Nashville SE 20 –  Power Supply Unit
The power source unit (fig. 1 to 3) might look complex at first sight though it comprises very basic circuit technology only. However, specifically SE amplifiers show poor power supply rejection ratio (PSRR) which is the capability to suppress any power supply variations such as hum to its output signal. Thus the plate voltages as well as the drain voltage must be carefully filtered DC without any ripple. Fig. 1 to 3: Power supply for the SE power amplifier Mains transformer is a Lundahl LL1649 delivering 230V at its secondary coils. The tube section and the transistor section get their AC voltage from the same secondary coils but are decoupled through separate bridge rectifiers (UF540). After rectifying the high voltage for the tubes is 300V. It is then filtered by a C-L-C-R-C network. Voltages for both the KT66 and the 6SN7 is 290V. Additionally the voltage for the gate of the cascoding MosFet is taken from the high voltage and stabilized with a simple 60V Zener diode (1N3000B). The tubes are simulated by 60mA and 9mA current sinks in the simulation and in the schematic (fig. 1). The high voltage network for the transistors lacks the charging capacitor. It is a L-C-L-C network and delivers 180V. The diodes protect the circuit from voltage peaks in cases of a fuse blow. The heater voltages are taken directly from the transformer and are not rectified. Two voltage dividing resistors R4 and R8 deliver +70V to elevate the heater potential for the 6SN7 twin triode. Heaters are grounded through 47Ohm resistors.   When I switched on the power supplies the first time I was impressed for it was dead quite, no hum, no vibration, just plain quite with all voltages standing. I also was surprised that the simulations matched the real life voltages and currents within the limits of the component tolerances. LT spice truly is a great help.    Fig. 4: Rectifier with UF540 fast recovery diodes and charging capacitor