MMPDS material strength vs thickness for stock mtl

[sigma38]

If a part is made from 1 ½” stock material does it fall within the [.250-1.500] column or the [1.501-2.000] column of MMPDS? And no, you can’t answer “the more conservative one”

The stock material is ink marked "1.500" yet almost always comes in slightly oversize (and hence could fall into either column).

Is there a definitive (referenced) answer?

Thanks.

 

[jetmaker]

Depends...

If the raw stock is called out in the drawing as 1.500" or less, then you use the 0.250-1.500" data. If the final part is 1.500" thick, then it must be made from a larger stock as machining must remove a little from the stock to get a good part.

Now if the stock size is oversized, then you are not using 1.500" stock, regardless of what it is part marked. If the stock does not meet the specs called out in the drawings, then it needs to be rejected unless the analysis supports the larger thickness values.

Another possibility is to use the "more conservative one".

jetmaker

 

[sigma38]

Actually, I'm talking about going the other way around.... if I know the part was made from 1 1/2" stock then which column does it fall into? IOW when the data for MMPDS was created did they assume that 1 1/2" stock (which is marked 1.500 but comes in a little larger) is 1.500 or 1.501+?

The machinists are telling me that the stock almost always comes in a bit oversized even though it's ink stamped "1.500".

Make sense?

 

[jetmaker]

sigma38,

I understand were you are comming from here. However, I can not comment on how the tests were done in MMPDS; however, the issue here is that if your stock is over 1.500", it falls into the higher category. It would be one thing if the nominal thickness were <=1.500" with local thickness exceeding 1.500" due to tolerance issues.

A materials guy might answer this better, but what I was told was that as the plate thickness increases, the rolled properties in the middle are different than those nearer the surface (might actually be the other way around). So in theory even 0.001" thickness difference can result in material property changes. I reality, the difference in material properties between these 2 thicknesses is likely negligible, but you have to create a range limit somewhere. So, without the raw test data for the MMPDS, you will not be able to tell if they actually tested 1.500" thick stock or 1.50+" thick stock.

I think that a DER might have issues if it were discovered that you made parts from thicker stock than accounted for in the analysis. That is why it is often best to specify material thickness below the cutoff point if possible to ensure you are working within the range.

jetmaker

 

[jetmaker]

Suggest you look at Section 9 of the MMPDS and look up the specified ASTM standards to see how they are tested.

 

[der8110]

Here is a tie breaker: use the actual mill certification from the lot (or "melt") of material involved. You may be surprised how much it exceeds MMPDS "guaranteed minimums". If it is just a onesy-twosy quantity, you can shed a little weight by using the actual material strength (14CFR25.613(e)).

Nominal vs. actual thickness is mostly academic; the mill was shooting for the properties in the nominal thickness. Many mill certification forms will indicate the minimum specification values, which steer you towards which thickness category in the MMPDS tables. This would be regardless of actual thickness.

Also look in the AMS-QQ specifications for guidance on thickness tolerances.

 

[sigma38]

I sure hope it exceeds "guaranteed minimums", at least 99% of the time with a 95% confidence (for A-basis)

I really can't use actual as-tested material properties or S-basis since I need to compare to what was tested and then specify material condition accordingly on the print. IOW I don't want to specify a strength requirement, only a thickness requirement.

Section 9 is spot on and where the question really derived. For instance, MMPDS says in 9.2.1, "The product used for the determination of minimum design values for incorporation into MMPDS must be production material." yet it also says that they base their thickness on actual heat thickness (but using a sample taken from the middle of the material). For example, in 3.1.2.1.1 it says, "All of these properties are representative of the regions from which production quality control specimens are taken..." which applies to aluminum in this section.

Maybe I'm just wrapping myself around an axle on this one and need to take the strict "1.500" written on the material literally. I would like to see how they tested though and might have to look up the ASTM standards.

Thanks for the discussion.

 

[jetmaker]

der8110,

Using the mill "guaranteed minimums" would qualify only as "S-basis" data and would need to be bought off by a DER. Is this not correct? Furthermore, would it not require specific lot testing if it were to be used on a continuous basis?


sigma38,

You probably are making a bigger issue of this as I'm sure most stock thicknesses are not overly scrutinized before being machined... if it close enough to 1.500", then it is 1.500". Even more so, I doubt that most machining shops let the engineer know if the stock slightly exceeds callout. However, I still see this as a violation of my specs. If I callout "material thickness not to exceed 1.500", then that's what I expect. The issue now lies in where you draw the line as to what is too thick/thin to be considered 1.500" stock.

Have you had a chance to look the ASTM standards up? They might say that you are to test samples in widths of 1", 2", 3", etc. That might be how the lines are drawing for estabilishing ranges.

jetmaker

 

[rb1957]

personally i think it's little short of nonsense to quibble that a 1.501" thick slab is different to a 1.499" thick slab. I guess when "they" started MIL HDBK 5 they wanted thin plate and thick plate, so they drew a line at 1.5".

And in anycase, how different are the structural properties between < 1.5" and > 1.5" thick plates; i doubt it's more than a couple %age points. and why not use the more conservative numbers (as a default) ?

personally too, i'd never use the actual properties of the material, that's something i'd keep in my pocket to cover say a 5% overload.

 

[der8110]

By "guaranteed minimums" I was referring to A-basis. There are manufacturers who use terminology which sounds like that in their promotional literature, problem being they rarely specify where it came from.

S-basis and the like (Matweb, Metals Handbook, metallurgy textbooks) are only good for low risk/low hazard applications, or as a forecast until a proper test program can be run.

Yes, specific lot testing would be required to use "premium selection" on a continuous basis. Many companies have internal specs which already do this. Most modern quality systems require mill certifications anyway; so the actual mechanical properties (and chemistry) of that batch of metal are at hand (not sure how you'd handle this for an overload, unless it's the mfr). The trick is to make sure the test methods are consistent, and they usually are.

Another way to look at it is the mill certification is a sample size of one, part of many data points that long ago went into the number cruncher to come up with MMPDS A- and B-basis values. Your only uncertainty is the calibration of the mill's lab, and that is a minor concern.

I do lots of one-offs and rapid jobs, and actual material properties have saved the day. For instance, I have seen a mill cert of 7075 with Ftu of 109 ksi AND elongation of 12%; coulda swore it was steel. It helped out when there was a small negative margin (in comparison to MIL-HDBK-5), after repair was in place. With the actual lot strengths in hand, voila, problem solved with a change order to specify a minimum strength.