In what situations are these different network designs most appropriate? I know that we want to make sure dynamic headphones see a low enough impedance for the correct damping factor, and that we want to make sure tube amps see either a 4 or 8 ohm load so as not to ruin them. But what about running orthos from a solid state amp -- is there any reason I shouldn't just use series resistors? Does anyone know if the Emotiva performs better if I trick it into seeing an 8 ohm load instead of something much higher?
A tube amp will NOT be damaged if it doesn't see a 4 OR 8 Ohm load.
In Rob's spread sheet, in the three resistor version, you could allow R1 to be infinite (i.e just omit it).
If the tube amp sees a load of 8 Ohms or higher (say approx. 32 or 50 Ohm or 64 Ohms or 43.73945 Ohms), it should be OK. Almost any SS amp should be rather indifferent to a higher or lower load impedance, just don't go TOO LOW.
As an example, R2 = 12 Ohms and R3 = 4 Ohms would be OK for dynamics,
from the headphone's point of view, the output impedance is less than 4 Ohms, so we have good damping
form the amp's point of view, the load is approx. 16 Ohms. A tube amp or SS amp would be OK with this.
Attentuation is approx. 12 dB, so hiss is reduced, if you need to reduce hiss!?
hiss can be reduced more by increasing R2.
But it's all a trade off: if R2 = 12 Ohms and R3 = 4 Ohms, you also have to be comfortable with the travel of your volume control, i.e. for your system would this avoid having all volume control travel between 8:30 and 9:00 o'clock? Hopefully the resistor network would also get the travel more between 9:00 and 1:00 o'clock.
With headphones and a power amp and a resistor network it's all a compromise between reducing the hiss and improving the volume control travel.
Don't over think it, folks.