Cable tension systems are very tricky to engineer, and Arecibo was built quite a while ago. If you use fixed, non-adjustable cables, you still need to tension them again after installation to allow for settling to balance tension on all the wires. I don't know much about Arecibo, but having strain gauges on each wire, some can be built into the assembly, would allow monitoring and possible automatic tensioning systems to balance the loading. Even in a system of many cables, the load shifts with wind and geological shifting of foundations, and sometimes for brief periods a few of the many cables take the majority of the load leading to differences in strain or stress.
It's cool that something good can still come out of this "disaster" - it's a shame that it wasn't kept up, but if this sort of forensic analysis reveals ways to improve cable manufacturing or maintenance routine improvements, that feels like a pretty good win
It is well known that zinc, like all metals with low melting temperature, flows slowly under high stress.
So this failure was easily predictable. I assume that the cables have been designed for a much shorter lifetime, so it was expected that either the cables would be replaced or the radiotelescope would be decommissioned many years ago.
Nevertheless, the choice of pure zinc for the cable sockets is somewhat weird, because it guarantees a short lifetime. Had a zinc alloy been used, like ZAMAK (Zn-Al alloy), the flowing speed would have been much less and the lifetime of the cables would have been greater.
Alloys are much more resistant to plastic deformation and flowing than pure metals, because the atoms with different sizes cause defects in the crystal structure that prevent the easy slipping of the atom planes over each other.
They show the zinc doesn't continue to creep if the load is below a threshold-- technically the zinc doesn't creep if it's not subject to persistent shear forces, which is won't be if the cable load is below a threshold as the wires take up the load. This essentially results in a long term capacity which is different from the short term capacity. Unfortunately the ratio between the two depends on the wire splaying geometry which isn't well controlled, resulting in wide differences.
The report concludes that if it had been built to a safety factor of 3 rather than 2, the issues wouldn't have been experienced. Alternatively, the failure could have been avoided by reacting to flow over some threshold (which was noticed well in advance, but not reacted to)
Presumably we don't see similar failures in sockets in suspension bridges because they're built at a safety factor of 5+ and usually that's a SF over their rarely reached maximum load rather than a factor over their 24/7 load.
Any handbook about the strength of materials has a chapter about creep a.k.a. cold flow.
The handbooks from immediately after WW2 already included such a chapter, but I believe that the first studies of this problem must be much older.
When any metallic structure is designed, it must be verified that it will not fail in any of the possible modes, including due to flow over the intended lifetime.
By the approximate rule mentioned there, zinc begins to have non-negligible creep already above minus thirty Celsius degrees, so at normal ambient temperatures you must always compute the creep of zinc for any structural design.
On the other hand, metals like iron or copper have negligible creep at room temperature, even when pure.
Aluminum and magnesium begin to have non-negligible creep at temperatures only a little above normal ambient temperatures.
Hard alloys can have much lower flowing speeds than the metals included in their composition.
In integrated circuits, the metal connections are affected by electromigration, which is the flowing of the metal due to electrical current instead of mechanical stress.
The electromigration properties and creep properties of a metal are closely related. In the beginning, the ICs used pure aluminum for interconnections, but when their size was reduced, the connections began to fail after a too short lifetime.
The first solution for this problem was the replacement of pure aluminum with harder aluminum alloys, including small quantities of copper and/or silicon.
When the ICs became even smaller, the aluminum alloys had to be replaced with a metal having a higher melting temperature, i.e. copper, which fortunately also has a lower resistivity.
Sometimes it’s worth pausing a moment when seeing a “It’s well known that…” to check the timeline because the construction of Arecibo might well have taken place (or planned) before it was known (let alone well known).
Edit: a bit more searching suggests that this was studied in 1947 (Andrade’s Creep Law and the Flow of Zinc Crystals, by AH Cottrell)
This feels like the facility was in operation for so long, that the people who knew about this potential problem all retired and with them went that knowledge.
Andrade is the namesake of Andrade creep due to his work in the early 1900s, but the existence of creep had been known for a long time by then. I'd imagine smiths have been aware of creep throughout the history of metallurgy.
TIL that creep can even happen at room temperature (albeit Caribbean room temperature) if the stresses are great enough and the time span long enough.
I always thought that it was more a thing for objects like jet engine turbine blades, that get very hot as well as having very great forces acting on them.
Transport properties all "happen" at any temperature. The rate is exponential with T, though, so at some point you dismiss it. Kinetics usually (always? Who remembers? ) is continuous with T.
Phase changes, on the other hand can happen abruptly wrt T (first order, melting/boiling ice), or gradually (second order, curie transition, boiling mixtures).
Of course, first order transitions require a pure substance that doesn't actually exist since the chemical potential is infinite at infinite dilution. But that's entirely academic.
EDIT: I forgot to mention that creep has many proposed mechanisms and their relative contribution is not fully understood; however most depend on a transport phenomena, typically diffusion.
Diffusion of atoms through solid materials is one of those things I intellectually know happens, but still don't really grok. Sintering is one of the weirdest things.
