i was answering a question on the askelectronics reddit. the question was whether changing the gain bandwidth product of an op amp reduced the current draw of a CMOS op amp. so, i fired up LTSpice and started looking through the op amp models for a CMOS op amp with external compensation. i couldn't find one (and wasn't so sure it would include a behavioral model for the supply current, even if there was a model). so. i "built" one. the obvious answer to the original question was that changing the diff amp current will affect the GBW, but changing the compensation cap does nothing to change the standing currents in the diff amp or the voltage amp. the model i built verified this quite easily. after posting my answer to the question, i started playing around with the model, and this thing looks like it would make a very good headphone amp. so, i'm going to post it here. feel free to mess around with component values etc... the best thing about a discrete op amp, is that you can change component values, add stuff to it, change components (replace a mosfet with a jfet, etc...). and test it immediately. if something doesn't work right you can troubleshoot it because it's not on a tiny sliver of silicon.