Yeah, surface grinding is certainly an art. So many things can affect the outcome. From what I know, ball bearing ways are generally more accurate than V bed ways for supporting the table. What's used to move the table can also have an effect. Hydraulics are generally thought to be best, and usually allow for some controlled power feed. Cable actuation is next in line for smoothness and the last choice is rack and pinion as the action is coarse enough that you can sometimes see the rack pattern come through on the work surface.
I'm sure it was my incredulity that Fitzo was referring to as I was incredulous at the thought of a home built aluminum machine holding a tenth (.0001). Every shop I've ever worked in and every machinist I've ever spoken with would understand a tenth, or a couple tenths to mean ten thousandths of an inch. I haven't played with my little old used Yuasa 6x12 surface grinder enough yet to be able to judge it's accuracy, but I'd be shocked if I can do much better than .0005 as it's not the biggest and most rigid machine around. For a surface grinder it's pretty light weight at around 700-800 lbs. I believe it's 3/4 HP but it's three phase 220 running with a static phase converter off single phase 220.
An aluminum head might not warp, at least when torqued down correctly. Get the torque out of sequence and they will warp pretty quickly. They certainly do experience thermal expansion and contraction. There's a pretty common practice in machinig and fabrication known as a shrink fit. When you need two parts to go together really securely, you machine them to pretty tight tollerances and then freeze the smaller one and heat the larger one to a couple hundred degrees in an oven, then slip them together and when the temprature equalizes the parts are joined so tightly that it takes a pretty good sized press to seperate them again. This is well documented. One of the fellows over at
www.practicalmachinist.com even posted the formula for expansion sometime back so you could figure out exactly how much variation you could count on at various temps. I think it was Forrest Addy who is an old machinist that should be looked upon as a national treasure. That guy's forgotten more than most machinists will ever learn about the trade.
I'm curious enough now that I might have to take a 2" aluminum block and freeze it, then measure it with a mic and then heat it in the oven and measure it again to see if I can measure the variation.
Good luck with your project.
John