Why does squishing things together AND warming them up cause the particles to stick together more than when you just squish them together or just warming them up?
> Why does squishing things together AND warming them up cause the particles to stick together more than when you just squish them together or just warming them up?
>> The equations describing these laws are special cases of the ideal gas law, PV = nRT, where P is the pressure of the gas, V is its volume, n is the number of moles of the gas, T is its kelvin temperature, and R is the ideal (universal) gas constant.
Did you miss the part where I was talking about sintering and diffusion of atoms in materials in their solid phase? A quick search in the article you linked shows zero results for either "sintering", "diffusion" or related terms.
The ideal gas law is very nice but hardly applicable to describe things like creep in metals under stress, let alone things like the grain boundary diffusion mechanism for sintering metal particles together.
What is a solid but a, relatively speaking, very, very cold gas?
Put two solids together with no intervening medium, and heat everything up, and you'll see some mixin' and mingling. Often, the reason you don't see more of that sort of thing is because everything is immersed in an oxygen rich fluid with a tendency to create large, relatively reluctant to mix and mingle oxide coatings before anything fun can happen. Hence why another word for cold welding is vacuum welding.
Materials, when you really look at em' are not as 'solid' and 'stable' as you may think. Hence why I have a mechanics of materials book I peruse from time to time.
Its not hard to understand once you see the mechanism in full (vacancies, interstials, etc) but its hard to write as txt.
Suffice to say, there's plenty of room to squeeze.
"Squeezing"
When you squeeze you get rid of air bubbles, right? Well, when you squeeze you get rid of boundaries which are really volumes at small scales.
Well nature hates surfaces, so if you squeeze surfaces close enough together and hot enough you'll give the material an opportunity to rearrange atoms to get rid of boundaries.
"Ostwald ripening"
If you thought sintering and diffusion are hard, I had a respected mechanical engineering professor working on a refractory materials project claim, when he first encountered it, that Ostwald ripening cannot occur.
Squishing is important, else the forces involved are negligible. Warming them up increases their movement and thus the likelihood of a number of things (including random interactions with the thing they're squished against).
Failures due to creep deformation are pretty common in solder joints - driven by residual stresses created during assembly which persist over a long time, and eventually lead to cracks forming (stress relief - but not so good for the connection).
Usually not an issue at room temperature, but it can happen - solder has a low melting point.
In the case of the observatory, the failure happened in zinc components, which has a fairly low melting point.
Turbine blades are just one of the most extreme examples, it's really hard to make a metal that resists creep in that environment.
Trying to think about the forces on the ends of the cables; when the wind is blowing and the big weight they support is bucking and the resonances of all these waves are interacting among all those cables... I kinda suspect there were instantaneous loads at times that were several multiples of the static weight. kinda expect that would count for more. Not that I know anything about it :)
It stood for as long as it did, and had more weight added than the original design called for: thats a pretty good win for the original design, i think.
The metal technology is outdated for this kind of structure
This should be rebuilt with far lighter weight carbon fiber composite structures for the suspended equipment and far higher-strength and lighter weight carbon fiber rope/cable.
Both technologies are very well developed and proven. Structural carbon fiber is used in everything from aircraft [2] to buildings to bridges [1], and it is only carbon fiber ropes that allow elevators to work in new kilometer-tall buildings [0].
Merely reducing the weight alone reduces the stresses involved in supporting the required equipment, and even greater benefit and safety margin is gained with the higher strength of engineered composite materials.
Time to rebuild with the next generation of technology.
I really doubt that's a good idea. The actual carbon fibers (not the plastic they are encased in) are very brittle. You could try using a combination of UHMWPE and Kevlar, but steel is probably still better. Blow dry air through the cable conduits to prevent corrosion.
The most common carbon fiber composite is carbon fiber reinforcing an epoxy matrix. The carbon fiber, as individual fibers 5-10 microns in diameter (vs human hairs ~20-180 μm), and as tows or yarns of 3000-24,000 fibers are actually very flexible. I can take fibers or yarns and fold and crush them back on themselves without breaking (of course, other high-performance fibers such as Kevlar or Dyneema, can take even more such abuse).
It is the epoxy matrix that is brittle, and when it finally breaks under load, the load is concentrated at the fracture and breaks the fibers.
This is not an issue in high-performance fiber ropes, as they do not use an epoxy matrix. They have much higher fiber ratio (so much less matrix) and the matrix is of course highly flexible, e.g., some urethane formulation, and often a wear-protective jacket; the matrix is basically a binder to keep the rope/cable together. This is why composite fiber cables have replaced steel in high-performance yachting, tall elevators, racecar wheel tethers, and many other applications. The entire point is that they are much MORE resistant to breaking and wear than steel (and in some applications will work in long-length when the steel cables would literally break under their own weight).
That said, while I haven't done a full analysis, it is possible that Kevlar or Dyneema or even Zylon as the main cable reinforcement, sacrificing some strength-weight for increased toughness, may be better in that particular application. But any of the high-performance fibers will outclass steel.
