s-n curve 301 spring temper

[strokersix]

Can someone point me to a source for fatigue limit curves for 301 stainless strip? I'm working on a part .01 inch thick, .25 inch wide and about 1 inch long. It's a spring detent function. Loading is in one direction only, lifetime cycles only a few hundred, normal operating position approximately half deflection, static load. I calculate bending stress at max deflection of 160ksi. Would a material with 200ksi proof strength and 0 to 120ksi fluctuating fatigue limit (Sandvik 12r11 218ksi for example)work? I have a tight space constrait to work in, hence my dilemma. Thanks in advance for any advice.

 

[TVP]

For as common as this grade is, and that it is used extensively in spring applications, one would think that there is a wealth of fatigue data available. I have not found that to be the case. The best document that I have seen is the following from Outokumpu:

http://www.outokumpu.com/files/Group/HR/Documents/Fatigueproperties.pdf

 

[EdStainless]

The only thing that I can find offhand is the A-L Blue Sheet with an endurance limit.

http://www.alleghenytechnologies.com/ludlum/pages/products/xq/asp/P.17/qx/product.html#

First ASTM A666 only takes you to 185ksi UTS at full hard.
For this material A-L give an endurace limit of 80ksi.
You are either going to need an actual fatigue curve, and/or a stronger material.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[israelkk]

You can draw the S-N curve yourself as tought in every machine design book such as from Shigley, Norton, Faires etc. Then use Goodman formulation to calculate the life cycle for your case.

 

[strokersix]

Thanks all for the info. I've asked one of our test engineers for assistance. We have extensive experience with iron, steel, weldments, etc. such as commonly found in agricultural machinery. Hopefully we will be able adapt that experience to my design.

 

[TVP]

strokersix,

If this part is statically loaded at ~ 80 ksi, and then periodically dynamically loaded to 160 ksi, then I would be more concerned with force loss/compression set then fatigue life. Probably Sandvik 12R11 would be ok for fatigue life based on several hundred cycles to 160 ksi.

 

[strokersix]

TVP,

Any feel for how much compression set might be expected? Say 20%? More or less? I'm just after a gut feel for now so I can get the design close. Then we'll test it.

I was thinking that a slight yield during initial operation might be advantageous in that it would redistribute stresses and accomodate tolerance stack-ups. Is this a reasonable thought or am I asking for trouble?

 

[israelkk]

Test can be misleading. Unless you know that the piece of strip you test is in the minimum properties allowed by the material spec the test can be too optimistic. Assume that the piece of strip you have has the maximum tensile properties according to the spec and the test results give you 500 cycles before failure and the permanent set is 10%. This doesn't mean that you are off the hook, because future batches of strip material can have inferior properties (and still be within the spec).

There is no substitute for calculations for the minimum properties of the strip. When your calculations for minimum material properties show that it will do the job, then repeat the calculations for the actual material properties and compare it to the tests. If you have a match then you can assume that your calculation for minimum properties are OK.

 

[israelkk]

Just wanted to add that if for minimum material properties you need for example 500 cycles and 10% set, then for better properties your calculation should give larger number of cycles and less permanent set. Therefore, if your actual tested strip (with probably better properties than minimum) should show higher number of cycles and less permanent set and your calculation should predict it. Otherwise, your calculation can not predict what will happen with minimum material properties.

 

[NickE]

I have an old Sandvik spring steel book (1987) and it gives an S-N curve for 12R11.

It also gives the S-N curve for differing tensile strength levels.

(reverse bending stress at 20C Mean Stress=0)

For 0.4mm thick 1500N/mm^2 strength:
Prob  --  Stress Level
50%   --  +/-550N/mm^2 
10%   --  +/-485N/mm^2
0.4mm  /  1700N/mm^2 
50%   --  +/-575N/mm^2
10%   --  +/-525N/mm^2
0.4mm  /  1900N/mm^2
50%   --  +/-595N/mm^2   
10%   --  +/-530N/mm^2

IF you find that your design is robust enough as far as max stress and relaxation you may be able to use a better steel such as 11R51 (17-7PH) or even 7C27Mo2 (MOD420) both of which have higher tensile strengths and higher endurance limits. However they are way way more expensive. One thing to remember when designing with Sandvik grades of steel is that some grades (11R51 & 7C27Mo2) are extra clean for fatigue applications. Also you may be able to use a carbon steel if the environment is dry and clean, or oil immersed.

Sandvik steel along with Hitachi and Uddeholm are some of the better steels for fatigue applications. They seem to have better control on properties adn gauge.

Nick
I love materials science!

 

[NickE]

Oooooppsss.....

11R51 is not similar to 17-7, the sandvik name for that is 9RU10. 11R51 is roughly the same chemistry as 12R11 just with a 0.7 nominal Moly addition. This "...has better tempering properties. The Mo additive also gives it a slightly superior corrosion resistance.

 

[strokersix]

israelk,

Good comments toward consideration of minimum properties. I hadn't considered that issue.

NickE,

Those figures are close to what Sandvik currently posts on their website. I headed directly to 301-type material because we already use it in other (less demanding) applications in the same outdoor agricultural machinery environment. Plated carbon steel could perhaps be an option. This spring will be inserted in a die cast aluminum housing so I do need to consider galvanic action. The other end of the spring will be rubbing against acetal. Does the S_N curve you have go down to 100s of cycles or does it start at 10,000 cycles? Any chance you could post an image or pick off some points for me?

Thanks.

 

[NickE]

You might be able to request the curve from sandvik, I dont think however that you can use this curve for US grade 301. Sandvik has specifically made steel for fatigue service for years.

Only down to 10,000cyc.

Nonetheless here are the starting points:

1500N/mm^2 strength
Prob--Cycles--Stress
50% -- 10^4 -- +/-1025N/mm^2
10% -- 10^4 -- +/-950N/mm^2
1700N/mm^2
50% -- 10^4 -- +/-1075N/mm^2
10% -- 10^4 -- +/-1012N/mm^2
1900N/mm^2
50% -- 10^4 -- +/-1125N/mm^2
10% -- 10^4 -- +/-1075N/mm^2

The curves all have the basic S-N Curve shape, the knee starts around 1x10^5 and the curve goes flat again around 6x10^5.

You will have a glavanic problem if the aluminium and the steel take any baths together. And ambient moisture will cause trouble after time. I tend to believe in the "design for roubustness" theroy, so I would use SS of some sort, or maybe another spring alloy if the corrosion issue is severe. Even these austenitic hard rolled grades will corrode if the environment is right.

Be careful of crevices! also surface condition is going to be important. Some sort of finishing process should be applied after stamping.

(oh and I should have ended the quote in my correction post above, Thats directly from the sandvik book. They do have a heat treatment for 11R51!)

Nick
I love materials science!