Nashville SE 20 - Some basic Math on the SuperTriode Concept Basic calculations require one component to start with. I chose the output transformer as the most crucial component. For the project the single ended version of a Lundahl LL1693 seemed to be a good match regarding quality, pricing, technical data and size (fig. 1). The SE version of the 4.5kg transformer has a core  air gap of 450µm. Recommended idle current is 230mA with all primary windings in series. Primary inductance is 16H. With a secondary connection B (4Ohm, see datasheet) the reflected primary impedance is 1kOhm. Inherent Output Power Limit The recommended transformer idle current of 230mA allows a theoretical current swing from 0mA to 460mA peak to peak (p-p). According to I 2  * R = P a simplified power calculation neglecting transformer losses reveals a maximum output power of (0.23A / sqrt2) 2  * 1000Ohm = 26.45W ≈ 26.5W (equation 1) which is the inherent output power limit, not too much but fairly above the requested minimum of 20W. Hence I stayed with the LL1693 for further calculations. By the way, the push-pull version of the LL1963 is capable to handle around 300W output power. Output Voltage Swing (fig. 2) Considering a „worst case“ estimation of 30W output power and 1kOhm primary impedance the power output voltage swing p-p at the primary coils according to U 2  / R = P is sqrt 2 * (sqrt(30W * 1000Ohm)) * 2 ≈ 490V p-p (equation 2) With an estimated minimum residual voltage of 100V across the triode the power supply voltage would have to be around 100V + 490V / 2 = 345V ≈ 350V (equation 3) Thus the maximum voltage peak expected to be at the plate is 100V + 490V = 590V (equation 4) Adding a comfortabe safety margin (speakers unintentionally disconnected) the circuit components should withstand 900V to 1000V peak voltage, nothing special for a tube but pretty much for a transistor. Power Dissipation From above equation 3 and the given idle current of 0.23A the theoretical power dissipation is 350V * 0.23A = 80.5 W. If an idle current of 30mA for the triode and an idle current of 200mA for the transistor plus an plate / collector voltage of 350 V are assumed then power dissipation at the plate will be 10.5 W and power dissipation of the transistor will be at least 70W. 350V and 200mA idle current must be within the DC safe operating area of the transistor again with a comfortable safety margin. This will rule out BJTs because of their inherent secondary breakdown and leaves power MosFets. Power MosFets meeting the specifications above are available for instance the IXFH14N100 or IXFN32N120P. However, these MosFets are designed for industrial high power switching applications. They are quite non-linear at low drain currents and have rather large gate capacitances. Below some data on the IXFN32N120P are discussed (fig. 3).                                    Fig. 3: Maximum ratings and transconductance of the IXFN32N120P. At low currents g m  is a linear function of I drain . Estimated transconductance g m  at 230mA is around 0.6S. U max  of the IXFN32N120P is 1200V, P d  is 1000W, C iss  is 21nF.  With a load resistance of 1kOhm the effective or Miller capacitance is   C M  = C iss  * ( 1 + gain) = C iss  * (1+ (gm * 1000)) = 0.021µF + (1+(0.6*1000)) = 12.62µF       (equation 5) 12µF (!) dynamic input capacitance is difficult to drive without compromising high frequency response even without the essential gate stopper resistor. Thus the IXFN32N120P or similar devices with even lower gate capacitances are hardly suitable for a SE power amplifier according to the SuperTriode concept. Core Magnetization The SuperTriode configuration (fig. 4) suffers from an additional shortcoming: Even at large signals the plate current remains almost constant. Thus in a single ended design the tube just magnetizes the output transformer‘s iron core but does not contribute to the output power. If 200mA idle current are assumed for the transistor, the inherent power limit calculation from above equation 1 drops to 20W and will be less if transformer losses are taken into consideration. Conclusion With regard to the estimated minimum specifications of the transistor it may be concluded that the SuperTriode concept as illustrated in figure 4 is not suitable for SE power amplifiers if an output power above a few watts is requested despite its beauty and simplicity. On the other hand the concept is far too fascinating to be simply discarded…..
 Fig. 1: Lundahl LL1693 output transformer
                Fig. 4: SuperTriode concept.
 Fig. 2: estimated voltages at the plate /             collector of SuperTriode circuit             discussed