Gudmundur wrote:
The op's statement that
the problem follows the tube shows two things, the tube is gassey
and the grid circuit resistance is probably to high. Some of my
old books show 47k as the maximum recommended control grid
resistance.

Just as a compare/contrast:

With 100K grid resistors to the bias supply, the current through the
resistor to a "GOOD" tube is zero. Less than a millivolt drop across
the grid resistor.

And with the "BAD" tube, the current through a 47K resistor to -25V
bias is so high that the drop across the resistor goes from 0 at
poweron to 5V after 10 seconds, and steadily increases over the next 30
or 40 seconds to the point of 300-400mA plate current (at about 0V on
the grid, and 25V drop across the resistor) when the B+ fuse blows.

So, as I see it: with a good tube it doesn't matter much what the exact
value of the grid resistor is. With a bad tube it doesn't matter much
what the exact value of the grid resistor is.

Tim.