Hacker News Reader

Comment by damoncali | Hacker News Reader

Cool, but I wonder about stiffness. When I was an aero engineer, we tended to use aluminum because most of our designs were limited by stiffness, not strength. Since aluminum is lighter than either steel or titanium, you can take advantage of geometry (which greatly impacts stiffness) without sacrificing weight. Strength was only rarely the bottleneck. Interestingly, the stiffness to weight ratio of aluminum, steel and titanium are all approximately equal.

We did occasionally use titanium, but usually some sort of steel was a better choice when strength was the issue. It's just the way it works out. It's also worth noting that for the work I did, cost was never an issue (NASA) - material costs were basically insignificant. We could use whatever we wanted.

View on HN

The problem with aluminium is its resilience, not its strength. You are right that overall structural performance is governed more by geometry than by material strength (resisting moment is m^4 (area times quadratic distance of center of mass), although there are gotchas like lateral torsional buckling, which require bracing) the main problem with aluminium is that cyclical tensile loading leads to material fatigue with a effective material strength of 0%, while other metals can be counted on for at least 50% strength.

This is what he is referring to:

https://en.wikipedia.org/wiki/Fatigue_limit

I'm not sure if you are familiar with it, but I believe that was the core problem with the aluminum-hulled Type 21 frigates of the Royal Navy. The cyclic stress imparted on the hull by wave action resulted in severe hull cracking and a relatively expensive refit to brace the hull. It also contributed to one of the frigates being sunk in the Falklands.

And also (IIRC) the reason why they couldn't fire the ridiculous gun with which they were equipped.

Did the front fall off?

https://www.youtube.com/watch?v=GgrX7uOZqHI

That's a really good point. Interestingly, the Nature Letter (paywall) [0] doesn't seem to mention the aerospace industry at all. It's more targeted toward structural and automotive applications. I'm not sure where The Economist picked up the aerospace narrative.

From the conclusions of the Nature letter: "These findings provide a new alloy-design route to lightweight steels, demonstrating that the combination of specific strength and ductility accessible to steels is greater than previously thought, and increasing the density-compensated tensile damage tolerance of structural metal for terrestrial applications."

[0]: http://www.nature.com/nature/journal/v518/n7537/full/nature1...

Yeah, I don't know much about the auto industry. The basics of stiffness driven designs, though, I suspect would be the same. As would the desire for lower weight. The obvious difference being that when we made a single satellite, material costs were negligible. When you make millions of cars, they're not. That said, there are probably quite a few steel parts in a car that are not stiffness driven, so that could be a big deal if this material works out.

As I understand it, at the macro level auto design is mostly stiffness-driven. But since they use stampings everywhere, strength matters since stronger materials can be made thinner without buckling (hence your roughly tubular section is stiffer for the same weight).

Just as an aside, when I was working at a bike shop, we had a guy come in and was looking for a road bike. After some small talk, he told me he owned a cryogenics lab and said they could freeze anything, including bike parts. Doing so, made them permanently in the range of 20-30% stiffer than the original material. Most guys at the shop thought about frames, cranksets, drivetrains, forks etc. Places on the bike which experience high levels of torsional flex and stress.

I'm wondering how many engineers ever thought about using cryogenics to enhance the stiffness properties of metals to achieve higher levels of performance.

The cost to do it was around $1,500 if I remember correctly. A pretty small fee to get what I'd consider a big gain in performance.

I'm not aware of any cryo process that increases stiffness. Usually, it's more of a machinabily/stability thing they're after. But I've seen many miraculous claims in what I'd call "hobby industries" for cryo treatment. There are guys who freeze gun barrels, for example - with wild benefits that never seem to really bear fruit. I'd be highly skeptical of such claims.

I don't have much experience with cryo treating, but heat treating to gain strength is quite commonly done, especially for metals that are too hard to work in their strongest state, or for metals that lose their heat treat when welded.

Here's a watch industry instance of cold hardening steel: http://www.damaskousa.com/technology-watchcase.asp

These are standard processes for high strength alloys:

http://en.wikipedia.org/wiki/Quenching

http://en.wikipedia.org/wiki/Cryogenic_treatment

I'm just wondering, how well is the influence of frame stiffness on racing performance studied and understood?

I have no personal experience in this matter, but recently I was reading some random stuff on the 'net and I found these guys swearing by flexible steel tubing:

https://janheine.wordpress.com/2011/10/03/science-and-bicycl...

https://janheine.wordpress.com/2014/11/23/what-is-planing/

I don't know what to think about this.

You have way more insight than I do. What you describe reminds me of a complicated webbing of aluminum to hold everything in place. If I understand you correctly having access to a titanium equivalent could do two things. One, you could reduce part count, replace an array of aluminum supports with one titanium part. If the strength and weight are about the same, the simpler design seems better to me. Two, you could free up some space for other stuff. I'm thinking fuel tanks in wings, but i'm not even at the level of armchair aerospace engineer. There are probably many cases where aluminum is lighter but bulkier. Opening up a channel to simplify wiring, or hydraulics could be helpful.

