OK, got my numbers about the gyro mass and size wrong - it's been a long time.
And of course I should have precised:
"when it disintegrates", rather than "
when it fails". Obviously the normal failure modes are different - nutation, bad signals, whatever. That was not my point. The individual gyros I had in mind had failed catastrophically during ground testing, after very long storage periods and quite possibly earlier damage; and they came from a design used on UAVs, not manned aircraft.
Just a short extract from "Aircraft accident reconstruction and litigation" by Mc Cormick, Papadakis, and Asselta, though:
"When a gyro platform crashes and the case is crushed, the gyros typically disintrgrate the internal mechanisms and almost explode due to the kinetic energy of rotating steel gyros. Even though gyros ar small, they are heavy (2 to 3 pounds), and their outer diameters at 25,000 RPM are moving about 300 miles per hour.
These steel gyros will not be simply captured in place easily. It is extremely rare to have a crushed platform of this variety with the gyros still in place. In such case, the gyros are not running at the time of impact because at the high RPM, the gyros scatter mindful of an exploding hand grenade."
My point is, in any case, a turbine disk is larger and heavier again, although the rotation is slower - but if it goes off, it's hard to stop.