does anyone remember a similar case (impact into water, no fire, relatively quick recovery of the FDR/CVR) with such an outcome, especially with solid state based recorders ?
I am not a pilot. I am, however, a retired electrical engineer who long worked for a semiconductor manufacturer in design, reliability, and manufacturing data analysis jobs.
Not very long after our first flash memory chips went in to one of the earlier solid state memory recorders, my reliability lab colleagues had quite a fire drill in assisting data recovery. It seems that in that case the recorder manufacturer's suspension had failed to provided a low-enough shock environment to at least one of the chips, and there was mechanical damage to the extent (if I recall correctly) of fractured wire bonds (yes--this was so long ago that chips were still connected to their pins by wire bonds). Anyway, the lab guys fiddled up an arrangement, the chip was read out, the accident investigation proceeded, and the reputation of solid state recorders was protected.
I suspect the housing/suspension was not very well done in that case. The old specs for bond wire integrity involved momentary accelerations in excess of 10,000 g, if memory serves me correctly.
They are not invulnerable. Actually as a crash survival medium, I'm pretty sure solid state memory of any kind is mechanically inferior to the old aluminum tape with scratchings, and probably also inferior to magnetic tape. But the survival has a lot to do with how well the recorder designer did his job, not just the medium. If it is going back to Honeywell, you only need to trust their greed to expect they will try really hard to get the data out. Their business and reputation have a stake in that outcome.