Electronic – the temp velocity of Nichrome

Temperature extremes will damage the physical package itself and interconnecting wires. It is the different expansion rates or silicon, metal wires (gold??), and plastic housing that will effectively destroy the part. The silicon itself will normally survive to a limit. That is why earlier on in electronic development there was a difference between plastic dip and ceramic dip packages which had different temperature ranges. The silicon was the same, just the packaging was different. It was this difference in packaging that allowed the 'military' ceramic packages to have a greater usable temperature range than commercial plastic package.

With SMD packages, this has changed slightly and they all tend to have greater temperature range, but you can still get commercial, industrial and military temperature ranges of 0 to 85, -40 to 125, and -55 to 125 as a general rule.

That is why devices have lower limits and medium range upper limits, due to packages constraints.

There are also phsyics limits for the upper end where the silicon itself may become damaged and the dopants might defuse? I am not sure of the exact function here or temperature limit.

Have you 'scoped the output? The LM35 is susceptible to EMI (which you have addressed with shielded cable) and load capacitance - which may be an issue. If there is instability due to load capacitance (>50pF) you can isolate this capacitance with a series resistor. If external EMI is still an issue, an RC damper can be added. Both effects are addressed on p7 (figs 3 & 4) of the datasheeet.