Re: How square are PECs? *VIDEO
craigd
>Can't see much of anything.
Sorry about that - I got in as tight as I could. If you have a large screen (mine is 23") it is better.
>comment section says that Mit Class Bs are off by a thousandth at 5".
Actually, it can be off by that amount to meet spec (the spec is 0.00080" + blade length/10,000). How do you know your tablesaw fence is that square? Over what distance? That is the idea, to be able to repeatedly verify an object in relation to a known standard, whatever that standard might be.
To beat the horse a bit more, how do you know that Starrett is good? Drop it a couple of times and maybe not. Or use it and the little bumpers in the stock that keep it square wear down and if the wear is uneven, the blade tilts (I've seen this). Starrett is a good brand because they publish the spec they meet and we can trust them to meet it.
I did this to test my new testing tool and I like numbers. I also do metal work where sub thousandths of an inch matter (the eye can see gaps of well under a thousands of an inch - light will actually start to turn blue as the longer wave lengths are filtered).
Do we care in wood working? Sometimes and we often do work to those tolerances without measuring. If you cut a tennon out of square, a couple of thou is very obvious during fit up - that gap sticks out like a sour thumb. Or you clamp it tight and that long rail doesn't run level; cut it 0.040" out (over 5", which is half a degree) and it will be almost a quarter inch at 2'.
Errors like these are obvious and easy to correct by eye; if it fits good enough, it is good enough.
On the other hand, Bill Tindall posted data on gaps, glue and strength. There, it can be hard to see or judge the gap so you gotta measure to ensure the glue is actually acting as an adhesive.