Service Life of Steel H-Beam Rods

[GRTfast]

This makes a ton of sense to me.

Is the conclusion then:

If the stress is below that limit, never replace steel rods unless there’s a specific reason.

Yes. The kicker is, how well does one understand the actual stress state a rod is subjected to?

Since it’s mostly an issue of tension, isn’t that a mass vs cross-sectional area vs. acceleration computation? In any case, you’d think that the rod manufacturer can compute the number given the component weights, geometry, and engine speed?

Yes, but there are shape factors, material issues, etc..

You are right though, the calculations are fairly straight forward given good materials and manufacturing processes, especially using finite element methods to understand the stress state(s).

[lekid]

Let me ask another question along the same line:
How often should rod bolts be replaced? Is it based on heat cycles or how many times they were torque? If it is torqued related (which I would assume it is since that is when you are really stretching the bolt), how many re-torque is too many?

[swampbuggy]

How many re-torques are too many ? When you mic. (measure the length of the bolt) when it is new, it will be a given length. Each time you remove the rod bolt at disassembly you measure it/them. When the bolt does not check to its original new length it is said that the bolt has yielded (to give up under pressure) Webster's new world dictionary/second college edition. This is when you should consider the rod bolt as junk. Mark H.

[GRTfast]

How many re-torques are too many ? When you mic. (measure the length of the bolt) when it is new, it will be a given length. Each time you remove the rod bolt at disassembly you measure it/them. When the bolt does not check to its original new length it is said that the bolt has yielded (to give up under pressure) Webster's new world dictionary/second college edition. This is when you should consider the rod bolt as junk. Mark H.

I’d agree with that as a good rule of thumb. The bolt load should be designed such that the tension in the bolts exceeds the cyclic force they are subject to during operation (thermal stress, mechanical stress) . If that is true, then the only stress cycles they see are from assembly and disassembly. If the bolts aren’t yielded, but the stress on them (from torquing) is above the infinite fatigue life line on the Goodman curve, they will eventually fail from fatigue even if they’ve never yielded, however the number of assembly/disassembly cycles it would take to cause fatigue failure is probably high, like in the hundreds or thousands. THAT said, I’d replace them every time.

[CharlieB53]

No doubt mic’ing rod bolts should indicate any change in length suggesting eminent failure the same may be true with rod length.

I don’t have the tooling to measure accurate rod length but I’ sure someone does. The cost could be well worth the peace of mind.

[digger]

How many re-torques are too many ? When you mic. (measure the length of the bolt) when it is new, it will be a given length. Each time you remove the rod bolt at disassembly you measure it/them. When the bolt does not check to its original new length it is said that the bolt has yielded (to give up under pressure) Webster's new world dictionary/second college edition. This is when you should consider the rod bolt as junk. Mark H.

I’d agree with that as a good rule of thumb. The bolt load should be designed such that the tension in the bolts exceeds the cyclic force they are subject to during operation (thermal stress, mechanical stress) . If that is true, then the only stress cycles they see are from assembly and disassembly. If the bolts aren’t yielded, but the stress on them (from torquing) is above the infinite fatigue life line on the Goodman curve, they will eventually fail from fatigue even if they’ve never yielded, however the number of assembly/disassembly cycles it would take to cause fatigue failure is probably high, like in the hundreds or thousands. THAT said, I’d replace them every time.

thats not true they see every acceleration event at TDC regardless of the preload so they are subjected to a variable tensile load which can fatigue them, its just not the full amount of tension when preloaded and designed correctly

[GRTfast]

How many re-torques are too many ? When you mic. (measure the length of the bolt) when it is new, it will be a given length. Each time you remove the rod bolt at disassembly you measure it/them. When the bolt does not check to its original new length it is said that the bolt has yielded (to give up under pressure) Webster's new world dictionary/second college edition. This is when you should consider the rod bolt as junk. Mark H.

