It’s not well studied. Some discussion over here though.
A 24 minutes -TL is inadequate to call stability. De facto standard’s 12 hours but with diurnal and seasonal temperature swings when those 12 hours occur can matter even in hot testing. In an office environment it’s not uncommon I get peak SPD temperatures just before HVAC starts up in the morning after the rig under test’s spent all night pumping up its local hotspot.