>Dan:
> Someone suggested that you use 8x811A in parallel. Before starting to
>punch out the holes for all of those sockets (not to mention the 200W
>filament transformer) I would suggest you consider the following basic
>electronic facts that would confront you.
>
> Assuming anywhere nearly a normal operating regime, the optimum plate
>load for that bank of tubes operating anywhere near their maximum, will be
>in the range of 500ohms.
? Close enough. With 1800v on the anodes, I calculate 590-ohms (?) of
RL..
>At 10 meters, for a properly designed PINET plate
>circuit, you will end up with a target L value of arounf 300nH, around 500nH
>for 15M.
>
> Next consider the fact that with that physically large bank of tubes,
>the total anode lead from the distributed anodes to the output capcitor, and
>through that to return to the GG, is not likely to be less than a foot long.
>With reasonable size conductors, you will be facing the nominal 25nH per
>foot of inductance, which when multiplied by your ca. 1 foot of wire, uses up
>your ENTIRE budget of inductance.
? new math ?
>In other words, most of the L in your tank
>circuit will be the anode wiring itself. This is necessarily a low-Q way to
>build an inductor and will result in considerable losses.
? I can layout a 3 by 3 square for 8, 811As, including the anode choke
that would need c. 8-inches of anode lead between the collection point
and the Tune-C. 8" of Cu strap has c. 80nH of L. For a Q of 15 at
28MHz, one needs 145pF of Tune C. Since the 811A has c. 5.6pF of
anode-C, the total anode-C for an 8-holer is c. 45pF. Thus, a variable
Tune-C of at least 100pFmax would be needed. The needed load C would be
484pF, and the needed L is c. 290nH, however, the 80nH of anode lead
transforms the 590-ohms to a slightly lower value, so 290nH should be a
bit too high.
>Hence, your 10
>(and possibly 15) meter efficiency will stink.
>
? only for those who ignore skin effect when choosing tank conductors.
> I have found this basic issue slightly troublesome in the 4 tube
>designs, which always tend to show decreased efficiency on 10.
? The only amplifier that I have seen which used tank conductors based
on skin effect is the TL-922. However, even if the tank efficiency was
so low on 10m that it caused an additional 20% loss due to heat, the guy
on the Rx end would never notice the difference.
>Good designs
>get the 10 meter tank coil as close as possible to the anodes to reduce this
>effect.
? True, Eric -- at 28MHz, the Z-transformation process pretty much
starts at the anode(s).
>
> The solution is obvious - you need to raise Rplate, by raising the
>plate voltage and lowering the peak current. This will lead right to the 572
>or similar.
? The above calculations show that this project is do-able with
1800V/1.6A.
>This also reduces the drive requirement considerably by not
>having to drive the plate current as high (for a given tube).
>
? a FT-1000D will easily drive 8, 811As.
- R. L. Measures, a.k.a. Rich..., 805.386.3734, AG6K,
www.vcnet.com/measures.
end
|