Stronger steel than 4340...... 20

[Tagger]

I have a Ø6" shaft that necks down to Ø4.70" which is splined. The shaft is subjected to torsion and bending.The splined shaft goes through a cyclical bending as it rotates, and this eventually lead to the shaft breaking after a little over two years of operation. The shaft broke right at the point where the usable spline length ended. Basically the Major Ø of the spline is the Ø of the necked down portion of the shaft (Ø4.70). Torsional-wise the shaft can easily handle the torque, it is just over time the bending fatigued the shaft. Without getting into detailed analysis of the application, what class of steel would better handle the bending (give longer life). Assuming the operating parameters are the same as the previous shaft, (I can't guarantee the customer the shaft will last longer), what would be a better material? The current shaft is 4340 Q&T (don't know the details of the Q&T) to 269-321 BHN. The length of the spline was nitride case hardened to a depth of 0.015-0.020" and Rockwell 15N hardness 90 minimum. Thank you for your support.

 

[saldanha]

It apears to me that the strength of the material will have little effect on the failure if this is by fatigue. it is possible, looking at the location of the failure, this may have been caused by stress concentration due to the design. therefore i would suggest that you look in to the possibility of changing the design

 

[Tagger]

I agree...the other alternative is the increase the diameter of the spline to closely match the diameter of the shaft (Ø6&quot. This would involve changing another component of the design to match the increase spline diameter.

 

[NickE]

saldanha is right the stress concentration of the necking and the splines is the cause for th efatigue failure to occur at that point.

The tensile strength of the 4340 steel is what will predict longer or shorter life for the shaft. The case you describe is the perfect bending fatigue case. Looking up your hardness on my chart gives a tensile strength estimate of 140-150ksi. I see from

http://www.matweb.com

that higher hardness/tensile strengths are possible w/ this steel.

If higher tensile strength material is used you will have longer service life. However there are many other things to consider with this change. Some are: service temperature, shock loading, corrosion, SCC (not sure but usually as strength goes up this get easier), etc...


nick

 

[TVP]

As NickE stated, 4340 can be quenched & tempered to higher strengths. It is often used at ultimate tensile strength levels ~ 260 ksi (> 1750 MPa), which is ~ 490 HB. Keep in mind that higher strength will make this grade more notch sensitive. The consumable electrode, vacuum melted grade (SAE AMS 6414) has better fatigue strength than the regular version (SAE AMS 6409), but if the fatigue is surface dominated, then this may be much of an improvement.

300M is a steel grade based on 4340 with ~ 1.6% Si that enables higher strength for a given tempering temperature. Tensile strength can be in excess of 300 ksi (>2100 MPa). The standard for ordering this grade is SAE AMS 6419.

Another grade to consider is Aermet 100 from Carpenter. It is a very high strength grade that has been used to replace 4340 and 300M in demanding applications like aircraft landing gear. Typical properties are UTS = 285 ksi (1965 MPa), but since it has better fracture toughness than 4340 or 300M, it can be used at maximum strength even in notched conditions. This grade is referenced in standard SAE AMS 6532, and you can find more information at http:/

http://www.cartech.com

However, you will likely obtain more significant improvements in fatigue life by investigating the following:

1. Reduction of stress concentration.
2. Improving surface finish.
3. Introduction of residual compressive stress (shot peening, laser shock peening, roller burnishing, Low Plasticity Burnishing, etc.)

 

[unclesyd]

How is the transition from the 6" to the 4.7" handled, is it stepped with a radius or is it a tapered transition?

Where is the stop off of the nitrided area in respect to the break? There might be a metallurgical notch at this point.

If you have a bending moment on the shaft can you use a more flexible design in the 6" shaft.

You might want to look at Astrally-V Bar for you shaft. You can contact the people at the following Url with the problem. We made a considerable number of our shafts from this material.

http://www.astralloy.com/index2.asp

 

[macmil]

unclesyd beat me to it. Astralloy-V is used for 'problem' shafts and typically outperforms 4340. That said, the shaft design and manufacture should try to avoid stress risers like sharp changes in section.

 

[Tagger]

I will check out Astralloy-V. Also I will try making the step down section of the shaft match the minor diameter of the splines, and have a radius at the end of the splines instead of a typical chamfer. In other words try to eliminate (or reduce) any stress risers at the point of fracture by making a smooth transition, and improve the surface finish.

