Harold,

Having worked from prints for what seems my entire life, it wasn't uncommon
for me to redraw (on scratch paper) and re-dimension parts so they were
meaningful in the way they'd be made. No changes were introduced, just
dimensions provided in a way that was useful.

My early engineering education happened just before CAD took over
completely. It existed, but our colleges had not yet ripped out all of
their drafting machines. As a result of that, I learned graphical
methods from books that probably had their first several printings in
Sanskrit. My descriptive geometry class started with a grinning
instructor saying "... all but one of these guys are dead" - referring
to the author. He of course was not taking joy in their demise, but
emphasizing how long the book had been in print.

Ok, so the book was ancient. However, we were clearly being taught how
to pass the buck on responsibility for mistakes. Over-dimensioning
introduces opportunities to get something wrong, so it is strongly
discouraged (IMHO) to let the other guy/company be the one who absorbs
the cost of any mistakes.


By working to nominal
dimensions at all times, you are able to use any portion of a machined piece
as the datum point. That's important when you start changing reference
points.

Understood. I ran into this just yesterday, having taken ten thou too
much off of something. The good news is that I picked a good dimension
for the mistake, but it lead me to have to start thinking about what
could/should and must not move as a result. Basically the parts are
rectangles with holes near the corners and a window (weight reduction)
inboard of the holes. The hole spacing has to match another part, so I
left it alone, and slid the window back ten thou in the affected direction.

One could argue that I should simply eye-ball the window, but the next
parts won't be so forgiving, and I (clearlyg) need the practice
holding tolerances. So, I try to build what I design down to a thou or
two, and use common sense in what I scrap. Of course, I made the
mistake on a stack of five plates, and after all other block dimensions
were correct - too much time and metal to toss over a matter of pride.

As for how I blew it, I had noted I was about 0.050" from final size, so
the plan was to take 0.020 twice and then measure again. I was
flycutting a stack, which is not the best way to hold tolerances on my
mill-drill, but if I stay on top of it, it works. Perhaps my real
mistake was to be emailing and making phone calls between cuts, but
somewhere either something slipped or I made a cut I forgot about (I'm
starting to think that is what happened). At one point, I was expecting
to have to take off ten, but measured 20+ to go. I measured, looked for
burrs and chips, re-measured a couple of times. Shortly after that I
measured again and found it undersized.

Bill