To Neil Deshpande - re: M5 caliper thread (LONG) (archive)

The complete automotive resource for buyers, sellers, and owners like you.

The reasons why M5 caliper bolts cannot be �stretch� bolts, also known as TTY (Torque To Yield) bolts, are as follows: First of all, when you talk about TTY bolts, they do not follow the same standards as the other bolts and studs used in the automobile industry. The way that bolts and studs are defined in standard engineering practice is that they have a diameter (for example M10) they have a length, and they have a thread pitch. They also have a strength rating. This is usually listed on the head of the bolt, with only some specialized applications without this characteristic. All of these have to follow the machining tolerances and load abilities set forth by different sanctioning bodies such as the ASME.

TTY bolts follow a completely different set of tolerances and loads requirements. To understand these you must understand what a TTY bolt is and how it differs from a regular bolt. While a TTY bolt must also have the diameter, length and thread pitch, it must also include uniform deformation characteristics and also the amount of load to properly stretch the bolt to maximum efficiency. Sanctioning bodies like the ASME also set these.

Now that I have said this, it is time to get onto the differences between stretch and normal bolts. First of all it is important to understand what is meant by stretch or TTY bolts. Stretch means that you tighten up a bolt to the point at where to starts to deform. Once deformation starts, it cannot be reversed without having to completely rework the metal. I am going to use the following graph to explain some terminology about metal materials.

[http://members.roadfly.com/cgraff/graph3.bmp, 200k]

This graph shows 1018 Annealed Steel in a standard tensile strength test. What can be witnessed here is everything required to understand a TTY bolt. First of all, every bolt in the world has an elastic region. This is a region where the bolt can be stretched using an applied load and once that load is released it will, in theory, return back to its original length. (What can cause the bolt not to return to its original length can be many things, from oxidization to fatigue stress.) This is the initial near vertical region of the graph. You will notice that it is linear, and this linear portion characterizes all elastic regions. The point at which the graph goes from linear to horizontal is known as the yield point. That is where the �yield� in Torque To Yield comes from.

In this graph it goes horizontal because we are using annealed steel. If you were using something like cold rolled steel, or a steel with fewer internal dislocations, you would have a curve form at the yield point, rather than the horizontal step. This point signifies the greatest amount of force that can be applied before the bolt�s shape starts to permanently deform. Once past the yield point, the bolt starts to plastically deform. This means that the bolt actually significantly changes shape and will not be able to go back to its original shape and strength if the stress is relieved.

The maximum point on the graph is known as the UTS or Ultimate Tensile Strength. This is the maximum stress that the bolt can withstand. This is also the point where necking begins. Necking is when a material can only stretch in length if it reduces in cross sectional area (due to the fact that the material only has a set constant volume). The initial plastic deformation is characterized by the removal of dislocations, a.k.a. imperfections, within the metal�s microstructure. Once these are removed, the metal reaches the UTS, but to deform even more it must reduce cross sectional area. As noted by the graph, the tensile stress that the bolt can withstand actually decreases to the final point of the graph, which is known as the fracture strength. The fracture strength is how much force the metal will take before it will break.

Now that I have explained to you the graph, I will explain how a TTY bolt compares to a standard bolt and how they can be modeled by this graph. A standard bolt, when used properly, remains in the linear region at all times. This means that when the stress is relieved, it will return to its original shape and will retain its original characteristics. A TTY bolt on the other hand is tightened into the curved region, and so if the force is relieved on the bolt it actually becomes weaker if an attempt is made to reuse/retorque it. So, the TTY bolt will actually become longer when it is torqued properly. This property is irreversible and so that along with the other reasons are why you are told to never reuse TTY bolts. Also if you retorque them, it is VERY likely that you will pass the UTS and the bolt will fail much more easily (because the UTS becomes lower than the pre-stretched UTS), and that is why you do not retorque head bolts.

Some more differences between TTY bolts and Standard bolts is the heat that they can withstand before failure. Unfortunately I do not have a graph for this, but when a bolt becomes heated, it actually cannot withstand as much stress. But a classic example of this is the lug bolts on BMWs. When you go to torque your lug bolts, if you haven�t driven the car, you can torque them without any problems. If you have just driven the car hard (or a track session), or have gotten the bolts quite warm, you will notice that if you try to re-torque, you will destroy them, because they will plastically deform. You will also notice that it is impossible to tighten the bolt to its previous torque setting. The bolt will just continue to deform until it fractures. This is why people who track their cars a lot replace their lug bolts more often than your average BMW owner. The more heat cycles heat the bolt goes through and the more times you torque the bolt, the more fatigue stress you apply to it.

A TTY bolt that is heated, will actually become slightly weaker allowing for more deformation, just like a standard bolt. But the TTY bolt cannot withstand as much added load before it breaks, versus a standard bolt in its correct application. If the amount of stress on the bolt is not significantly changed, or the heat is kept relatively low compared to the recovery and recrystallization temperatures, this deformation is reversible.