I’d agree with that as a good rule of thumb. The bolt load should be designed such that the tension in the bolts exceeds the cyclic force they are subject to during operation (thermal stress, mechanical stress) . If that is true, then the only stress cycles they see are from assembly and disassembly. If the bolts aren’t yielded, but the stress on them (from torquing) is above the infinite fatigue life line on the Goodman curve, they will eventually fail from fatigue even if they’ve never yielded, however the number of assembly/disassembly cycles it would take to cause fatigue failure is probably high, like in the hundreds or thousands. THAT said, I’d replace them every time.

thats not true they see every acceleration event at TDC regardless of the preload so they are subjected to a variable tensile load which can fatigue them, its just not the full amount of tension when preloaded and designed correctly

The entire bolt sees (essentially) the same load from acceleration. Without doing any calculations, I’d guess that load is minimal, even negligible. I could be wrong. Do the math. I’m just providing some general principles.

[digger]

I’d agree with that as a good rule of thumb. The bolt load should be designed such that the tension in the bolts exceeds the cyclic force they are subject to during operation (thermal stress, mechanical stress) . If that is true, then the only stress cycles they see are from assembly and disassembly. If the bolts aren’t yielded, but the stress on them (from torquing) is above the infinite fatigue life line on the Goodman curve, they will eventually fail from fatigue even if they’ve never yielded, however the number of assembly/disassembly cycles it would take to cause fatigue failure is probably high, like in the hundreds or thousands. THAT said, I’d replace them every time.

thats not true they see every acceleration event at TDC regardless of the preload so they are subjected to a variable tensile load which can fatigue them, its just not the full amount of tension when preloaded and designed correctly

The entire bolt sees (essentially) the same load from acceleration. Without doing any calculations, I’d guess that load is minimal, even negligible. I could be wrong. Do the math. I’m just providing some general principles.

i'm not sure what you mean but whatever sinusoidal tensile load that the rod is subjected to throughout a cycle the bolt will see about 1/4 the magnitude would be my guess based on relative stiffness's

[GRTfast]

thats not true they see every acceleration event at TDC regardless of the preload so they are subjected to a variable tensile load which can fatigue them, its just not the full amount of tension when preloaded and designed correctly

The entire bolt sees (essentially) the same load from acceleration. Without doing any calculations, I’d guess that load is minimal, even negligible. I could be wrong. Do the math. I’m just providing some general principles.

i'm not sure what you mean but whatever sinusoidal tensile load that the rod is subjected to throughout a cycle the bolt will see about 1/4 the magnitude would be my guess based on relative stiffness's

If the bolt preload is above the tensile load on the rod, the load on the bolts will barely increase (maybe something like 1% of the load on the rod, because of the composite elasticity of the system). This is why bolt preload is so important.

A thought experiment. Imagine you have two plates bolted together with a bolt and nut. You tighten the bolt/nut until it has 500 lbf of tension, which is also clamping the plates together with 500 lbf.

Now, you grab the plates and try to pull them apart. Say you pull with 100 lbf. What is the tension on the bolt? Still 500 lbs. all you’ve done is relieve the reaction force where the plates interface each other. It was 500, now it’s 400, and that 400 plus the 100 you are applying is adding up to the 500 lbf tension from the bolt preload. In order to increase the load on the bolt, you have to exceed the preload value.

I’ve done this experiment. I have a couple patents on some high current bolted electrical joints in the end windings of large turbine generators. It’s counter intuitive but it’s true.

http://www.boltscience.com/pages/basics2.htm

[bill jones]

---this left side picture is a Carrillo rod with one crack---and the second picture is a Carrillo rod with 3 cracks between the red marks.
---two different engines from good late model stock car 8000rpm engines in the late 1970's.

---both engines had two full seasons of short oval track racing.
---this was when we could only get about 13 honest races out of any engine with fully prepped stock rods and making about 500hp.

---so we were happy to be able to get two full years and finding these cracks were our best way of saving our engines.
---those cracks go over under the rod bolt head and down into the hole about the same distance

[digger]

The entire bolt sees (essentially) the same load from acceleration. Without doing any calculations, I’d guess that load is minimal, even negligible. I could be wrong. Do the math. I’m just providing some general principles.

i'm not sure what you mean but whatever sinusoidal tensile load that the rod is subjected to throughout a cycle the bolt will see about 1/4 the magnitude would be my guess based on relative stiffness's

If the bolt preload is above the tensile load on the rod, the load on the bolts will barely increase (maybe something like 1% of the load on the rod, because of the composite elasticity of the system). This is why bolt preload is so important.

A thought experiment. Imagine you have two plates bolted together with a bolt and nut. You tighten the bolt/nut until it has 500 lbf of tension, which is also clamping the plates together with 500 lbf.

