Dave, thanks for letting us get inside the head of a failure analysis specialist. Much like detective work, it's fun to get in on the thought process as you follow the trail to the end resolution. And with this example, it shows you really do have to sweat the small stuff.
When one is tightening bolts on a car, there are torque specs in most cases. Thus one can use a torque wrench to ensure that you're meeting the spec and not putting undue stress on the bolt-plus-nut assembly (and also not cracking the metal parts that bolt and nut are clamping together). Of course we all know that in many cases, in repairs and particularly with home mechanics, bolts are just tightened and the "spec" is just done by eyeballing it (i.e., no torque wrench used). So my question is, is there any analogy for screws? In other words, how to you ensure a screw is tightened properly but not overtightened?
@Alexander: Yes, the situation is absolutely analagous. There are torque specifications for screws, just like there are for bolts and nuts. In either case, you are applying a torque in order to create a specific clamp load. This is the force which keeps the joint from coming apart.
Whether the fastener is a screw or a bolt, a big part of the torque which you apply goes into overcoming underhead friction. If there is a washer in the assembly, this is friction between the head of the fastener and the washer. If there is no washer in the assembly, it is the friction between the head of the fastener and the mating surface.
Another big part of the torque goes into overcoming thread friction. For a bolt, this is the friction between the threads of the bolt and the nut; for a screw, it is the friction between the screw threads and the mating part.
The smallest, but most important, part of the torque goes into stretching the screw or bolt. This part is responsible for the clamp load. The more the fastener is stretched, the greater the clamp load. Of course, if the fastener is stretched too much, it may yield or break. (As I discussed in my article, the mating threads also need to withstand the clamp load -- and if it is spread over too few threads, they may yield or break).
This brief discussion of torque-tension relationships shows why proper thread lubrication is important. By minimizing thread friction, proper lubrication allows you to produce more clamp load with the same amount of torque. It also shows why it is often advantageous to use a washer. Using a washer helps to make the underhead friction predictable and consistent.
Clearly, the torque-tension relationship for a given joint is highly dependent on the materials and finishes used. For example, if you replace a zinc-plated steel screw with a stainless steel screw, you shouldn't expect to be able to achieve the same clamp load with the same torque, because the friction coefficients will be different.
In order to come up with a torque specification for a specific joint, it's often a good idea to do a torque-tension study. Generic torque charts can be misleading. It's important to understand how much variation can be expected in terms of underhead friction and thread friction in your application. It's also important to understand how much variation in torque you can expect from your torque wrench or driver. When you have a good understanding of all of the variables involved in the torque-tension relationship, you can specify an installation torque which will provide the clamp load your application needs.
Once at a previous job, we had issues with threaded holes stirpping out when molds were hung on the aluminum platens of a molding machine. We could not understand this as we had a regulation on the maximum torque allowed on the screws. Observation of the process showed that we had one tech on one shift that would tighten until the torque wrench clicked then would tighten and additional half turn. When asked, he said that he just wanted to make sure that it was tight enough. We explained the reasoning behind the maximum torque and we really did not see the issue again. Someitmes you can have too much of a good thing.
@Tim: "Too much of a good thing" is a common problem. When I worked in a foundry, there was a melt operator who consistently poured steel as much as 100°F above the recommended temperature. When I asked him why, he said that it was because other operators who poured at temperatures below the recommended temperature had too many defects. This was true, but he was getting plenty of defects, too. (In fact, the finishing department had asked me to talk to him, because they were tired of welding up the defective castings he poured). I asked why he didn't pour at the recommended temperature. This thought didn't seem to have occurred to him. I guess nobody ever read him the story of Goldilocks when he was a kid.
Ford has gone to Torque-To-Yield (TTY) fasteners for critical parts. a Typical TTY might specify 40 lb-ft, then loosen 1/2 turn, torque to 80 lb-ft, then tighten 90 degrees.
