Recommended Shaft Surface Finish for Fatigue 2

[natebott]

Looked through a number of posts related to surface finish and fatigue life in shafts. I was unable to find the specific answer that I am after. Are there recommendations or suggestions for the surface finish to have on a shaft (transmitting up to 1200HP) to best function in fatigue? Are there any industry standards, studies, etc. that support a certain range? I apologize ahead of time if I missed a post that answered this.

Thank you,
Nate

 

[MintJulep]

Smoother is better - all other factors unstated.

 

[hydtools]

This may lead you to another source of information.

http://www.asminternational.org/content/ASM/StoreFiles/05224G_Chapter14.pdf

Ted

 

[TVP]

Can you provide some basic information like the size of this shaft, expected operating temperature, environment, cost sensitivity, etc.? Are the loads primarily in tension, bending, torsion, or contact? An automotive camshaft may have a rough as-cast (Ra ~ 10 to 25 µm) or as-forged surface (Ra ~ 6 to 10 µm) except for the cam lobes, which are surface finished to Ra ~ 0.1. Conversely, the piston rod for a hydraulic cylinder that must seal high pressure oil along its length in addition to resist bending stresses is induction hardened then ground, hard chrome plated, and polished to Ra < 0.1 µm. Both require fatigue resistance, but the details are different.

 

[rb1957]

like MJ above ...

smooth is good, smoother is better, smoothest is best !?

unless of course you shotpeen the surface as a fatigue life improvement,
then cheapest surface finish is best (as the shotpeening will destroy it pretty well !).

but surface finish is pretty low down the (my) list of things to worry about as far as fatigue life is concerned. i'd start with cyclic stress, then material.

 

[natebott]

Shaft material is 4340, currently machined on a lathe, it is a steped shaft with bearing surfaces on each end.

 

[dhengr]

I agree with rb1957, but to add to his list: geometry and shaft size are critical, and at least as important as surface finish, as they set the stress magnitudes; then shape, and generous radii and transitions at dia. changes, etc. Look in a good Machine Design text for shaft design and the various stress raisers due to steps, etc., and means of reducing these stress raisers, etc. Maybe pay more attention to surface finish in dia. transition regions, since these are typically stress raiser regions.

 

[dvd]

Seems like surface finish won't matter in areas of lower stress, and might be controlled by the attachment devices in higher stress areas. What is attaching to the shaft? How is it attached? Is the area of maximum stress coincident with the attachment(s)? Is 1200 HP a high speed shaft or low speed? Splined? ....

 

[natebott]

Please see attached pdf for app info. The area I have highlighted in pink is the area I am most concerned about the finish requirement.

 

[rb1957]

i imagine the bearings are press-fit onto the shaft, yes? so that controls the surface finish, 64 ? 128 ??

away from the bearing, why would you want more than stadrdard smooth finish, 32? 64 ??
but then i guess that's your question, what is the fatgiue benefit for increased surface finish.

i remember seeing somethng awhile back quantifying the fatigue life improvement with improved surface finish. i don't remember the details but thought it was an expensive way to increase fatigue life.

how close is your fatigue life to the goal ?

or is the question "how do i defend the smoothness i've selected ?"

 

[hydtools]

This ref proposes some quantitatve fatigue modification values for surface finish and other surface effects.

http://www.eng.cam.ac.uk/DesignOffice/cad/proewild3/usascii/proe/promec/online/fatigue.htm

Ted

 

[Tmoose]

page 82/83 here-

http://www.timken.com/en-us/Knowled...s/Timken_Practical_Data_For_Metallurgists.pdf

As RB1957 said, how are the bearings, especially the bearing near the sheave/pulley mounted and secured to the shaft? Unless this is a fractional HP device, I hope not with commercial setscrew or eccentric collar. Their holding performance in the presence of rotating varying side loads like belts is poor, and they eat and etch the shaft right where the moment may be highest.

http://lejpt.academicdirect.org/A20/049_060_files/image011.jpg

Also I'd do my cyphering using belt tension 3 or 4 times higher than than any prediction from the belt manufacturer.

Rather than machine tapered shafts, hubs and bushings, I'd try to used a commercial tapered locking hub ( ringfeder, etc) sleeve on an easy-to-make straight cylindrical shaft

 

[geesamand]

4340 is strong steel but also more notch sensitive than a less strong steel. 4340 is more geometric and surface finish sensitive in fatigue. There are rotating fatigue applications where I prefer the low carbon steel material.

That said, elimination of scratches that are transverse to the rotating axis (as in turning marks), the largest possible radii at the shaft steps, and shot peening all improve fatigue life.

David

 

[mfgenggear]

for aviation application, I have seen 32 micro. on bearing journals

Mfgenggear

 

[bcd]

Google "MAchine Design Karpov Curve" and you should get to a graph that shows the various types of surface finishing methods and the resulting reduction in endurance strength. Many of the good Machine Design textbooks include this curve in the fatigue section. The impact of surface finish is nearly negligible with lower tensile strength materials. However, the difference increases with the tensile strength.