Watching Magnets Rust | Hacker News Reader

Watching Magnets Rust

kjmagnetics.com

44 pointstrynewideas14 comments

KJ Magnetics is an interesting example of driving traffic for online shopping.

I could buy similar products on amazon, but since I'm occasionally there reading the interesting blog, its kinda my first choice.

Also another common amazon problem is poorly proofread unprofessional ad copy. Like if they can't bother translating full specs into English, I don't feel they can be bothered to properly ship on time. So the attention to detail and high quality details in the ad copy also drive sales above and beyond typical amazon.

They could use amazon as a back end fulfillment platform, but the amazon web UI and amazon search engine just don't cut it compared to the KJ online store. Its an interesting alternative strategy to observe in the era of "all retail purchases will inevitably eventually move to Amazon". Someday, in theory, the amazon UI could encompass these features and adsorb all this traffic and commerce, but its not happening yet. Working the opposite direction places like McMaster Carr and MSC and similar have been pushing downward from corporate B2B only to semi-retail sales.

Excerpt:

"While you may know the chemical formula of neodymium magnets as NdFeB, or Nd2Fe14B, they are mostly iron and neodymium.

By mass, they’re roughly two-thirds iron and one-third neodymium. Only a tiny fraction is boron, plus a few other elements."

That phrase had me slightly puzzled - I thought chemical formulas were for molecules, i.e. combinations of elements that actually react with each other and form bonds, like H2O or CO2? While a neodymium magnet is actually an alloy of mostly two metals, plus a little bit of boron, but without chemical bonds?

Most metals are crystalline, so while the atoms are bonded together they're all bonded into big lattice structures rather than neat little ones with covalent bonds.

(Changing the distribution of crystal size and shape is an important part of metal forging as it affects strength and ductility)

The notation is also used to describe the ratio of elements used in alloys and crystals. In such a case the elements are also bonded together, but form much larger structures so writing out a specific number for each 'molecule' becomes meaningless (the numbers are huge and also vary massively).

Before this point in time, I always thought that Neodymium magnets were made of... well, Neodymium <g> (and yes, I claim complete and total ignorance! <g>)... it never occurred to me that they were mostly Iron... Amazing just how wrong my thinking about them was, prior to this point... (but I forgive myself... as JFK so eloquently put, "An error does not become a mistake -- unless you refuse to correct it."...)

With alloys, the formula given is akin to the 'empirical formula' for molecules, i.e. the simplest integer ratio. For the mass fraction, the relative masses of atoms of iron:neodymium:boron are approx 56:144:11, so Nd2 Fe14 B should be 27% Nd, 72% Fe and 1% B by mass, I think.

> but without chemical bonds?

Metallic bonds. My understanding is all bonding is atoms sticking together electrostatically, with (insanely fast) "slow" lumbering nuclei, nudged around by their local e-field, and wacky electrons, adopting weird spatial distributions based on the current nuclei positions, setting that field. The bond taxonomy descriptively sorts by characteristics. Here, one is delocalization of valence electrons. Perhaps picture https://www.youtube.com/watch?v=oSCX78-8-q0&list=PLaFe0BJiho... , STM imaging of cold CO molecules sitting on a copper surface, IIRC. You can't see the individual copper atoms, and there are extended ripples in surface electron density.

Found a laptop hdd near a forest last year, took it home to see if I could do anything with it. To my surprise the head magnet corroded to the point of becoming soft material and had no magnetism left.

I am wondering: shouldn't the head be a solenoid? If the head is permanent magnetic it would destroy the stored information.

Permanent magnets don't do much to hard drives[0]. Well, with a strong enough magnet, it will, but the magnet will mechanically damagethe internals before they affect the actual data.

Anyways, it can't be a solenoid because solenoids are unidirectional. They actuate in one direction, and have a spring return in the other. The actuator in hard drives needs to be able to actuate in both directions, so they use a magnet for the coil to repel and/or attract.

https://www.kjmagnetics.com/blog.asp?p=hard-drive-destructio...

You mean magnet AND a solenoid, surely, the actuator can't be just a magnet.

That's why I think you both are right.

For modern hard drives the head actuator is linear motor made from permanent magnet(s) and solenoid. the arrangement is called voice coil, because it is mostly the same thing as coil of an speaker (although the geometry is 90 degrees rotated such that it is flat instead of cylindrical)

so much that a Guy on YouTube modded a head to oscillate from audio signal and get a 'speaker'

Reply on news.ycombinator.com