One modification that reduces the heat stress in "SSB" mode is to replace the factory-original 5.6-Volt zener diode with a series string of thirty (30) 3-Amp rectifier diodes. Each one provides about 2/3 of a Volt of bias. Thirty of them in series adds up to about 20 Volts. This reduces the zero-drive tube current so that the tubes only push about 30 Watts of heat apiece with no drive. This leaves some margin for running AM in SSB mode without melting the tubes.
It's no big trick to build from perfboard. We make a DIY version, but there's more to it than just this alone.
Reducing the tubes' idle current will now change the input impedance of the tubes. The SWR on the amplifier's input when it's keyed will rise if you install this trick by itself.
We install this input-matching circuit directly under the tube sockets for a customer who wants an 11-meter only AM hot rod.
To make room, we move the fat blocking cap from where it was over to the input pole of the relay. Makes sense to also replace the coax feeding from the relay to the tubes' cathodes with teflon coax if you have it on hand. The capacitor on the input-side of the coil is 68pf at 2500 Volts. The output-side cap on the coil is 47pf at 3kv. A lower voltage rating would be vulnerable to breakdown if you get the tubes too hot and one of them arcs inside. The coil is 7 turns on 1/2-inch diameter, spread for lowest input SWR.
Beefing up the parasitic chokes is also a big deal. The harmonic current in the tubes will be higher running hot-rod AM, and overheat the stock parasitic chokes. We use three 100-ohm 5-Watt resistors in place of the original single 47-ohm 2-Watt carbon resistor.
There are additional measures to take that beef up other weak links in the chain for hot-rodding the SB220. Bypassing the band selector's output side works best if you tap the output coil one turn farther down from the original 10-meter tap point. This compensates for the stray inductance of all those wires on the band selector. Using the original 10-meter tap point may increase the Q factor of the output circuit in a way that makes the tubes run hotter than necessary.
But that's enough for today.
This is one of those "You just gotta want to" modifications. Trying to justify the cost by how much higher you'll drive up anyone's S-meter is not so easy, either. This sort of setup makes the SB220 safe to run 2000-Watt AM peaks or more.
Doesn't make it a good idea.
73