What's new
What's new

Metallurgists: I need advice on material selection to avoid stress-fracture

You're right that it's hard to evaluate all the factors like specific loads and stress concentrators.

But, you can probably draw ok conclusions from S-N curves from slightly different conditions. Remember that these are log scales, so if the curve for 17-4 looks just a bit higher than 304 (or whatever) it might actually be like 5x the life.
 
The 5/8” 304SS is no match for that 1 3/4” sch-40 steel pipe when it comes to load/displacement, prolly a wise feature for a solid foundation with a moving whip though.

3XX stainless doesn’t like to be upset/bent back & forth (not cool with austenitic iron based steel). 1045 TG&P will get you similar strength, 4340/300M a bump up in fatigue & strength. Yo mama would be 9310 carburized & HT, kinda like a tube in tube structure (be cashy too).

I’d also agree with Modelman that it may be folks fooling with it. Kids do some ass-hat things.

Good luck,
Matt
 
Thanks, guys.
Mark, you referenced pipe/tubes but the part that failed is the solid SS top shaft. The pipe, whether properly long or un-authorized short has not failed.

Matt, if it were kids they'd have to give up too much phone-time to cause these fractures. I'm under the impression they are typical long-finger breaks along one plane, with sharp edges. No bending of the failed area is visible and from what I understand, its not an area where pedestrians would go unnoticed. I'll see if I can get photos of the actual break.

Three of you have suggested 17-4, thank you. I've never used it and we'll look into that. If suitable it would avoid the nickel plate operation.

The bike-ride was phenomenal! First ride in almost six months due to another shoulder surgery and I was going nuts without my fix.
 
One thing came to mind, but don't know if it is relevant.

The thing on top of the stinger has both mass and aerodynamic drag in the wind, and so this plus the stinger may be resonating at some wind speed.

To check, calculate the cantilever resonant frequency with the bottom clamped and the top end (with the top mass) free.

The formulas are in Roark's Formulas for Stress and Strain.

Then calculate the aeolian (vortex-shedding) tone frequency.

.<https://en.wikipedia.org/wiki/Aeolian_sound>

If this is the problem, there are some simple ways to solve it, like a spiral on the outside of the stinger.
 
Thanks, Joe.
It is surely relevant. We're aware of the probability of resonance but the formulas and calculations you list are way beyond my cranial capacity. I almost flunked out of high school algebra! I'll forward the info though. For this one installation I think we're ok with trying a material change and watching it for a while.
 
Yeah, about the only thing you can do is use an overkill material which can sustain the higher stress without fatiguing. Do you have a good way to recalculate the load for the deviant shortened installation? The good news is that you have lots of room to improve from 1018/304 in terms of strength. If it's just for this one time, I'd go with something seriously strong (according to Engineering Toolbox, 18-8 stainless gets you up over 70 ksi fatigue, and also mentions a rule of thumb for most steels that the fatigue limit is near half of its ultimate tensile strength).

Another bit of good news is that going with a stronger material shouldn't have much impact on the stiffness, if that's relevant to the way the assembly works.


WTF??? !8-8 will be stronger than 304????? What planet are you on? What is 304 if not 18-8?
 
WTF??? !8-8 will be stronger than 304????? What planet are you on? What is 304 if not 18-8?

Hah, it sure sounds like I said that. Brain fart on my part since I didn't recall that they are functionally the same.

What's actually interesting though is that 304 annealed is pretty wimpy, but it can in fact be conditioned up to the really solid numbers listed under 18-8 in that Engineering Toolbox page. Where the plot really thickens is that it's done by cold working :scratchchin:

But the way that 304 Condition B is made would be rather different than work-hardening in the field, likely as it work hardened itself there were some geometry changes and stress concentrators formed within the material from uneven bending, especially because the load direction is variable and somewhat random.
 
Thanks, Joe.
It is surely relevant. We're aware of the probability of resonance but the formulas and calculations you list are way beyond my cranial capacity. I almost flunked out of high school algebra! I'll forward the info though. For this one installation I think we're ok with trying a material change and watching it for a while.

If the cause is this kind of resonance (an example of aeroelastic flutter), making the material stronger won't help. Won't hurt either. You must instead raise the resonant frequency and/or break the vortex-shedding mechanism up in sufficient degree.

The basic problem is that when aeroelastic flutter is dominant, the amount of energy available to twist the structure is unlimited, so the amplitude will increase until something is yielding.

