Soldering chips and transistors. How hot? How fast? Sockets?

[petemoore]

  Requires frying a couple things.
 I don't recall being sure I ever fried an active or diode with 'normal' soldering.
 I fried an LM386 before I'd heard there are IC sockets.
 Have installed opamps [that then worked] without sockets, and even stacked many opamps [1 pair became damaged], use socket.
 But I've pushed the envelope numerous times, the Diode and Si transistor B/C junction diodes can be tested with DMM and then for performance with the sophistication of the testing equipment here.
 The diodes clipped the signal as intended and the transistors performed their tasks after moderate to medium-extreme heat-abuse.
 Ge transistor installs get socket or heatsunk soldering.
 If a burning sensation causes a retraction-reaction, the silicon probably has a second or so of heating before damage occurs.
 Burning the fingerprints numerous times with varying degrees of minor burn helps with understanding heat transfer/times. Once the smoke just begins to exhibit outgassing [very hot], chip damage should be suspected.
 The socket helps keep heat out of the chip to some degree, removing the chip from the socket prior to soldering prevents all heat from entering the chip.