geebee509 wrote:
I have a Starrett precison level I think is not perfectly flat on the
bottom.
What is the proper procedure for checking the flatness of a part using
a surface plate? For example:
Do you coat the part with Prussian Blue? How, a finger tip, what
thickness?
How do you test it on the surface plate - press it down gently, slide
it, rotate it, what?
How would you prepare the surface plate - clean with alcohol, use a
damp kem wipe, what?
Do you test flatness on different sections of the surface plate?

I hope these are not to basic. Any other hints, tips, cautions, would
be sincerely appreciated.

Gil


Put a tiny speck of prussian blue on the surface plate after a very
thorough cleaning. Wipe the plate 3 times with your fingers to detect
any speck of dust before applying the dye. If you can find a "speedball
brayer" (used in the printing and silkscreen industries) this is the
best thing ever to spread the dye evenly.

The coating will be about .0001" thick for this type of work!
The part to be tested has to be completely clean of dust specs,
again, use your fingertips to make a last test of the part before
placing it on the plate. Place the part very gently on the plate,
do NOT press down at all, and move it sideways by pressing against
the bottom. You want to slide it, not rock it. Then lift off from
one corner, or just straight up for small items like this.

You can also test the part on the clean part of the plate (no dye
needed) with the "spin test". Place the part and gently turn it
from one end. If it pivots around the far end, the part is concave,
which is good for scraping. If it pivots around the middle, it is
convex, and will rock and give a false reading when tested with dye,
unless you take precautions to prevent it.

I have some chemical surface plate cleaner. I think it is mostly
detergent and water, but I don't know exactly. I suspect alcohol will
work fine. Watch out for all types of wipes, they leave lint. You have
to remove the lint after the plate dries. A tiny speck of grit or lint
will completely foul up a reading of this type.

If your surface plate is in good condition, you don't need to test at
different locations. Most granite surface plates, even cheap ones, are
incredibly flat. You have to test diagonally across the entire plate to
find any error of significance. Cast Iron plates, on the other hand,
can warp and wear over time, and can't be trusted until you have
verified their accuracy.

Jon