# Aluminum thickness tolerance and structural calculations

### Help Support HomeBuiltAirplanes.com:

##### Well-Known Member
When calculating strength/stiffness, is it conventional to assume that all sheet metal is at the worst case thickness given the sheet tolerances (so 0.025" aluminum is treated as 0.0225", assuming 48" sheet width)? Or is there some fudge factor to account for most of the sheet not being at the absolute worst case? Does this then get pessimistically combined with worst case material properties (295 MPa tensile, 245 MPa yield for 6061-T6, vs. 315 MPa / 280 MPa nominal)?

##### Well-Known Member
I'd say yes. Where it matters mostly (buckling), it's only at the most critical position (the midpoint) where it really matters what the skin thickness is and that's what matters.

I take the minimal values too. Testing parts, you're usually very close to the nominal values. Problem is that once a single part starts to yield the rest of the structure is affected (weakest link in the chain)

#### Matt G.

##### Well-Known Member
In industry, B-basis allowables are used for redundant structure, i.e. skins, frames, stringers, ribs, etc. and A-basis allowables are used for anything that does not have a secondary load path and that would affect safety of flight if it were to fail.

At work I use nominal values for calculations, but I suspect the tolerance may be built into the allowables somehow. Using A-basis allowables for everything and using the minimum thickness is probably excessively conservative. I guess it depends on how much conservatism you want. It's built into the allowables and analysis methods, and you can always add more conservative assumptions along the way. Too conservative and you'll be carrying around some extra weight though. Up to you to find the balance that works for you.

If you are working with clad aluminum, don't forget to subtract the cladding thickness from any buckling calculations you do.

##### Well-Known Member
Matt, can you define those terms (B-basis and A-basis)?

• A-Basis: At least 99 % of the population of values is expected to equal or exceed the A-basis mechanical property allowable with a confidence of 95 %
• B-Basis: At least 90 % of the population of values is expected to equal or exceed the B-basis mechanical property allowable with a confidence of 95 %

Last edited:

#### DeepStall

##### Well-Known Member
+1 Matt G. Beware of cascading safety factors, it's an easy way to feel good about a design, but an equally easy way to get heavy too. 10% knockdown on thickness -> 10% higher stress + 10% knockdown on allowables and now your design might be upgauged and 20% heavier, or using a different alloy or material that may be more expensive or harder to work with. That's the art in aircraft structures -- finding the lightest solution that will work great 99.99% of the time, and get you home that one time you actually encounter true design limit loads. Risk vs. reward -- maybe upgauging only costs you a pound and lets you greatly simplify the design and sleep well at night.

I'd be asking myself: How confident am I in the loads and loadpaths for this structure? Is that already reflected in my design factor of safety? What are the consequences if this structure ends up with a 10% lower than designed factor of safety? If the part yields, is the airplane still flyable? If it buckles? If it fails outright? OR, I'll design to a little thinner than spec, and satisfy myself that the actual as-delivered sheet meets or exceeds this thickness before I start fabrication. OR, I'll design to spec, then proof test the actual design and see if passes. OR, I'll design to spec, then monitor the part in-service for cracking etc that potentially might result in some painful rework, or it might never actually be a problem.

##### Well-Known Member
Follow-up: Is the aluminum that I'd buy through (for example) Aircraft Spruce or Online Metals commercial tolerances? Looking at Alcoa's price lists, if I'm reading them correctly half-commercial tolerances are ridiculously cheap. By those price lists, 0.024" half-commercial tolerance sheet would be a drop-in replacement for 0.025" commercial tolerance (assuming designing to minimum thickness), cheaper per sheet (the 4% reduction in weight more than offsets the cost of the higher tolerances!), and lighter (4% is around half of the gains that can be had from going to exotic aluminum-lithium alloys!). Given this, why are thin sheet tolerances so high (or, alternately, why do buyers whose customers presumedly care about this stuff, like Aircraft Spruce, not get tighter tolerances on their materials)?

#### Matt G.

##### Well-Known Member
I would assume the AMS spec is different for the alternate material? If it is, do you still have allowables?

##### Well-Known Member
I would assume if someone ordered a tight-tolerance load of 6061-T6 from Alcoa, it would fall under the same standards as the default tolerances, and the same numbers from the MMPDS handbook would hold, no? (Real question, I have no idea how this works.)

#### TFF

##### Well-Known Member
Because it is priced per pound per shape. The industry wants the aluminum on the thin side. You order a 025 2x4 piece and that is what you think. They wantit thin so they can buy more but they sell it at the thicker end of the spec.

#### Matt G.

##### Well-Known Member
When you buy aluminum, it'll have, for example, "6061-T6" and something like "AMS-QQ-xxx/xx" printed on it, with the x's being numbers. If you look at a page out of MMPDS, the top of the allowables table will say "specification" which is where the AMS spec number is listed for that material. Sometimes there are multiple specifications for a particular alloy and temper that have slightly (or not so slightly) different properties. Usually it's just for multiple product forms of the same alloy, i.e. 2024-T42 extrusion vs. 2024-T42 sheet, but sometimes there are multiple specs for the same alloy, temper, and product form.

Example: some of my employer's products use machined fittings made from 7050 plate. There are two specs for the same alloy, temper, and product form, yet one is several times more costly than the other, and has slightly better properties.

I'm not sure how often this situation occurs, but I figured it might be worth mentioning. All in all, if the AMS spec number on the material can be found in MMPDS, you're set.