Can you elaborate on this, pls?
I got to wondering about this a while back when I discovered that most stones for sale are not supported properly.
Remember this golden rule of machining - everything moves when subjected to a force. Yes, EVERYTHING - even solid Granite!
Now, think about this. If a stone is lapped flat on 3 points at the manufacturer, the edges droop down from the weight of the granite. Pure physics. If the user then mounts it in an edge frame, it droops in the center - the exact opposite of how it was made. That's MAXIMUM deflection from design.
According to my math, the maximum deflection due to its own weight for a 2ft stone 3" thick is about 0.0005. You can probably double that for a reverse flexture (lapped on 3 points vs mounted on edges). A thou is not a lot. But the flatness specification for a two foot B grade block is approximately 3 tenths. So mounting it incorrectly triples the error from its certification. That's for B grade. Quite obviously it is much much worse for the tighter grades. Also consider that for most applications you are using references across a length. A reverse slope from one edge to the other magnifies the effect (just plain old geometry).
For reference, granite surface plates are graded based on their flatness tolerance, which determines their precision and intended use. The common grades are:
Grade AA (Laboratory Grade): The most accurate grade. Used in controlled environments like metrology labs for ultra-precision measurements. Flatness tolerance is the tightest.
Grade A (Inspection Grade): Suitable for quality control and general inspection work. Flatness tolerance is twice that of Grade AA.
Grade B (Shop Grade): Used in workshops and production areas for less critical applications. Flatness tolerance is four times greater than that of Grade AA.
I know that the foundation of all metrology is a granite surface plate. Any metrology lab knows that too. But I reasoned that maybe a small stone wasn't as susceptible as a big one. So I did the math. And it turns out that this is true. But not nearly as much as I had suspected. The smallest you can safely go without a meaningful impact (greater than the cert specs) is 12x18. Any bigger than that and the natural deflection of the stone due to its own weight
exceeds its quality certification. HOWEVER - If the stone was originally lapped on a flat surface or on its edges, you can safely go bigger. Perhaps even as big as 24 x 18. I didn't do the math for this - I wouldn't even know how to go about it. The important question is, how was it made?
You can go to any of the stone makers websites and find lots of info about how to properly mount a stone. They make a really big deal about using their 3 point system, and they all tell you why.
If you don't care about the accuracy of a granite surface plate, then why buy a granite surface plate? Just buy a piece of granite countertop! In the end, you have to ask yourself why you want a stone in the first place. If it's just to do some simple height measurements, the stone quality prolly doesn't matter. If you want to calibrate your metrology equipment, or do some precision comparison measurements, it probably does.
For me, it boils down to this. If I just want to do good measurements, my mill bed is good enough (and was for many years). I don't need a stone at all. It will just consume valuable space in my shop. If I want to do precision measurements and calibrate my metrology, I need a good surface plate and it needs to be reliable. You need specialized equipment to verify the reliability of a stone - which I don't have. So the only way I can get close to being reliable is to use a certified stone the way it was designed and lapped to be used.
I also remember this from my days in industry - the shop machinists and the guys in the quality lab NEVER got along. It was always like the Hatfields and the McCoys.
Obviously, your mileage may vary. Just pick your poison.