[MrFixr55]

Hi Dale, 

Though the cathodes are grounded, there is something more negative than ground.  Resistor # 66 in the drawing goes from chassis ground to the center tap of the B Supply secondary, making the B supply center tap the most negative point in the radio. Note also that the driver and 1st audio stage get their grid bias off the same tapped resistor. Note how the 2 main filter cap negative terminals do not go to ground but to the B supply secondary center tap.  Although it looks like fixed bias, variation in current draw will vary that voltage.  This is why I am calling it cathode bias.  Newer radios usually have chassis ground as the most negative point and resistors from that point to the cathodes of those tubes that require bias. It may be less complicated and cause less hum that way.

See the annotated pic below:
   
In contrast is this diagram of a high power Heathkit W6M Amp below.  Note that the cathodes of the 6550 output tubes are also at chassis ground potential, but so is the B Supply Minus terminal.  The negative grid bias comes from another winding on the power transformer (I should say the extremely substantial power transformer).  This is an example of fixed bias:

Heathkit W6M Williamson Amp.pdf Size: 30.77 KB  Downloads: 164

BTW, I have one of these. They are rated at 60 watts But I don't think RMS, but even more conservatively. I bet that this thing does at least 100. I haven't fired it up in almost 50 years. One of my projects is to recap it. I may sell it. I wish I had 2.

IMHO, the difference between fixed and cathode bias is not whether or not the cathodes of the output tubes are at chassis ground, it is whether a separate power supply of good regulation supplies the grid bias.