[Amps] Grounded B+ amplifier

[R.Measures]

>Years ago I recall reading about an amplifier where the B+ supply was
>grounded and the cathode was "hot" with the B- supply.  I'm wondering if
>that might have some advantages for high power amplifiers.
>
>Specifically, the pi-network output could be simplified by eliminating
>the plate choke.  Here's how the output would look:
>
>http://www.dslextreme.com/users/teeaye/B+.jpg

**  Such a circuit is high-pass, although it does eliminate any RFC in 
the anode circuit.  For a low-pass configuration, a RFC would be needed 
to keep the anode at DC-ground potential.  
>
>Believe it or not, this is a pi-network output circuit, just redrawn.
>The same component values would be used as with a conventional circuit.
>
>I see some drawbacks:  
>
>1.  The filament transformer would have to be insulated for the full
>plate supply voltage.
>
>2.  Likewise, the grid current meter would be at the full plate supply
>and would have to be carefully insulated from the chassis.  Metal case
>meters would not be a good idea.  

**  Insulating the entire meter and metering circuit would probably be 
needed, even with plastic meters.

>In fact, everything in the
>cathode/grid circuit would have to be well insulated, such as the zener.
>
**  At this power level, my guess is that there is no cathode zener 
because the amplifier would probably be AB1 grid-driven -- i.e., 8973, et 
cetera.   

>3.  Switching between standby and operate modes would require a
>well-insulated relay, perhaps a vacuum type or reed relay.  
>
**  Switching could also be done electronically with an optoisolator - 
such as in:
http://www.somis.org/D.a.04.GIF

>4.  The load capacitor would have it's frame grounded as always, but the
>tune cap would have to be insulated on both sides.

**  Good point.  And the Tune-C bypass would be required to carry the 
full circulating current in the tan.  For an 8973 using a Q of 10, the 
bypass would need to be able carry c. 700A-rms.  This is probably why the 
low-pass configuration tank is typically used on elevated-cathode 
amplifiers.  
>
>And there may be others, but eliminating the plate choke with it's nasty
>resonances would be a big plus.  

**  Indeed, however, this problem can be addressed by bandswitching two 
HV-RFCs - or even by parking the resonances in unused spaces of the 
spectrum..

>While it appears at first glance that
>the coupling capacitor is gone, it's function is actually shifted to the
>B- connection at the cathode.
>
**  Good point.  However, the cathode must be well bypassed to circuit 
common in a cold-cathode grounded-screen configuration.  Thus, there are 
tradeoffs with both the hot-cathode and the cold-cathode circuits.  

>Comments are welcome.
>
>--
>Bill, W6WRT