billski
 

PMD

Well-Known Member
Joined
Apr 11, 2015
Messages
1,185
Location
Martensville SK
Let's check that against field experience. Most, but not all wheels I have removed from cars over the last 48 years slip off the hub easily when lugnuts are removed. Some are firmly held by corrosion products at the wheel contact with the hub. Most recently, I pulled four alloy wheels from my Focus ST for replacement with other wheels and tire, and two have light witness marks in the pilot bore. The other two have corrosion products in the bore that interfered with the pilot on the hub. So maybe we have some support by pilot, but it does not look like much...

Let's review the loads on the wheels. There is the weight of the car downward trying to slip the wheel upward along the hub. There is braking or acceleration trying to slip the wheel fore and aft along the hub. There is torque applied either from power or braking that attempts to slip the wheel in rotation on the hub. And there is cornering which is compressing the joint between wheel and hub for part of the rotation and increasing tension on the joint during the other half of rotation.


Billski
Of course, your assessment is correct, but you missed the one extremely large load that one must design for in wheel ends - hitting the bump stops. Wish I had some hard numbers to inject, but of course a function of the somewhat randomly chosen jounce velocity, vehicle mass and rubber hysterisis. One way or another, mush greater value than regular deflection within limits, cornering or rotational torques.

I have removed MANY alloy wheels from iron hubs (actually rotors) that I know for certain have been hard onto the travel limiting devices. IF the hub pilot or the stud in bending had actually carried the load, the corrosion product woutd have been fractured and the wheel(s) would just fall off with fastener removal. They don't/didn't. To demonstrate just HOW much I trust the clamping force of wheel bolts, I place a coating of zinc-rich paint ("cold galvanizing) on the face of the hub and rotor - drastically reducing the co-efficient of friction between wheel mounting face and hub/rotor face - and there is never any evidence that movement has taken place (but they DO come off without resorting to a BFH to fracture the bond at the interface).
 

JimCrawford

Active Member
Joined
Nov 30, 2019
Messages
30
Location
Oxford, UK
Back to the 'bolts in tension' debate. I thought the wise owls of 'Homebuilt Airplanes' would be amused to see this arrangement which is a Tipsy Nipper. Shots 01 & 02 are details taken from the drawings and show transverse and longitudinal sections of the main spar at the attachment bolts. The one piece wing is retained by two 5/16" bolts at the longerons. Pitching moment is reacted by a rear spar with much lighter fittings. The main spar is made up of spruce laminated booms separated by spruce blocks (grain vertical) and faced with birch ply. Sorting out the spring coefficient for the spar involves a lot of detail!

Photos 03 & 04 show the top and bottom terminations of the attachments. The bolt is actually a stud as a bolt was found to have cracks, probably from manufacture, and there was a mandatory directive to either have them crack tested or replaced by studs provided by the type design rights holder. Straight replacement with studs was the simplest solution as the bolts are very long (~7"") and not listed by any of the usual suppliers.

The top of the stud distributes the load across the full chord of the spar, it is shaped to account for the wing dihedral. The lower face of the spar rests on a flat welded onto the top longeron. and the stub passes through to the weldment you see below the longeron.

The first run photo is included to show the scale of this diminutive aerobatic aeroplane which in this case is powered by a 1834 VW. The guy by the cockpit is acting as safety man as he has immediate access to the throttle, switches and sight of the oil pressure. The guy behind the wing is actually minding a fire extinguisher which is out of shot. The engine came with magnetos but this is the first run with dual electronic ignition. After a bit of a faff getting the priming right it now starts on first compression, hand swung of course.

Best Wishes

Jim
GAVTC first run 03feb2022edit_Moment.jpg Nipper spar 04.jpg Nipper spar01.jpg Nipper spar02.jpg Nipper spar03.jpg
 
Top