Tensile Test Variability and Load-Time Curve Discontinuity 2

[Dyskolos]

Hello,

I am reviewing testing certs from a potential new vendor for metallurgical testing, and have identified a potential issue. The material is 4130, approximately 21 HRC. I have two questions on which I would very much appreciate any assistance y'all would be willing to give:

1) The yield strength is somewhat higher (10%+) than it ought to be based on our previous tests at other labs from this same bar, though the ultimate tensile strength, elongation, and other properties are right in line. How much variability is normally to be expected in a tensile test, and is it unusual for a single property to vary so while the others remain constant?

2) The Load-Time curve was provided, and appears to have a discontinuity prior to the yield point (see attached photo). I've not seen this before, and I'm concerned that it might indicate a malfunction of the testing machine, and might have contributed to the issue from question 1. Is my intuition correct, or is this a non-issue?

Thank you!

 

[EdStainless]

It looks like slippage, either of the sample in the grips or of the extensometer.
I wouldn't have sent this curve to a customer, though I would use it internally.
Since the Yield strength is not a true physical value I would be well satisfied with 10%.
Do you know what strain rate was used?
Do you know the calculated Modulus?
We use a fixed stain rate in our lab, and it is required for outside labs to use the same values.
And by reporting the Modulus you know if there has been a good fit to the curve, after the value is 29.5 Mpsi. If it is off by much then the fit is bad.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[Dyskolos]

Ed,

Thanks for the information, that helps a great deal. I can see I need to spend some quality time with ASTM E8 to better specify our procedures. I unfortunately do not have the strain rate or the modulus on the provided cert, but I can certainly ask for them. Thanks again!

 

[EdStainless]

The new version of E8 has removed a default strain rate.
The aeropsace specs are adding it back into theirs as many alloys are very SR sensitive.
Though in many cases is the Yield and elong that are effected.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[MagBen]

The discontinuity in the curve seems to me a slippage too, but I do not expect that will cause yield increase.
The hardness level indicates this is annealed material, the strain rate before yield can make a different. normally the higher the rate the higher strength. note there are two strain rates, before yield and after yield from yield to breaking. the latter is much faster than the former.
another thing i seldom see is the curve from 0 to yield is all not straight! so there is not a portion on the curve to be used to calculate E modulus. How did strain-rate curve look like?

 

[EdStainless]

The reason that I asked about the modulus is that most software takes two points and bases the modulus and offset for yield from those. With this curve shape they could have ended up with some very odd values for E. And in turn this could skew the Yield value.
We usually ignore the first 10% or so, and then try to pick a second point that is about 50% of expected yield. If the calculated E is out of range (I think that I used 26-31 Mpsi) the technician needs to select new points manually.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[MagBen]

We are doing a similar thing to determine yield: offset .2% from E straight line. This is load-time curve, i think there exists a straight portion at the beginning if it is converted to stress-strain curve. it is hard to imagine the slippage occurs from the beginning to end, such that it seems no elastic deformation for the alloy.

 

[Dyskolos]

Thanks for more good info! I'm reaching our for more info from the vendor, but they've not been particularly responsive so far, so it could be a day or two.

MagBen,
I know I would expect a straight line in the elastic zone of either stress-strain or load-elongation diagrams, but is this true in a load-time curve? If a constant strain rate is used over that portion of the test I suppose it would be. Regardless, to convert from load-time to stress-strain or load-elongation, I'll need to know the strain rate that was used, correct? Or am I looking at this wrong?

 

[EdStainless]

I don't know why someone would ever generate a L-t curve, useless.
In reality most curves are load vs crosshead displacement, or sometimes they use the extensometer output.


= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[Compositepro]

Stress/strain curves that are gathered without using an extensometer are actually load/time curves, which presume constant crosshead speed.

 

[EdStainless]

I have used hydraulic machines that don't have fixed speed (not closed loop control) but still charted using the actual crosshead position for the x-axis (from a linear encoder), so you do need to ask what the data really is.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[Dyskolos]

Got some answers, but they don't make much sense. I'm setting up a call with the vendor to clear some of this up; it appears e-mail is insufficient.

They stated that they were unable to provide a stress-strain curve because the machine only provides L-T. The machine is listed on the cert. as model WAW-1000E, if anyone's interested.

Strain rate provided was .001/s. From what I can tell in E8, this is quite low, correct?

Modulus was given as 24.92 MPa. I have asked for clarification in the hope that this is a typo.

 

[btrueblood]

Googling that machine shows that it comes standard with a crosshead displacement sensor (linear cable pot), and also shows two (mechanical and laser) extensometer attachments. When originally looking at the time history plot, I thought the hiccup in the curve was a pause to remove the extensometer prior to proceeding to fracture of the sample. But there are other weird things happening in the plot, with the nonlinear area at the start being one (likely a bent sample, or the crosshead control servo valve is wonky...)

Find a different lab. These guys don't know how to do tensile testing, and don't appear to even know their own machine, or to have read the manual. Link to the manual below:

http://www.aoblab.com/dosyalar/waw-1000e-111.pdf

 

[Dyskolos]

We've decided to look for an alternate supplier, and I'll be tightening up our testing requirements for future qualifications. Thanks all for the info and assistance!