Hacker News Reader

Comment by dogleash | Hacker News Reader

"Not at all, because carbon fiber is better compression then tension. And that's what nobody understands. It's completely opposite of what everyone else says. Everyone's, oh, carbon fiber can't handle compression. They're full of shit, and I've proven they're full of shit."

- Man crushed by custom carbon fiber submersible

View on HN

"Now, if it fails, then you have to stop, and it's -- again, this is not something that just happens all of a sudden. It doesn't just implode. It screams like a mother before it implodes."

- Man crushed by an instantly imploding submersible

Management classes will write textbooks about Stockton Rush, but not in the way he would have hoped

Why? There’s already a plentiful list of disastrously wrong arrogant cocksure assholes. Still doesn’t stop people from thinking they’re the Steve Jobs flavor of arrogant cocksure asshole.

Steve Jobs, who very famously died of pancreatic cancer because he delayed surgical treatment in favor of fruit juice? That Steve Jobs?

the disastrously wrong arrogant cocksure assholes don't hear that part, they just hear "people told him he was wrong" and "millions of dollars"

Right next to the O-rings will survive in temperatures colder than they were designed. At least Rush's backside cashed the check his mouth wrote where as the NASA engineers were still around after their decision. Does that make one better/worse than the other? Rush took other people with him though when he cashed that check.

For the record, NASA engineers recommended against launching, but were overridden by, as usual, management.

And again when Boeing's management overridden engineers in decision that Max 8 was safe to fly and new flight stabulity electronics were good enough.

For the record, Rush was management as well

To be fair, it did scream before it imploded.

They just disregarded the screams because they couldn't afford to make sure it was still working.

The funny thing is that it's trivially untrue of carbon fiber, which has a compressive strength ~40% of its tensile strength.

Being weaker in compression than in tension is not the same as being unable to handle compression. 40% of carbon fiber's tensile strength is still comparable to the compressive strength of high strength steel.

> Not at all, because carbon fiber is better compression then tension.

That's still false though.

So there's a NASA/Lockheed report on this.

> https://ntrs.nasa.gov/citations/19810045017

As 95% of the carbon fiber market is T300, and as T300 hasn't changed in the years since, I think it holds up.

Long story short, raw carbon fiber has a high compressive strength, but it's still just ~40% of its tensile strength. Carbon fiber composite single-plies or tapes have a reduced, but still reasonably high, compressive strength. When you get to multi-ply 0/90° or 0/45/90° composite laminates, "allowable" A-Basis (<1% likelihood of failure) compressive strength is very low -- around 400MPa -- well under the compressive strength of good steel.

And that's under ideal conditions. You've got to apply knock-downs for ply angle, moisture, through-thickness shear, impact damage, and manufacturing defects.

...Hence the Titan sub failed at a compressive load of something like 200MPa.

I'm not accusing you of this, but one of the issues with carbon fiber is that people see on a specsheet that it has, e.g., a "tensile strength = 4000MPa" and they assume that composite parts will exhibit a practical tensile strength of 4000MPa. The properties of CFRP -- CF composites with epoxy or PEEK or whatever -- are always way reduced, and sometimes quite difficult to pin down.

That said, the tensile strength of CF is always high, as a rule of thumb. Its compressive strength really isn't.

I would not call 400 MPa "very low" that's better than the best structural steel. Further, the important thing is the specific strength. CFRP has a density 5 times lower than steel, meaning you can use 5 times as much for the same weight, bearing 5 times the load. A good steel has a compressive strength around 1000 MPa, the best superalloys have compressive strength around 1500 MPa, the same weight of CFRP can withstand the equivalent of around 2000 MPa.

Sure this is idealized, but so is the strength of steel.

The concern with carbon fiber is its potential for delamination, not its compressive strength. Titan's failure was after delamination.

400 MPa is not better than the best structural steel. And 1000 MPa is way higher than the best structural steel (compressive strength and tensile strength are essentially equal for steel). Most structural steel has an ultimate tensile strength of around 500 MPa. Ultimate tensile strength is the comparable strength parameter when discussing rupture/fracture.

400 MPa is better than the best structural steel. 1000 MPa is not for structural steel, it is for high grade steel, the sort you make submarines out of. Why would I list 400 and 1000 for the same value?

Steel is weaker in compression than tension. It's more isotropic than say concrete, but the difference is meaningful in practice.

We're not discussing rupture here. That's for when the pressure is higher internally than externally. We are discussing a submarine, which is a pressure vessel under compression which must also remain buoyant. The specific yield compressive strength is the value which matters.

When someone says "structural steel" they are normally talking about something similar to A572. A572 comes in multiple grades, with grade 42 being the lowest. Grade 42 has a yield strength of 42 ksi (hence grade 42) and a rupture strength of 60 ksi by spec. 60 ksi is 414 MPa. Even A36 (which is basically the weakest structural steel commercially available nowadays) has a rupture strength of around 60 ksi and a yield strength of 36 ksi. Hence, even the weakest "structural steels" have a rupture strength of around 400 MPa.

When I use the word "rupture" I am talking about the material property, not the specific submarine loading condition at play. When comparing steel, which is a ductile material, to carbon fiber, which is a brittle material, you should use the steel's rupture strength instead of yield strength. Steel is for all intents and purposes an isotropic material, and the difference between tensile strength and compressive strength is not material in practice (because steel in compression is nearly always governed by macro-scale geometric issues leading to buckling rather than the material strength in compression being exceeded).

The '~' is important: composites are in general very difficult to predict, and under compression even more so.

(To build some intuition: carbon fibre, is just that, i.e. a fabric. No-one expects to be able to push on the edges of a sheet of fabric get any real resistance at all. If you then embed that it a plastic, i.e. make a composite, the plastic is mainly what makes it hold its shape. It's only really made stronger in tension by the fibers in it, why would you expect the compression to be better?)

"what are you gonna do, stab me"

- Man crushed by custom carbon fiber submersible

“Now, you can Google this on Google”

Reply on news.ycombinator.com