[Amps] Sanity check

[Vic K2VCO]

I am doing something dumb, but I can't figure out what.

My new amp is a pair of 813's, class C. On 20 meters, it loads to 500 ma at 2050 volts and 
produces about 800 watts output. The tuning capacitor and tank inductance are close to the 
calculated values (from GM3SEK's spreadsheet). With power off I connected a 2050 ohm 
resistor -- (2050v x 0.5a)/2 -- from the plates to ground and used my MFJ analyzer 
connected to the output to get a 1:1 SWR at the same point that it tunes up under power.

Everything is perfect on 20 meters. But on 160, it's not.

When I used a tank coil of the calculated inductance, the loading was much too heavy with 
a practical value of output capacitance. I added several turns to the coil. Now it loads 
properly, but the tuning capacity is considerably less than the calculated value. Also, I 
get only 750 watts with the same input. I expected that it would be more efficient on 160 
than 20.

On 10 and 15 meters, the tuning capacity required for a match is considerably more than 
the calculated values (and of course the inductance is lower). I haven't checked the 
efficiency yet.

Here is one fact that might be related: instead of the usual RF choke of about 200 uh, 
I've made the 10-20 meter tank coil out of tubing with a wire through the center. The wire 
is connected to the plate side of the blocking capacitor. After it exits the tubing it 
goes to a large 1.7 mh choke.

The idea (courtesy WC6W) is that the inner conductor of the coaxial tank coil will be 
close to ground potential on 20-10 meters, so it's possible to use a choke that will do a 
good job on 160 meters without worrying about series resonances in or near the higher bands.

But I have a feeling the choke has become part of the tank circuit!

Any thoughts? Should I just order a choke from Ameritron and use the conventional system? 
Or is something else going on?
-- 
Vic, K2VCO
Fresno CA
http://www.qsl.net/k2vco/