SB-200 input impedance correction on 10 and 15 meters

I had known that both the 10 and 15 meter band input circuits on my old Heathkit SB-200 linear amplifier were not well matched to my transceiver.  Using this amplifier with a newer radio, would cause the transceiver to fold back on these bands with high SWR.  Referring to the schematic we see that these employ an L-network to attempt to match the input impedance to standard 50 ohm load.

Going through the mathematical exercise for a PI network, which I won't trouble you with here, it stands to reason this would probably work better just like it does on the lower bands.  This more sophisticated matching network filter can have more loss, but that's not something I was worried about.  The limited number of turns on the 10 and 15 meter coil meant that they required some fairly small capacitors to attempt to make the matching come out correctly.  The problem is that those size capacitors are about the same value of some of the stray capacitance I believed to be in and around the wiring.

The plan was to double the number of turns and increase the capacitance required on the Input side of the network.  I then stumbled on this article by PA0FRI who came to the same conclusion.

I stripped the 10 and 15 meter coils from their windings and re-wound with 1mm diameter magnet wire.  Scrounging through the junk box yielded the capacitors that I calculated and were confirmed by PA0FRI.  I didn't have any tuning caps available, so trial and error yielded appropriate values for the output capacitors which had to factor out the stray capacitance in the wiring.

The trick was how to validate and tune the slugs in the inductors.  I've lost the source from a forum post on eHam, but the procedure is to simulate the resistance of the tubes per their data sheet.  So we temporarily place a 220 ohm resistor between the filament lead and then to ground on the chassis.  While manually pressing in the transmit / receive relay an antenna analyzer is used to measure the impedance at the input connector.  After few tweaks of the slugs both bands yielded acceptable impedance at less than 1.5:1 SWR reading.

The final results are below: