Has anyone had trouble with the IME screws not fitting in all the way?
I recently upgraded a few of our motors to IME’s, and had no trouble with the screws. Possibly your IME’s have damaged threads inside, preventing the screws from going all the way in. You can always grind the screws shorter if you don’t want the head of the screw sticking out.
I recommend just using the optical shaft encoder instead of IME’s because of internal problems with them and gear stripping, but when i did use them i never had a problem with the screws. If i did i just used a dermel to turn it into a flathead screw. Just use zip ties to hold it and the motor together.
The problem with the optical shaft encoders is that you very quickly use up all your ports. You’ve got 12, and each optical shaft encoder or distance sensor requires two. Some of those 12 are unavailable to things requiring paired wires. So that leaves you with a maximum of five optical shaft encoders or distance sensors, and that’s if you only have two single-wire (well, single three-wire) sensors. If it weren’t for that huge limitation, I would be quite happy to use only optical shaft encoders.
You only need them for axles that make a full rotation. Potentiometers can be easily substituted on a lift, claw, etc.
That is not true. Potentiometers do not rotate 359 degrees.
And I never said they did. Most claws/lifts do not rotate more than 265 degrees.
You said “You only need them for axles that make a full rotation.” But in actuality you need them when you exceed 265 degrees, not 360 degrees (a full rotation). I happen to be dealing with just that issue.
I didn’t disagree with that, did I, and that wasn’t even part of the post I disagreed with at all, was it? At the same time, there are many lifts that use more than 360 degrees of rotation. Consider rack and pinion lifts and cascade lifts. Earlier you said you only need them (IME / optical shaft encoder) for full rotations as a response to my pointing out the limited number of ports becoming a problem with optical shaft encoders. I happen to be dealing with a non-full rotation greater than 265 degrees. A correct statement would be “you only need them for axles that rotate more than 265 degrees.” Even then, that wouldn’t entirely solve the problem I pointed out. In addition to rack and pinions and cascade lifts, consider what happens if you’re using optical shaft encoders with holonomic drives, where you’ll need one for each wheel. Again, you very quickly run out of ports using optical shaft encoders.
Then use a sprocket where you would mount the potentiometer, and have it chained to a larger sprocket, which has the potentionmeter on the same axle. I’ve done that before.
Yes. That could work nicely and should be able to fit well with low-strength chain.
Still out of ports, unfortunately. 
The whole port issue is the sole thing in favor of i2c. Well, space can be, too, as the optical shaft encoders are huge.