On the plus side, TTY cylider head bolts can be installed and forgotten where older style bolts may need to be retorqued after so many hours of operation. On the minus side, TTY bolts should never be reused since the first use has stretched them.
Is there a specific manyufacturing process TTY bolts must first go through, or is just figuring out the proper torque procedure?
@TJ McDermott: Many liquid threadlockers serve as lubricants during installation, so they allow you to get more clamp load for the same amount of torque by reducing thread friction. Dry threadlockers tend to increase thread friction, so you need to supply more torque to get the same clamp load. In either case, though, the retained clamp load stays high, i.e. the joint is prevented from loosening over time. Obviously, the point of using a threadlocker is the retained clamp load.
@fredsay: Generally, fasteners which are used in torque-to-yield applications are the highest strength grades (Grade 8 for English, Grade 10.9 for metric). This makes sense, since to get the most bang for your buck, you want a fastener with a high yield strength. But in terms of metallurgy and manufacturing processes, there is nothing different between TTY fasteners and other high-strength fasteners. In fact, while we regularly use the term "TTY fastener," torque-to-yield is really just a fastening strategy, not a type of fastener.
Some fasteners which are specifically designed for torque-to-yield applications have a reduced shank, which ensures that they yield in the shank rather than in the threads. In other cases, a "TTY fastener" is just a standard Grade 8 bolt.
@Brother John: Yes, torque-to-angle is also a good tightening strategy. As you point out, tightening to a specified angle takes the effect of friction out of the equation. We often install fasteners under angle control with torque monitoring (i.e. we tighten to a specified angle, but also monitor the torque to make sure it is within a specified range).
I don't like the idea of threading a high strength fastener directly into aluminum in the first place...especially fasteners that will be removed and reinstalled a few times over the life of the product, and definitely ones that will see big and/or live loads. I've found it better to use helicoils or something similar in aluminum.
The good insert manufacturers can provide strength data for design.
When using helicoils (or similar non-name brand wire inserts) you should not use thread locking compounds on the fasteners. Best case, they can adhere the screw to the wire insert so it backs out when the screw is removed. Worst case, it reduces the strength of the completed joint because the locker prevents the load from distributing properly over the length of the insert. Better to use locking inserts, which have one convolution deformed.
@ed_bltn: I respectfully disagree. There is nothing fundamentally wrong with using high-strength fasteners in aluminum, provided that the joint is properly designed. There is a long history of thousands and thousands of successful designs which use high-strength fasteners in aluminum. Heli-coil inserts are a wonderful thing, but there is no reason to include them in every single design, unless you don't care about cost. As you mention, they are a good fix in certain situations, such as when a very high clamp load is needed, or when the threads need to stand up to lots of repeated assembly and disassembly.
Another approach to developing clamp load is using angle control. After taking up the slack in the joint, the bolt is tightened to a predetermined angle that corresponds to a stretch that develops the appropriate clamp load There will be a direct correlation between bolt stretch and clamp load until the bolt yields. This is not the case with the torque-tension relationship.
When I worked on cars, for 30 years on my Hot Rod cars (all daily drivers), I always thoroughly cleaned all fasteners and all threads before re-assembly.All fastener threads were properly lubricated, "Loctited", anti-seize used (exhaust fasteners), or silicon sealant used (head bolts (screws) threading into the engine block).Every fastener (screw) that went into a tapped hole was sequentially tightened using a torque wrench.The nuts were always tightened on the bolts using a torque wrench for all critical applications such as suspension and brakes.Such thoroughness meant I rarely had a problem on the road, and had much better success than other home mechanics I knew.
@wb8nbs: I would use that rule of thumb with a lot of caution. For one thing, it doesn't take into account the strength of the material you're threading into. You need more thread engagement if you are threading into a soft material than you would need if you were threading into a hard material. It also doesn't take the clamp load into account. If you need to withstand a higher clamp load, you need to engage more threads.