Similarly, for the structure housing the suspended instruments, composites are a win on every parameter. First, reducing the suspended weight pays benefits across the whole system — lower weight to suspend reduces the requirements of the cables, which reduces their own weight, which reduces the loads on and so the weight of the towers, etc. This alone may overcome any initial cost differences. Moreover, lower weight means lower momentum generated fighting high winds, etc., composites are better vibration-damped, and more. Plus, composites don't corrode the way metals do. And, this is not a structure where we need to worry about impacts and brittleness (which can also be managed by reinforcement and matrix selection).
Doubt all you like, but composites would be a huge improvement in this application (and don't need active blow-drying, which seems unlikely to have been able to prevent the creep failure).
Source: I founded & run a high-performance composites design & mfg firm, with 20 years of direct hands-on experience designing & fabricating advanced composite components & products for motorsport, defense, aerospace, subsea, sport, & architectural applications.
And yes, some of your idea that carbon fiber is brittle may come from the fact that it has very low elongation and elongation at break, but this means that when it finally breaks, it will be a sharp event (e.g., compared to steel which starts by deforming), but that break is after far higher loads.
The main dish is fixed, they move(d) the receiver around to steer it, with a paraboloidal dish that would not work (because the focal point would be fixed).
Yes. Initially the receivers had very long linear feeds, because a spherical mirror focuses to a line, not a point. Later additional mirrors were added to partially correct for spherical aberration.
Metallurgy is super interesting. The fact that metals - these crystals that would be super brittle and stiff except for a few dislocations - can creep like this is amazing.
My (slightly) tongue-in-cheek proposal for Arecibo: drones. Instead of trying to hold one big antenna up with huge cables, you use an array of antennas flying above the dish on drones. This allows for low cost, easy beam steering (within the limits of the shape of the dish), adaptation for different frequencies, quick stowing when weather arrives, etc. Even station-keeping to the required accuracy doesn't seem that hard (you might have to do some active phase correction). Disadvantages are that you'd need to digitize on each drone (which might bust the whole scheme, SNR-wise), and that drones are quite electrically noisy.
Anyone that remotely touches any ODEs relating to this line of work just had a heart attack and fainted.
We don't need a roaring 20's prohibition-era jiggly jazz fest on our energy packets, this sounds absolutely nightmarish for anyone below a management position having to work with it. :(|)
Interesting - you might be on to something. Hmm, maybe it's not quite drones (because the limitations you mentioned like electronic noise, etc) maybe it's some other autonomous swarming unit. Maybe they are drones, but you use some kind of floatation like helium or hydrogen to reduce power needed and therefore noise. Maybe you physically connect each of them in some way in a compression/tension webbing that can be adjusted to control swam unit positions instead of using drone propellers. Fun idea :)
DJIs biggest drone is over 1m wide and can carry a 6 kg payload.
If you could make a suitable receiver that weighed 6 kg or even 600kg, you could just suspend it from motorized plastic cables on space frame towers. Drones aren't going to make it cheaper or easier.
Oxen are herbivores which need to eat a lot more to provide the necessary nutrition and that would translate to greater volume. OTOH chickens are smaller animals with greater surface area to mass and might require more energy to maintain body temperature.
Does it use ‘Bond villain’ rather than implicitly mention Sean Bean in order to avoid spoiling the demise of a character in a movie from almost three decades ago?
Fair assumption, although Goldeneye is easily one of the best Bond movies, if not the best. Personally I remember more about that movie than I do about the whole rest of the franchise combined.
Goldeneye is among the best, but for my money I think Timothy Dalton gave us the best Bond performances we're likely to ever get. Brosnan did it well, but Timothy Dalton just oozed cool killer vibes.
I remembered it was a villain, I didn't remember it was played by Sean Bean so, more effective headline for me personally. I'm not sure why you honed in on that though.
Rumor has it he might have ultimately been defeated by brushed steel Parker pen. Rumor also has it that those pens became sort of a must have in the late 90s. Maybe those incidents are related, we will know more when his Majesty's Secret Service declassifies the files in 100 years or so.
Or we ask Pierce Brosnan, I have the feeling he was involved in this somehow...
hahaha Not by far... I used to work with an Erlang programmer (probably not really the reason he was so eccentric), and this guy was by far so full of oddity that was almost impossible for anyone else in the office to tolerate him. A brilliant programmer, but utterly incapable to having other humans around.
> ...the long-term strength of a socket depends on how the cable wires are splayed out in the zinc during socketing.
> ...eventual cable failures would not have occurred if the cable system had been designed with a safety factor of at least 3.0...
Reaction 1: Sounds nice...but having a much-larger safety factor in something the size of the Arecibo telescope costs a lot of money. And there really is no safety factor large enough to for-sure protect you from shoddy workmanship. For far less money, they could have hired a couple workmen who really knew what they were doing, to handle the critical socket/zinc/cable joints.
Reaction 2: If you're stuck with so-so workmanship in those joints, then increase the safety factor - but only in those joints. Use larger sockets, with more zinc in contact with a greater length of splayed cable (within the sockets).