The trouble is that it's not a lot of space, and it's not as simple as stiffness/weight of the material. Think about something simple like a structural tube. The diameter of the tube has significantly more impact on it's stiffness than it's material. So if you use a 30% lighter, but strong enough material, you can add 30% to the cross sectional area of the tube- which means you can make it a bit bigger in diameter, which will have potentially huge impacts on overall structural stiffness. Note that adding 30% more material will add at least 30% more strength to the part (huge oversimplification here, but just for arguments sake, we'll use that simple tensile case).

Titanium was useful when we needed very high strength on a very large part - a good example is the trunnions that attach payloads to the space shuttle cargo bay. These are large metal bars a little over 3 inches in diameter and maybe a couple feet in length. Aluminum is not strong enough. Steel would be punishingly heavy. (do the math - these things are beasts). So titanium can be a good fit for stuff like that.

But most steel parts were small - things like bolts, rod ends an the like. They tend to be specialized with very specific detailed requirements for things like fatigue, frictional characteristics, and that sort of thing. Because they are small, the weight just isn't that big an issue. Bigger stuff was almost always aluminum, which is actually a lot stronger than most folks realize. At least, the good stuff is.

Titanium also has some other odd properties that make it troublesome in space. I wasn't a materials guy, but they were always bitching about "hydrogen embrittlement" and the fact that large titanium parts do not burn up on re-entry like aluminum does, which presents a debris hazard to people on the ground. (I don't know how big a problem that really is, but people didn't want to be dumping space junk on top of people's heads. Honestly, I just took this on faith - people were concerned about it, but it always struck me as odd.)

If I had a nickel for each time I heard a guy bitching about "hydrogen embrittlement", I'd fund this space program myself!

By 'aluminum' do you mean 'duralumin' or just aluminum? My uncle had a few tiles of duralumin in garage, and I remember it was definitely a lot harder than plain aluminum.

Almost everything we used was plain old 6061 aluminum - the "standard" alloy everyone uses for everything. If a little more strength was needed, 7075 would usually do the trick.

Funny that I've started to think of Apple's easily CNC'd aluminium as the standard instead of 6061. That Apple stuff is about as strong as wax, not even as tough as others' plastic shells. Sure looks good on my iPhone and MacBook, though.

I'm glad to have 6061 on my bike instead.

Oh, i think i get it. The materials are different, but not different enough for a whole different design - you can't generally simplify out parts. Some strut being 1cm across vs 1.2 cm across doesn't buy you much.

Thanks!

To better understand the comment above you need to understand stiffness. When you take a wire of steel, which is very strong, you'll find you can actually bend it very easily. If this kind of stress is what the material will face, you could replace the wire by a thin tube which will be much more rigid -- you probably don't want your structure bending and deforming. In this case, where stiffness is more important, it's more important that the material is light so you can make it larger yet proportionally stiffer per unit weight by using an adequate structure -- e.g. a tube or large diameter (this is because the bending stiffness increases more-than-linearly with diameter).

Now suppose instead you were trying to traction the wire (pull it). There's no structural change that will make it stiffer -- in this case all you may want is that it is very strong -- in which case you simply go for the best strength/weight or just pure strength.

> which is very strong

How are you defining strength? Tensile strength?

In this example, yes.

If you went for a compression stress however you'd probably want that same bending stiffness (by making a hollow tube) because rarely the force is perfectly axial.

Just the opposite. 1.2cm vs 1cm might be a world of difference. And if weight is a concern, you might be stuck with a 1cm tube, which might be less stiff than the 1.2 cm tube even if it's made out of a stiffer material.

The idea is that design is material dependent. You can't assume you can just swap out parts for lighter/stronger/stiffer parts and still have everything work.

Amazing and useful write-up to a non-engineer audience such as myself. This is very relevant to my participation in the Local Motors LITECAR challenge (goal: reduce curb weight significantly with developing tech for 10 year outlood) sponsored by ARPA-E. If you have any references for "materials" type people with composite experience, I'd love to have a technical partner with whom to expand on my entry and share the prize money. I can be reached through the contest site or rockstarguitaristforhire at the service of mail by the big G. My apology if this strays too much, I just really enjoyed this response and felt excited about pushing boundaries.

It's been years since I worked in the industry so I don't know anyone, but I did spend a couple years designing composite structures and testing the materials. The basics of composites are learned in undergrad mechanical engineering, and you can generally apply them in real life. It's not that complicated. Where composites get tricky is in the details - de-lamination, manufacturing nits, that sort of thing. I would look at engineering schools - someone in a masters program with an interest in composites would get you quite a bit of useful knowledge. Structural composites get tricky, though. Things that are easy with metal (for example, bolting parts together) get hard with composites, and you really have to think hard or you give up all the benefits of composites when you have to engineer around their limitations.

Reply on news.ycombinator.com