[Amps] 8877 - To Choke or not to Choke?

[Rich]

>As I've been sorting through my 8877 filamant issue, I have noted that some
>GG 8877 designs tie the filamant and cathode together while other designs do
>not.  The designs that keep the cathode isolated from the filament generally
>do not use a filament choke since what little RF that may be coupled to the
>filament can be easily bypassed to ground through judiciously placed bypass
>caps.   In these designs, I have seen one side of the filament grounded  and
>alternatively, designs where the filament is balanced and floating above
>ground with each side of the filament connected directly to the secondary of
>the filament transfomer.
>
>Given these design variances, is there a consenus of a preferred method? 

When one side of the heater is grounded, and an inadvertent hv+ arc to 
gnd temporarily raises the cathode to several kv negative, such will  
cause a heater/cathode arc that can burn open the heater as the hv filter 
is discharging.  When the heater is not grounded, and a jumper is 
connected from cathode to heater, this potential problem is avoided.  The 
tradeoff is that a 10a, 20 - 40uH heater choke is required.

 
>Is there a measured performance deficiency in the design that grounds one side
>of the 8877 filament? 

No