Mechanics Behind Fasteners 8

[Tsiolkovsky]

After reading a brief description on fasteners, their nature of loosening etc etc, its incredible the volume of science behind the simple bolt and nut. I have a few questions regarding this.

On the topic of bolt/nut loosening (of which is a problem on our machinery), I asked myself why cannot castellated nuts together with cotter pins be used wherever possible. After reading up, this configuration is to only be used in "low torque" applications. I ask, why low torque only?

Secondly, I do realize that a big problem is lining up the hole in the bolt so that the cotter pin slips in. But surely this can be corrected by incrementally adding/removing shim washers until a perfect fit is found. However I hear doing this compromises bolt strength as it creates a "weak point" due to numerous shims. TBH, im more concerned about bolt loosening than strength reasons. Also the bolt may stretch in its life or shims compress thus the torque tightness will drop off after time.

Ultimately, how true are the above points and the reasons behind it. I guess castellated nuts cant be a solution or else we would see them ubiquitously on truck hubs, all fasteners etc...

 

[hydtools]

I think you touched on the key reason when you said the cotter pin slips in. 'Slips in' implies clearance. Clearance which will allow the nut to loosen if the preload is not sufficient to prevent loosening. The cotter pin then only prevents losing the nut after it is loose.

Ted

 

[Metalguy]

Cotter pins should only be used for zero/very low prestress conditions, as should Loctite.

For high preload conditions where there is (and must be)a solid metal-to-metal contact with rigid members, the preload must be higher than any service tensile stresses.

Think of an engine connecting rod bolt. Severe alternating stresses on the rod itself, but almost nothing on a well-torqued bolt.

Alternating/varying stresses are required for starting fatigue cracks, and the number one cause of such cracking in fasteners is insufficient or loss of preload. Bolt head/nut imbedment can cause loss of preload, and such things as Loctite do not prevent it.

"You see, wire telegraph is like a very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? Radio operates the same way: You send signals here, they receive them there. The only difference is there is no cat." A. Einstein

 

[MiketheEngineer]

Everything said is true... But main reason so much study has been made on this subject - SPEED and ECONOMY. When GM installs about a billion bolts a month ---- try putting cotter pins on all those!! Or for that matter a NASCAR racing tire where 2/100s of a second can win or lose a race.

Most installations using cotter pins are typically what my dear old Dad used to call a "jezuz" bolt. If it let go - you said "Jesus" because he was the next person you were probably going to meet. Typically, the bolt(s) that hold the helicopter rotor in place or the prop on your airplane.

 

[berkshire]

Quote
""Most installations using cotter pins are typically what my dear old Dad used to call a "jezuz" bolt. If it let go - you said "Jesus" because he was the next person you were probably going to meet. Typically, the bolt(s) that hold the helicopter rotor in place or the prop on your airplane.""

Critical installations like this on aircraft are typically not cotter pinned, they are safety wired. The bolts are properly torqued, then a twisted wire is inserted in drilled holes in the head of the bolt and led to the next bolt or an anchor plate. The sole purpose of the wire is to prevent the bolt backing out if it does lose its clamping force.
B.E.

 

[MintJulep]

Safety wire doesn't do much either.

faq725-1549

 

[MiketheEngineer]

Berkshire - true but the idea is still the a same....

 

[drawoh]

I can see another issue with cotter pins.

If I install a castle nut and drilled bolt and I tighten them to the maximum allowed torque, I must back them off anywhere up to 120° so that I can insert the cotter pin. An alternate model of this would be that I would torque them down approximately where I wanted them, then adjust the nut ±60°

On shorter bolts, a 60° turn is enough to unclamp the bolt.

JHG

 

[MintJulep]

You might also be interested in thread404-280057

And the FAQ's in both this forum and forum404

 

[rerig]

drawoh,

Most castellated nuts have cotter pin openings every 60°.
You don’t torque to the Max.Torque.
You torque to the Min., then tighten to the next opening without going over Max. torque.

Rerig

 

[drawoh]

regig,

Oops. My bad. I should have said 60° and ±30°.

This is still a lot of tension. On a 1/4-20UNCx1 screw, a 60° turn works out to .008in extension, and a strain of .008. If the screw has an elastic modulus of 29,000,000lb/in², the resulting stress is 232,000lb/in².

Compression in the joint will reduce this a bit.

JHG

 

[Metalguy]

Pretty high strength screw there! Even an SAE Grade 8 bolt will go past yield ~130ksi, which is an elastic strain of ~4 mils/in.

"You see, wire telegraph is like a very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? Radio operates the same way: You send signals here, they receive them there. The only difference is there is no cat." A. Einstein

 

[drawoh]

Metalguy,

One hopes it will be a strength screw.

JHG

 

[gadkinsj]

The company I work for has invested thousands of dollars per machine in Nord-lock style washers and have seen the return on this investment. We have found it works better than anything else when it comes to keeping a properly torqued nut/bolt in place including chemical, spring washer, and wire tieing.

They are expensive (on the order of magnitude of half the cost of a bolt) and must be replaced every time you remove/reinstall the bolt but they are well worth it.

And make sure your maintenance people are using the torque wrench every time they do something. We find our biggest fasterner issues now are with under/over tightened bolts.

 

[Metalguy]

Unfortunately, torque can be a very poor indicator or the all-important preload on a fastener for rigid connections, and critical applications should not/do not use them. Very large preload errors are frequently experienced when using them.

Something as simple as the turn-of-the-nut method can be more accurate, and bolt stretch measured directly or via ultrasonics are far superior. Many engine builders have given up on torquing con rod bolts and use direct measurement of stretch.

Loctite, etc. and the various lockwashers/trick nuts have their place for non-rigid connections, but for true rigid connections they can lead to fatigue fractures if the preload is insufficient.

"You see, wire telegraph is like a very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? Radio operates the same way: You send signals here, they receive them there. The only difference is there is no cat." A. Einstein

 

[btrueblood]

Metalguy,

I can agree with most everything you say, but how does Loctite make a bolted connection more prone to fatigue?

 

[Metalguy]

Loctite etc. doesn't actually cause fatigue cracking, but if the preload is lost via bolt head/nut embedment, the bolt/nut doesn't back out/fall off. In a multibolt assembly where a loose piece isn't a problem by itself, the missing bolt can be observed, thus alerting people of a problem.

We had an embedment problem with some 1.5" main steam support bolting, and the heavy nuts were falling down ~20 feet. Without telling engineering (me), maintenance decided to fix it themselves.

Too hot for any Loctite, they WELDED the back side of the nuts to the bolt--ultra high-temp "Loctite". For a while nothing happened, then a nut fell down WITH a piece of the bolt inside! Happened twice before I found out--classic fatigue cracking right across the bolt.

Sure glad one of those things didn't hit anyone in their hard-hat!

Moral: It's better to know you're losing preload when you are, and things like Loctite can hide what's happening.

"You see, wire telegraph is like a very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? Radio operates the same way: You send signals here, they receive them there. The only difference is there is no cat." A. Einstein

 

[btrueblood]

Fair 'nuff, good story, glad I asked.

 

[ricklts]

One consideration might be to use one of the Bellville washer designs. These have been successful in many petrochem applications to avoid loosened nuts.

 

[Guest102023]

I just recently researched 'thread locking' for a recent failure analysis I did. Loctite (I use the term generically) is apparently an accepted form of locking in environments where vibration might loosen the bolt. HOWEVER, it is no substitute for correct preloading of the bolt.

Too tight, and more commonly too loose, are by far the most common causes of fastener failure.