Thanks all. I was taught to torque the bolt to the manufacturer's specification (aviation) then walk away. The idea of turning the bolt further, even just a little bit, is very strange to me.
The thing to keep in mind is that the torque on a fastener is not what really matters, it is a secondary measurement to get to the tension in the fastener. The torque roughly indicates the tension, but only under certain conditions, and you will notice different torque tables for lubricated threads vs dry threads, with lubricated threads requiring less torque. The roughness of the seating surface of the nut or bolt can have an effect, as can the calibration of the torque wrench and the way it is used.
What correlates exactly to the tension of the fastener is the amount that it stretches, and that can be determined by its rotation. The torque value specified is not enough to put significant strain on the fastener, and is used to ensure that all of the flex or compression in the assembly is removed. If you look at the torque tables for a fastener of the same size and pitch you will notice that the table will give you a much higher torque value. By rotating a fastener of known pitch a certain amount after the joint is tight, the amount that it stretches and therefore the actual tighness is known with great precision.
Torque-to-yield is actually a subset of torque-angle in which the stress applied to the fastener is enough to stretch it past the point of elasticity and thereby achieve the absolute maximum holding force. Torque-angle can be used for joints that require high precision but do not require ultimate strength.
When I worked on critical machinery for the Navy, we used two additional methods for ultra-critical fasteners. The easiest was for bolts that went through a flange, where we would measure the bolt with a micrometer prior to assembly, then tighten it until a length measurement indicated that it had stretched the correct amount. For blind holes with large enough bolts, a hole was drilled past where the threads engaged for a depth micrometer and the same process was followed.
