What is really being discussed is not loss of temper, but a relaxing of a percentage of the stresses that are established when the spring is compressed, and not when the steel is heat treated. Relief of these stresses tends to occur not only with changes in temperature, but also with shock (recoil, for example) or when simply waiting long enough for it to occur.
In the last case, the compressive stress relief is generally referred to as "taking a set". I read recently that the valve spings in piston aircraft engines loose about 15% of their compressive strength in the first couple of weeks they are installed and run (I don't recall the number of hours of operation), then remain compressed fairly constantly until engine temperature and fatigue finally do them in a few years later. The speed of the onset of the set in these valve springs is greatly accelerated by engine heat, but given enough time, that 15% loss would eventually occur anyway.
But here's the thing, as an engineer I can tell you I would not design anything that depends on a compressed spring without leaving some margin of extra compressive strength. A military magazine? Make the spring 15% stronger than it needs to be, and bingo, you have something that still works even after taking that initial set; even decades later, if it hasn't been fatigued by constant use? The aircraft engine keeps working because it was designed this way.
Leaving a spring uncompressed will keep it stronger longer, but falling below minimum required compressive strength is another matter and depends how critcal a particular pressure was to the design? Other than trigger springs, I can't think of any springs in personal firearms that need to hold critical absolute values. Triggers do need periodic readjustment, so you just live with that.
If it worries you, get your springs cryo-tempered. In the case of the aircraft engines, cryo-trempering (treating, processing, etc.) is reported to reduce the amount of iniitial set they take to about 7% from 15%.
Nick
