The issue I was concerned about isn't how much power it wastes, but the dropout voltage.
I was looking at some graphs for 9v and AA batteries and at 500mA they'll both drop below 6.3 volts after only 60 to 90 minutes. And that is like 1/4 the run time I thought I was going to get out of the batteries based on my original calculations using their rated mAh.
With a regulator that has a dropout votlage of 2v, the situtation looks even more dire. By my calculations I would only get around 30 minutes before the voltage drops so low that the regulator can no longer provide 6v.
Of course all that assumes the servo is in use 100% of the time, and I'm using 500mA, which I won't be, so I should be fine and have a decent run time.
When concerned with battery life, it strikes me that worrying about dropout differences between different linear regulators is like worrying about what color to paint the walls on a house with a bad foundation.
Have you thought about using a switching regulator? It'll give you a vast improvement in efficiency (probably in the 90% range as opposed to 50 - 60% for the linear regs you're talking about), AND (likely) a very low dropout voltage. You could even do a SEPIC topology that'll keep working when battery voltage drops below your desired output voltage, though by that point the battery will likely be pretty flat. At any rate, runtime will be vastly better than ANY linear regulator in this sort of application.
Of course it'll be a more complicated design, but that's the tradeoff for better performance.
I was looking at some graphs for 9v and AA batteries and at 500mA they'll both drop below 6.3 volts after only 60 to 90 minutes.
The AAs OUGHT to have MUCH longer run time than a "9V transistor" battery (although I would think that the latter would barely be able to provide 500mA at all.) What graphs are you looking at? Using the sort of rechargeable NiCd or NiMH batteries typical in remote control, you ought to be able to do a lot better. (and NiMH has a flatter discharge curve as well.) There is a bunch of data over at http://www.candlepowerforums.com comparing various batteries at "respectable" discharge rates. Run-times of 100+ minutes at a 1A rate seem pretty common for NiMH cells, and 150+ minutes for Alkalines at 500mA: http://www.candlepowerforums.com/vb/showthread.php?t=79302 http://www.candlepowerforums.com/vb/showthread.php?t=64660