Very definitely go for the 750! And there's lots more to it. (I wish Garmin would pay me to make a video on how to best exploit the G3X Touch architecture...)
Factors to consider:
* There is NOTHING that really needs to be displayed full screen. For starters, most of the information content is in the center of the screen, and on the right hand screen (MFD, if you will), single window, that good information is farther away than if the MFD screen is partitioned into two windows and the good information is in the left window. I customarily fly with the MFD split into two windows: the left window content is swapped around in flight, often traffic, and the right window is typically duplicate flight instruments.
* Consider the parameter bar as well. That's a wonderful place for all kinds of flight data. I have the left screen set up for next waypoint data and other information; the 650 has destination data, but a gotcha is that is along the programmed flight path, including the approach. If ATC gives you a shortcut, that destination data is flat out wrong. Don't fall into the trap of putting all of the radio controls on the parameter bar just because you can.
* On the right screen parameter bar, I have less commonly used data but I also duplicate the next waypoint data for the occasional right seater.
Here's a good exercise: list all of the available screen displays, and figure out which ones are used in what phases of flight. I tend to use traffic, approach plates, map and chart the most. I almost never use terrain, but that might change if I moved out of the flatlands. And since almost all of my clearances are GPS direct, I never use IFR charts. (Yes, I'm lucky).
You get the idea -- figure out what information is used when. Give priority to information that is used most frequently, and to information that is needed in a hurry. (This is similar to figuring out what functions earn their way onto control stick buttons.) Don't optimize your panel for pre-takeoff, as that's when you have no time pressures. Include the parameter bars and radio tuning in your design. For example, transponder ident is easily handled with a discrete pushbutton and can be done quickly that way, but changing codes is rarely done and there's no great hurry, so it doesn't earn prime space on a parameter bar.
Like everything else in human factors design, it's easy to get to a 90% good design, a lot more work will get you to 95%, and after that, lots more work will exploit different tradeoffs but you'll still be at about 95%. Maybe 96%.
For example: I went through several checklist iterations when my RV-9A came with a hodge-podge of glass panel stuff, and the checklist cleaned up when I went all Garmin. But as I learned more about how G3X system worked and the subtleties, and as I improved my techniques, the checklists continued to evolve. I just finished Rev 31... and in addition to airframe checks, I also have an approach checklist for both LPV and ILS, as there are things to remember.
I'd love to put a 750 into my panel, but it would mean new sheet metal, new wiring, and months of down time. And of course, lotsa money.
