Hi Ralph. I've heard it said that tubes are ideal voltage amplifiers and that transistors are ideal current amplifiers. For an input buffer the grid of a tube provides almost unlimited impedance. Why not use an input tube before a class D output stage?
Ralph
@Atmasphere can answer definitively, but here are a few points to ponder:
1. The Atma-Sphere Class D amplifier's input impedance is 100 k-ohms, which is comparable to many tube amplifiers (and higher than some). That is effectively "unlimited" for audio components since preamplifier (or other source) output impedance is usually (very) low.
2. "Ideal" in this context is subjective but it is true that transistors generally provide higher current capability and tubes greater voltage range. Higher gain and greater feedback allows transistor circuits to operate with lower distortion, though the intrinsic distortion series of uncompensated tubes is lower (but any practical circuit has
some feedback, even the "zero-feedback" circuits).
3. The input impedance of a bipolar transistor depends upon bias current and can be very high, but is very circuit dependent. Just as a tube circuit could have very low input impedance (but rarely does, at least for audio). There are trades such as gain and bandwidth for low bias current.
4. Transistors such as JFET and MOSFET devices have very high ("unlimited") input impedance, much like tubes, and like tubes there are trades for high input impedance. JFETs are often used for phono preamplifiers due to their high input impedance and low noise (as well as distortion characteristics similar to tubes). Designs must balance numerous variables to optimize the circuit (and thus product) for the application.
5. Cost and reliability of including a tube circuit in a SS design are among reasons to stick with transistors, but as stated above it is a design choice and a tube input may not fit the goals. A tube requires a high-voltage supply and buffering to drive a SS output stage, meaning additional power supply complexity, with added requirements for a class D amplifier to handle the potential high-frequency noise fed back from the output (as well as coupled/radiated internally) and work with the very high internal (loop) gain of typical self-oscillating class D amplifiers. The high impedance and low gain of typical tube circuits means high isolation from the power supply and any other (e.g. switching) noise must be achieved for high performance. Staying with SS devices can make that much easier.
FWIWFM/HTH/IME/IMO/etc. .- Don