Here are some poster children, a test aircraft and Galloping Gertie (#AIRBOYD #AvGeek Aeroelastic Flutter):

Aeroelastic Flutter - YouTube
 
Fascinating, Joe! In this case we're not dealing with twisting, just random variable bending. The sway is never consistent, probably never twice in exactly the same direction, speed or deflection. You've given me some resources which I appreciate.
 
Fascinating, Joe! In this case we're not dealing with twisting, just random variable bending. The sway is never consistent, probably never twice in exactly the same direction, speed or deflection. You've given me some resources which I appreciate.

Although the striking examples involve twisting, this is not required. The reason that many automotive whip antennas have a helical strake ridge along the length, is to suppress flutter due to airspeed past the vehicle.

See "Mitigation of vortex shedding effects" in the following:

Vortex shedding - Wikipedia
 
so you need stronger stainless steel for a flexing component? (sorry, i couldnt read the whole thread.)

springs are routinely made of 302/304/316. next step up is 17-7ph. next 15-5ph and then custom steels, e.g. carpenters 450, 13-8, 455 up to475 (2gpa uts). best damage tolerance: 465/h950 (k1c of 90).

fatigue strengths is a funny business (asm has compiled 2592 pages on it). they say, that the endurance limit seen for a steel is more an artefact of the test method then a real limit. thats because it depends on several variables, one of them being shape/geometry, that may result in a totally different ranking. (many engineers have fallen into the trap of using data for an unsuitable frequency.)

i believe your problem could be to have used 304 annealed. 304 40%cw will have a fatigue limit more than 2x higher. i guess its not that the material is workhardening but that its not workhardening enough. dont interpret any ductility into fatigue data. as a very rough rule of thumb use 40% of uts as endurance limit. so for mild steel 200 mpa, 304 aneealed 300, 304 40%cw 500, 17-7ph 600, 15-5ph/h1025 650 mpa. the most damage tolerant, common steels ar still aermet 100/310 (k1c for 100 grade 115 mpa/m^2).

i have no idea of availability of the steels and guess the best solution would be a mechanical one.
 
The basic problem is that when aeroelastic flutter is dominant, the amount of energy available to twist the structure is unlimited, so the amplitude will increase until something is yielding.

I forgot to mention one thing, that aeroelastic flutter has fundamentally the same cause as chatter in machining, where the power driving a moving media (be it air or metal) is tapped to drive a self-sustaining oscillation.

 
Milland, no torsion at all. The outside assembly rotates freely on dual ball bearings or a ball-bearing/bronze bushing pair, and on larger parts an additional thrust bearing.

Rickyb, pure bending. The amount of bend would in most cases be invisible to someone driving by, except in a very strong wind. The stress is repetitive, the direction from which it comes varies and the cycle count is unknown since it is determined by the whims of mother nature. I think we're ready to try the 4340 shafts and see what happens. The stainless failed after about one year in operation.

At present there is no plan to modify production with a different material, just trying to fix this one because it is so different.

You'd really have to put at least 2 seismic transducers at 60° angle of rotation at the bottom of that shaft & record for a period of time to see what if any movement (acceleration/velocity & vectors) is happening).

The 4340 isn't nuts expensive & actually machines well. I usually don't buy stainless on purpose unless I need the corrosion protection…

Just for laughs Gordon, check the broken end around the break with a pocket magnet. If you feel any attraction you've changed some iron properties in that area. When similar units run for years & this one fails around 1yr it's depressing. Also if some part's rotating around the shaft It should be fairly well balanced even if its small.

For laughs I checked ASM (vol 1, properties & selection of metals – the Section for Fatigue Resistance) & found a comparison of carburized through hardened & induction hardened hydraulic main shafts all HT’d to equal strength condition from 1035, 1137, 4320 & 4140. I’d have thought the 4320 (carburized) would outrun everything… Nope, the induction hardened 1137 resulfurized sewage ran 1.18M cycles before the test was terminated without failing… The 4320 ran 0.8M cycles & broke. The real kicker is the “all in” price for the 1137 was 31.8% cheaper per part. Out of the box 4140 could get the hardest if you wanted but it only ran 0.41M cycles. Some deep poo-poo out there guys.

Good luck,
Matt
 
What's the actual strength metrics needed?

Would 1144 anti stress/fatigue steel ("Fatigueproof") work?
 








 
Back
Top