I was musing about the post on using lower heater voltages and think this
could bear a little more scruntiny.

From the material and physics standpoint, the heater raises the temperature
of the thermionic valve until the phonon internal energy is sufficient to
overcome the surface work function. At that point the cathode can emit
electrons freely. This is quantum mechanical effect, so there is a very
sharp cutoff point (temperature) at which emission can occur. It's not
"brick wall" sharp because the emitting coating layer is not infinately
thin--thus collisions and other effects occur as the electrons seek to
escape the surface.

One would think, however, that there is a fairly small region where the tube
would be begin to work--and clearly it is quite a bit less than 6.3 or 12.6
volts. The next thing to consider is the variation in the manufacturing
process and the temperature range over which the tubes must work. NOS tubes
from the 50's and 60's--where military quality was needed--have the best
uniformity of mass produced tubes. And indeed some of the same tubes have
military version with lower voltages and wide temperature specifications.
So if a tube must function at 0 degrees, then the heater construction for
that filament must provide enough energy to overcome environmental cooling
losses at 0 degrees. This implies that at 25-75 degrees (C) more energy
than necessary is being used.

So it makes sense that running a tube a lower heater voltages, especially in
a climate controlled house, should be very easy to do. So what voltage
could be used? I recitfy and filter my heater supply through a CRC network
where the R is 2.2 ohms. This gives about 5.5 to 5.3v for the heaters--and
it has never been a problem. Because I can get surplus 200,000 to 500,000
uF 30V caps for about $1, this makes a brute force, but easy and cheap way
to rectify & filter heater supplies.

Well, this makes good armchair philosophy, but how does it work. Patrick
made the comment the gm would be reduced with lower heater voltages, but I
wonder it that is true. For older tubes, sometimes the gm measurement goes
up on a tube tester when the heater voltage is increased. I suspect that
this is because of the surface coating has partially evaporated, and
increasing the heater energy frees up more deep electrons that are being
occluded by collisions. However, I notice that on new tubes that are broken
in--changing the heater voltage from 6.3 to 7.0 does not usually change the
gm measurement. Again--this is not always true, but I like the idea of
running tubes at a lower voltage to give them longer life.

Clearly, there are a lot of factors and this only scratches the surface.
One interesting question is the difference between AC and DC for energy
required at the heater filament. And the answer might be different for
50Hz and 60Hz--as there is a cooling and heating period in each cycle. Of
course the time constant for cooling is fairly slow, but it might make a
difference. It would be interesting to test a series of tubes with AC vs DC
vs voltage on the heaters vs current flow through the tube at a constant
plate and grid voltagle. Then repeat for different plate and grid voltages.
Sounds like a great college lab experiment.

Thoughts?