Even tho the exact thing happened to me on an LJ crank (59 Super Rocket) over 35 years ago so this looks horribly familiar.....I concluded it was the incorrect radius at the edge of the journal.....the machine shop I used to regrind the ends on that crank did not seem to be aware how important that was, although to be fair, I did not either, so did not give the shop the instructions on radius required, or check the crank after the grind.
One thing to remember when buying a crank, it's generally not possible to correct an incorrect radius, once the metal is gone, you can't put it back!
On my SJ crank the machinest just did the best he could given on the prior grinds the radius had been made a bit small, and I only wanted a grind down to the next undersize.
Out of interest, does anyone know if this same mode of failure occurs on SJ cranks? Or is it just a problem with LJ?
From what I've heard, contrary to popular opinion, the SJ crank is not THAT much more prone to failure than the LJ, even under race conditions.
The SJ has wider journals than the LJ, so big end bearing area is not that much different to the LJ, and the SJ has a smaller sludge/oil supply drilling through the journal (not that that part of the journal adds much strength anyway).