Now, you grab the plates and try to pull them apart. Say you pull with 100 lbf. What is the tension on the bolt? Still 500 lbs. all you’ve done is relieve the reaction force where the plates interface each other. It was 500, now it’s 400, and that 400 plus the 100 you are applying is adding up to the 500 lbf tension from the bolt preload. In order to increase the load on the bolt, you have to exceed the preload value.

I’ve done this experiment. I have a couple patents on some high current bolted electrical joints in the end windings of large turbine generators. It’s counter intuitive but it’s true.

http://www.boltscience.com/pages/basics2.htm

the proportional of load seen by the bolt is proportional to the relative stiffness between bolt and rod. it is a lot bit more than 1% but alot less than 100%. draw a joint diagram and calculate the respective stiffness's it will be about 25%

http://www.boltscience.com/pages/basics4.htm

[GRTfast]

i'm not sure what you mean but whatever sinusoidal tensile load that the rod is subjected to throughout a cycle the bolt will see about 1/4 the magnitude would be my guess based on relative stiffness's

If the bolt preload is above the tensile load on the rod, the load on the bolts will barely increase (maybe something like 1% of the load on the rod, because of the composite elasticity of the system). This is why bolt preload is so important.

A thought experiment. Imagine you have two plates bolted together with a bolt and nut. You tighten the bolt/nut until it has 500 lbf of tension, which is also clamping the plates together with 500 lbf.

Now, you grab the plates and try to pull them apart. Say you pull with 100 lbf. What is the tension on the bolt? Still 500 lbs. all you’ve done is relieve the reaction force where the plates interface each other. It was 500, now it’s 400, and that 400 plus the 100 you are applying is adding up to the 500 lbf tension from the bolt preload. In order to increase the load on the bolt, you have to exceed the preload value.

I’ve done this experiment. I have a couple patents on some high current bolted electrical joints in the end windings of large turbine generators. It’s counter intuitive but it’s true.

http://www.boltscience.com/pages/basics2.htm

the proportional of load seen by the bolt is proportional to the relative stiffness between bolt and rod. it is a lot bit more than 1% but alot less than 100%. draw a joint diagram and calculate the respective stiffness's it will be about 25%

http://www.boltscience.com/pages/basics4.htm

I’ll take a look, I didn’t think it would be that high, but the joints I am intimately familiar with use high strength stainless fasteners to clamp copper plates together. I can see the higher stiffness of the base material (the rod) effecting the overall load distribution. Good conversation.

Edit:

Upon further reading, it turns out that high stiffness joints with lower stiffness bolts result in less of the external load being applied to the bolt. I’ll mess around with these diagrams when I get back to work later this week. Cool stuff.

http://www.boltscience.com/pages/basics5.htm

[Rick!]

---this left side picture is a Carrillo rod with one crack---and the second picture is a Carrillo rod with 3 cracks between the red marks.
---two different engines from good late model stock car 8000rpm engines in the late 1970's.

---both engines had two full seasons of short oval track racing.
---this was when we could only get about 13 honest races out of any engine with fully prepped stock rods and making about 500hp.

---so we were happy to be able to get two full years and finding these cracks were our best way of saving our engines.
---those cracks go over under the rod bolt head and down into the hole about the same distance

On the pictured rods:
How long is the run time per race?
1 or 2 10-15 lap heat races and a 25 lap main?
Timed laps for heat race placement?
asphalt? dirt?

When I back calculate a few assumptions on a NASCAR engine, they can easily get a million cycles on them in a two hour race.
Pretty sure they have pressure traces to calculate the the combustion load and the exhaust stroke load is pretty easy to model and apply so they know their fatigue life predictions.
Doing the same on my brother's drag race engine, around 750 laps is around a half a million cycles and I'm just getting a feel on what a replacement program looks like on that engine at that power level.
It would be interesting to take a swag at predicting the life of your conrods. The failure locations appear to reflect the tensile load of the piston during the exhaust stroke, combined with the stiffness change between the bolt "barrel" and the cap body, IMHO...

[econo racer]

just think how long those rods would last if they stopped the victory burnout session, might get a whole season.

[swampbuggy]

The burnouts aren't any harder on the rods than racing laps, probably not AS hard on them. This is my personal thought. Mark H.