If a valve passes an anode current of 1mA then this equates to 650, 000, 000, 000, 000 electrons a second. When you consider that an output tetrode like a KT66 can pass up to 100mA in anode current then as Cilla Black would say 'thats a lorra, lorra electrons!' I think this little fact demonstrates quite well that over a period of anywhere between 2000 and 8000 hours on average, a valve may loose a proportion of it's emissive properties.
During the early life of a valve, the effect of emission loss is realtively imperceptible but after a while, the multual conductance, gm, falls off significantly. The following graph shows the change in gm with negative grid current at three points in a particular valve's life; curve A is the valve when brand new; curve B after burn in and curve C when worn and showing less than 50% emission.
You can see also that the slope of the curve changes too and the sensitivity of the valve is measured by the steepness of this slope. The effect of shallowing of the slope means loss of volume, loss of sparkle and increase in distortion due to overloading. So you can see now why an emissive limit of approx 50% may be considered the bare minimum for continued usage and typically an emissive limit of 70% for a valve to be considered to show reasonably good emission.
In our next blog entry we will look at some other things that happen as valves 'wear out'.