Based on the approach I described in the column, I calculated that you need a minimum of 7 threads of engagement when threading a Grade 5 3/8"-16 screw with a 6600 pound proof load into an aluminum casting with a 36,500 psi tensile strength. In this case, three threads would definitely not be enough.
The 1.5D rule of thumb which I mentioned would dictate 9 threads of engagement for a 3/8"-16 screw, regardless of the clamp load or the internal thread material. This is likely to be conservative in nearly every situation, but it's still just a rule of thumb and your mileage may vary. The best bet is to go through the calculation.
Many companies have internal standards used for thread engagement required for end product use. These are basicly charts with diameters and loads that dictate the required engagement for a tight / lasting joint.
My brothers ran into the debate about grease applied to automotive wheel studs, or not.One recommendation was that you should never apply grease because this would cause bolt stretch and lead to failure.But while growing up, in our driveway it was common practice to put a dab of grease on the threads.So my engineer brain broke it down this way.A little grease applied to just the threads is good to prevent the thread from freezing up and causing cursing trying to change a flat on the side of a highway.But keep the grease off of the angled surface under the head of the bolt/nut.Once the head comes into contact with the wheel, that diameter and angle area is a lot larger than the thread flank cumulative area, and it is the dry head contact area that will take the applied torque.Avoid grease under the head to prevent too much of the torque transmitting into bolt tension/stretch.
There's another funny story about abusing a pry bar trying to get an old Alfa Romeo wheel off – luckily I figured out the LH-thread trick before we broke it.When my Dad got home he explained why LH....
I believe he is referring to left hand threads used on one side of the car and right hand threads used on the other side. While this is probably a good thing for axle nuts where the spinning torque can loosen one side if both sides use a standard RH thread, it's probably not necessary on wheel studs because of their arrangement off of the center of rotation.
If you don't know that a nut or bolt is LH, you might add a lever to your wrench and snap the bolt before it unthreads. RTFM, a good shop manual can be one of your best friends. The harmonic balancer (actually it's on the lower pully for my supercharger which is bolted to the actual harmonic balancer) on my Ford is a LH thread, but it is clearly marked on the pully. However I have run into LH bolts which are not marked. I guess that's the real trick.
I stripped out several studs on the exhaust manifold while torqueing it down. Made the mistake of using anti-seize. So happen at work, I was investigating clamping force and thread lubricant. At work, we also broke a stud because it was coated with moly sulfide lubricant. Turns out, lubricant have a significant effect on clamping force. Friction make up 80% of the clamping force on a screw. Did a webpage on it:
Using LH studs on the LH side of the vehicle was due to the idea of self-tightening of the threaded fasteners from rotation of the clamped (but almost unclamped) components.Raced pre-war and post war MG's had a little habit of loosening the left rear axle nut (RHT) allowing the drum to come off.
It's not a current practice and we have no serious issues with wheels falling off, so at best it could be seen as an extra expense and complication to solve a non-existent problem while creating a new problem for non-suspecting owners when the need to change a wheel.
It still is common practice to use a LH thread for mounting a bicycle pedal to the left crank arm.Just 2 months ago I had to hit the trail after only putting my left pedal on finger tight, but after the ride it needed a wrench to get it off.
Palmer is correct about Helicoils and similar products. But they only are beneficial when used on material thick enough to allow them to be installed correctly, which usually does not include sheetmetal. although it is common to put self threading screws into pierced holes in thin sheetmetal, on manyoccasions the result is an inadequate joint, only strong enough to hold correctly until the product is first used. When the hole for the screw is punched instead of being pierced the joint integrity is usually even worse.
Thread engagement in tapped holes in material much thicker than sheetmetal is a totally different case. Many folks assert that most of the load is carried by the first two threads, but they don't realize is that as soon as those threads deform just a bit that the load is then shared by quite a few more threads. Also, that conceptis most true for "avaerage quality" threads. It is not that much extra effort to produce threads that have greater engagement and can carry a larger load than standard threads. So it is a better approach to select threaded connections to be at least 2 screw diameters deep.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.