Playing with a buffer amplifier for a DDS…

I tinkered together a circuit for a buffer amplifier that I thought would be good to use with the AD9850 DDS board I have. In particular, for the purpose of this exercise, I was trying to preserve the nice sine-wave nature of the DDS input, and present it at the output into a 50 ohm load. I dug around and ended up creating this circuit in LTSpice.

Screenshot

(I didn’t do the design, it was from my notes, but I failed to crib where it originally came from.)

I know it is is ridiculous to look at this from the standpoint of efficiency, but this feedback amp is astonishingly inefficient. Overall power draw averages at 470mw, with only 2.6mw of power actually delivered into the 50 ohm load. That seems pretty ridiculous in terms of efficiency.

Surely there must be a better way (power wise) to preserve the original waveform over a large range of possible output frequencies, yes? Or maybe not. I feel like understanding this is one of those teaching moments for me.