Yes, I understood what you meant. I was trying to get you to reaffirm your concept of 'ground'.
No offense taken. I sometimes get frustrated when I am looking for intelligent dialog on a specific issue and get general comments or restating the question as a statement instead of adding to the discussion. I would much rather have someone tell me I'm a fool, and then tell me why, than tell me the problem is the same if looked at it from a different angle.
One solution to the problem was not to try to reference the voltage from ground but from a point half way between the supply and ground. With that answer and a voltage divider I have a workable solution.
As always, there are other solutions, but this one is simple and should work. I need to look at what will happen if my voltage still goes negative or above the supply voltage. It could go over +/- 2.5 V, in the example, if the current exceeds 2.5 A. Do I want to clamp that with production diodes? Which is more forgiving, the input to the micro or the input to the OpAmp/Buffer. Maybe one or both have some level of input protection built in.
Do I loose my protection or my range? I don't think I want to incur the expense of a rail to rail OpAmp to get that last little bit of range. Which, to me, means protect the front-end of my micro with a buffer or get the full range by reading the voltage directly.
I would like to think that the solution is not so obvious that I missing it or that I am the only one thinking about such things.
I didn't look at that diagram. Yes, you've got it right!
I'm glad you saw it. It took me some time to put that together. My CAD is not a good as "Dave CAD".