[To fully explain: Recovery, recrystallization and grain growth are what metals undergo at high temperatures and/or long times at temperature. For example, when you stretch a piece of metal, you are actually increasing the stress on the lattice that makes up the material. This actually makes it easier to fracture the material, as show by the graph. When you heat a metal, it actually first tries to remove all of the remaining dislocations which is the process or recovery. Recrystallization, is when the grains that make of the individual lattices of the material try to realign themselves with each other to try to make one continuous grain throughout the metal. Grain growth is what happens when the grains have aligned during recrystallization and grains start to combine to make larger grains. This relieves stress in the lattice, but also requires energy. This is why it only happens at temperature. It is a different temperature for different materials, but is also a function of time.]

If you have ever torqued a bolt into the plastic deformation region (which you have if you�ve ever snapped a rusty bolt, overtorqued a very small bolt, or put in new head bolts) you know that the feeling of torquing bolt into the plastic deformation region feels totally different than torquing a bolt into the elastic region. You can note that the amount of force it requires to rotate seems to be high, this is the region of the graph between the yield point and the ultimate tensile strength. If you continue past the ultimate tensile, you note is that it is easier to rotate, but in fact you are plastically deforming it until it snaps.

Also, you will note that in all tech manuals, that if you are using a stretch bolt, or TTY bolt, the torquing method is completely different than that of regular bolts. Initially you torque the stretch bolt down to a certain torque. Then you retorque (in BMW head bolt case), and then there is a degree measurement. This is also accompanied by different heat characteristics. E28 5-Series Bentley Manual, pg. 28, sec. 4: �Cylinder head bolt tightening: Stage 1: 40 +/- 1 ft-lb, wait 20 minutes.� This 20 minute wait allows the bolt to cool, since it has just been torqued and heated up from the friction. This heat could change the place where the bolt will plastically deform, i.e. thread, shank, etc. You do not want the bolt to stretch at the thread, and that�s what would happen if you continued to tighten without the wait. �Stage 2: 59 +/- 1 ft-lb, then run engine until fully warm for about 25 minutes.� Running the engine for 25 minutes, allows the total bolt to heat up to the same temp throughout. This is important because since the head bolt is at a raised temperature, it is now possible to stretch it plastically. It would�ve been much more difficult to stretch it plastically if it were at a lower temperature. This extra required stress (if stage 3 were completed without heating) could lead to head cracking and failure.

�Stage 3: After the running the engine for 25 minutes, retorque bolt to 35 +/- 5 degrees.� To properly a stretch a bolt plastically, you cannot give a torque setting to a proper amount of plastic deformation. The reason being that this could change depending on the bolt, and also that not every bolt will the stretch the same amount for that torque setting due to the bolt manufacturing specifications. Also like in the lug bolt case, it is almost impossible to reach a measured tightening torque. The degree measurement ensures that all bolts will stretch the same amount � which is required of stretch or TTY bolts. Now, as your previous statements, the BMW manual only lists a �-� in their torque settings. This means, it states a maximum torque setting, if you torque above that you will decrease the remaining elastic region that the bolt needs to be able to remain the elastic region during operating cycles. BMW could only use a stretch bolt in the calipers if they gave a degree measurement with the torque, since they did not this another reason why they are NOT TTY bolts.

There is one of reason why BMW and other manufacturers use stretch bolts for head bolts, however. When the bolt is heated to a relatively low temperature, compared to its recovery and recrystalization temps, [Note: Plain iron�s recrystalization temp is 840 degrees F, just as an example. Steel is actually higher and depending on the past heat treatment applied to the steel, that heat treatment can increase the recrystallaztion temperature.] it is possible for the head (made of aluminum) and the block (made of steel) to expand at different rates and not apply an increase in load to the components but apply it to the stretch bolt. The bolt will take up the extra force created by the difference of expansion rates of the steel versus aluminum, and keep the total clamping force relatively similar to its initial clamping force. This insures that the bolt will not fail, the clamping force remains nearly constant, and you don�t blow headgaskets, etc. Then you may ask, why do individuals suggest using non-stretch bolts/studs, like ARP�s head studs? The answer would be that the non-stretch bolt requires a lower initial clamping force (because the bolt doesn�t have to be deformed like the TTY bolt), and therefore uses it�s elastic region more effectively to take up the strain caused by the expansion of the head vs. block than a stretch bolt would. The initial strain is also lower (due to lower initial torque of the non-stretch bolts), and the added strain from expansion won�t overstress the system.

BMW says never to reuse many bolts in their shop manual. This is because when using your brakes in normal everyday applications, you cycle the loads on the caliper bolts. This cycling causes fatigue, which could eventually lead to bolt failure. The same applies to suspension bolts, and the like, where the stresses on the bolts are cycled. If you read a good manual it will always suggest replacing bolts that normally go through cycles. It does NOT mean that they are TTY, just the simple fact that if you cycle stress a bolt, it causes fatigue that can lead to failure. If BMW does not recommend to replace cycled bolts, then the lawyers would be screaming bloody murder. Since these specific bolts in the caliper are cycled, BMW always recommends using new ones whenever you do work to the car. However, since these bolts should remain in the elastic region, the general consensus is that unless they have deformed past their yield point, they are still reusable. Once past their yield point, they become stretched (which you can see and measure), and should be immediately thrown out. Any good engineering or race car book � e.g. Carroll Smith�s �Screw to Win,� a.k.a �Nut�s Bolts and Fasteners to Win.� � will tell you this EXACT thing � that if a bolt has stretched, it is JUNK. But if it is not stretched, it is still useable.

