[ARC5] Modeling the PA stage

[Ian Wilson]

I spent a few hours on this yesterday. Any numbers in the following are for
the BC-459 at 7.1MHz.

Tank coil: the business part of this is 6 turns on a 1.75" diameter ceramic
former, spread over 1". There are also a couple of additional turns
occupying another 1/2". Estimated inductance: 1.85uH. This would resonate
with 272pF at 7.1MHz.

Coupling coil: on a 1.25" former. Assuming 1 turn; about 0.075uH.

Roller inductor: 24 turns spanning 3" on a 2" diameter former.
	 25% of coil =  6 turns = 2.2uH
	 50% of coil = 12 turns = 6uH
	100% of coil = 24 turns = 15uH

Half-roller would resonate with 84pF at 7.1MHz.

Tube source resistance: my pair of 1625's draw about 150mA at 600V.
Rp (see ARRL Handbooks, etc) = 600 / (2 x 150mA) = 2k ohm.

If you ignore coupling coefficient and just think of the tank + coupling
coil as a tuned transformer, this provides a voltage ratio of about 6:1 and
hence an impedance transformation of 36:1. Using the actual inductances
instead suggests an impedance transformation of 25:1. This would match the
2000 ohm source impedance to 80 ohms.

The effect of the less than unity coupling coefficient is to increase the
transformation ratio. A coupling coefficient of 0.5 reduces the matched
output impedance to about 20 ohms.

The coupling coil has some inductance, and so does the roller inductance (if
not on the end stop). These need to see a load that looks capacitive,
otherwise the inductive reactance reduces the available output power. I did
some playing around with (simulated) components. A good recipe for 7.1 MHz
looks like:

  - series capacitance of 50pF (e.g. ARC-5 vacuum capacitor)
  - 10uH roller inductance setting (somewhat more than half)
  - 50 ohm load with 560pF parallel capacitance

When I get time (perhaps over the weekend), I'll see how the reality matches
the simulation.

73, ian K3IMW