 

[Carburize]

Have you considered Shot Peening the transition area to improve fatigue resistance?

 

[MOB1]

Many manufacturers of this type of shaft use an alloy steel such as 4340 and induction harden/carburise/nitride the splines. Some are also shot peened to further improve resistance to torsional fatigue at the spline roots.
You do not clearly identify if the cracking is rotational bending fatigue at the section change or torsional fatigue cracking initiating at the spline roots. It is important to make this identification as the solution to the problem will be different in each case.
In both cases it probably boils down to design rather than a metallurgical solution (apart from the benefits of shot peening).

 

[Tagger]

I want to still case harden the splines up to the shoulder where the shaft Ø becomes 6". I also want to shot peen the neck area where the splines end. Which should I do first the case harden, or shot peening? Or does it matter. I am also modifying the design where I am basically reversing the spline profile. The minor diameter of the splines will equal the neck diameter. I can then radius the ends of the splines to eliminate any sharp corners/edges. Then case harden and shot peen (or vice versa) the area. Opinions?

 

[Carburize]

If you are going to shot peen it needs to be as close to the last operation as possible. Any heat treatment after shot peen will destroy all or at least a major part of the beneficial residual stresses.

 

[TVP]

As Carburize said, shot peening must be after case hardening. You can mask off part of the shaft if necessary to avoid shot peening disrupting the surface. You may want to investigate roller burnishing as this can provide the same improvements in residual compressive stress without the negative effect of increasing surface roughness.

 

[yates]

TVP's suggestion to use Aermet 100 would certainly be beneficial, although probably the most costly alternative. We have been making Aermet 100 bolts for a few years now in the 285 ksi class, and these have shown to have good fatigue resistance and excellent resistance to notch initiation.
However, do not lose sight of the basics - exotic materials are fine but sound design is essential! There TVP has it right again, the three pillars of fatigue resistance are there:-
- no sharp changes in section
- the best surface finish possible
- compressive stress (shot peening or fillet rolling on a cylindrical radius)

 

[EdDanzer]

We used to make a torsion bar from 4340 that has a spline on each end. The diameter after the spline must be smaller than the minor diameter of the spline. FEA would better define the shape you need. Core tempering to 30 RC and induction hardening gave the best life. I can’t find the information, but the draw temperature has a major effect on fatigue life. The original temperature of draw work was done by an axle manufacture in the 1960’s or 1970’s.

 

[JSMAT]

Might be worth extending the case hardening beyond the end of the spline so that the edge of it doesn't coincide with the section change. You may also see benefit from using a tougher grade which can be used at higher strength whilst maintaining the toughness you get from 4340 at your current strength level.
JS30B has been used for shafts of a similar size and can be supplied at a strength level of around 1200MPa (~170ksi) min yield, with 50-60 J (35-45FtLbs) RT impact (~388HBN).

 

[71cj5]

The sae baja team I was on used 300m to make shafts of the shape you describe, Just alittle smaller dimensions. 300m is an expensive material, I believe it's also called 4340m. In annealed form, a driveshaft made of 300m, failed catastrophically. Poor surface finish and a improperly made relief for the splines was to blame.
Solution: Revise machining to a good surface finish, replace chamfer relief with radius, and through harden material. The hardening had the side benefit of allowing us to gun bore the center of the shaft for reduced rotating mass. Those shafts now have 3 years on them without a hiccup.

 

[Frederick]

There is a lot of good advice given in this thread. I would like to add one thing: Cryogenically treat the shaft after shot peening to increase the fatigue life.

 

[SMF1964]

Perhaps a maraging steel would be of use? They typically run higher in strength than 4340 steel, but I don't know if you can nitride them (the hardening mechanism is different). Any thoughts?

 

[jrhols]

I agree all this advice is yoda grade stuff, but I would like to add a practical approach. First define the problem.
This is best done if a formal approach is desired in books that discuss 'Quality Planning' or chapter labeled as such.
I would guess it is the radius. It is known that such parts are very sensitive to the exact shape of the radius but there have been very few studies, but such studies of the shape of blend radii do exist. Perhaps a master's degree progam in design engineering would have such a library or try going throught their PHD papers in the old basements. Second - plan on adding a coating to the area no matter how you change the quench or temper (the Q & T you mentioned). This will protect the finish where failure begins. I have studies dynamic failure and found surface polishing in the usual drum polish is very effective.