To give you an idea of just how much stress this bolt cycles through, lets do a rough �back of the envelope calculation.� ATE requires that during brake rebuilding, the pressure needed to press out the pistons has to be 10 bar. Now, all of their brake components are listed at 120 bar maximum (http://www.contiteves-am.com/english/indexmit.htm), bar a few different components. So, let�s roughly assume that you�ll be putting in about 50 bar during a reasonable braking test. So�there is 50 bar of pressure throughout the system pushing on the cylinders (let�s use the 4-pots as an example, and lets convert to PSI).

That's 725 PSI of pressure on EACH cylinder. To convert to force (given the diameter of the piston at 40mm, which gives a 1.95 in^2 area) you multiply by the area the pressure acts on. This calculates to about 1413.75 lbs of force on each cylinder.

Multiply by two pistons on each side, and you get 2827.5 lbs of force from the pistons pushing in on the disk from each side. This is just to give you an idea of forces involved.

Now�the reaction force of the disk on the piston/half-caliper system (since it is no longer moving once it has contacted the disk) is going to be equal to that. While there technically is a bending moment, it doesn�t have to be considered since the caliper acts as a rigid body once bolted together. Moments can only be taken around joint or hinge pin models. The reaction force of the caliper bolts holding together each half of the piston is the same, assuming equilibrium. If it weren�t you�d either drive through the disk or rip apart the caliper, which you don�t see happening.

There are 4 bolts, with 2827.5 lbs of force pulling them on each side. So, divide the total tension force (which is 2827.5*2 lbs) by 4 gives you the force exerted on each bolt. That�s 1413.75 lbs of force each bolt has to sustain.

If they are roughly an M8 size bolt, that means the shank is about 8mm in diameter, which gives a cross sectional area of 0.0779 in^2; let�s use 0.08 in^2 for ease of calculation.

Since stress of the bolt is measured in PSI for us, we divide the force the bolt has on it by the cross sectional area and get 18,135 PSI. The bolt has to withstand a cycle of about 18,000 PSI. And that�s not including full braking force, or other heat issues (remember that disks can reach as high as 1300 degrees F and calipers as high as 800 degrees on the track! And yes, I�ve measured the temperature of those calipers and rotors myself!).

There is a short and easy equation that relates torque and tension. Torque = (Tension)*(rho)*(bolt diameter). Tension is the force exerted on the bolt, in lbs. Torque is the torque setting in ft-lbs. Rho is a constant for the bolt, and the diameter is in ft. This equation is in no way perfectly exact, but gives a good estimate. In order to get the exact tension you�d need a lot of the fine specs of the bolt (thread angle, coarse vs. fine threads, number of threads engaged, etc.) But this equation is valid for most cases.

The bolt has an 8mm diameter, which translates to 0.02625 ft. The torque setting of the bolt is about 41 ft-lbs according to your factory manual. Rho is between .2 and .12 depending on the bolt, and whether it�s torqued dry, oiled, etc. For our purposes, it�s a dry, non-locktite (i.e. the simplist), and rho will be about 0.18. Therefore the tension put on the bolt by these parameters is: (41 ft-lbs) = (F)*(0.18)*(0.02625ft), F = 8677 lbs, or rounding up to about 9000 lbs of force.

To calculate the stress of 9000 lbs for force, you divide by the cross section area of the bolt. This comes out to an estimated 112,500 PSI. This equation, however, doesn�t take into account the friction of threads, and so forth when initially torquing the bolt.

The actual tensile force in the bolt is about 80% of the initial torque under these conditions (see any online bolt calculator to check parameters of bolt and conditions of tightening). Again, I�m slightly overestimating, so we�re making certain that these bolts will hold, and plus it'd take much more information about the bolt to get the most accurate measurements. So, in fact, the actual tensile strength is about 90,000 PSI on the bolt.

These bolts are grade 11.9 on the M5. Do you know what that means? It means that they have a UTS of over 165,000 PSI, and a Yield strength of over 135,000 PSI. They are stronger than SAE grade 8, or Metric grade 10.9 and 8.8 which is what is most often used on your suspension!

Summing the stresses involved (our initial torque setting of 90,000 PSI, and the cycling at 18,000 PSI, comes to 108,000 PSI) we are well within the yield strength, let alone the UTS. If these bolts were 8.8, they�d be near their UTS by now. If they were 10.9, they�d just about have reached their Yield strength and just start plastically deforming. And when these bolts start plastically deforming you�d be SOL. And you don�t see many M5/M6 drivers plastically deforming their caliper bolts, leaking brake fluid, and having their brakes fail on an everyday basis, now do you?

Questions, comments, or problems, please visit the Roadfly help desk.