Nice pair. The Applied Science build is worth watching twice, because its route is the same line as the best modern powder work: ammonium nitrate and iron powder, ball-milled, following Preparation of an α″-Fe16N2 Magnet via a Ball-Milling and Shock-Compaction Approach (Adv. Eng. Mater., DOI 10.1002/adem.201500455). His planetary mill plans are public on GitHub, which makes the synthesis repeatable outside a lab, and a Lake Shore application note on permanent-magnet measurement sits in his reference list. That last detail matters more than it looks.

A DIY magnetometer trace pins down what microstructure controls: coercivity, loop squareness, remanence ratio. The number the field argues about is different in kind: absolute saturation magnetization on a sample whose phase fraction and nitrogen ordering are known. Ball-milled nitride is the hardest case for that number, because disorder, residue, and magnetically dead surfaces all push emu/g down in ways a loop shape cannot disentangle.

That is the fork in the computed-versus-measured post that grew out of your 40-year-mystery thread: near 281 emu/g and first-principles magnetism of α″-Fe16N2 is missing something real; near 230 emu/g and the computed range holds while the giant moment is demoted for the ordered phase; near 165 emu/g and the ideal cell is the wrong end of the comparison. If anyone watching this ever gets a sample through calibrated magnetometry (a Ni or Fe standard in the same jig fixes the absolute scale) with an XRD phase fraction beside it, that data point belongs in our call for measured magnetic data. It would sit right next to the computed 230 emu/g.