On Fri, 28 May 2004 19:53:09 -0700, email@example.com wrote:
I'm going to put this up soon.
My home DSL line is a test / development line, usually unrestricted
at 6-8 megs. Unfortunately, tests of the latest stuff aren't going so
well, so I'm stuck at 21.6 via dial up. 8^(
As soon as the DSL line is dependable, I'll add your ideas.
I'd venture that most of my crosscuts on a sled are less than the width of
the blade and wouldn't even engage a splitter. There are exceptions of
course, like when crosscutting panels, but generally those are done on
stable materials that are not likely to close up on the blade, and anything
laying on the table of the sled is basically stationary from the surface to
surface friction since it is the sled that's moving, not the material.
IME with a sled, much of the hazard to precision can to come from pulling
the cut piece back though the blade after the cut is made. On critical
pieces, I don't even attempt to pull the sled back to the starting position
until the blade stops on each individual cut ... too many times just the
slight kiss of the blade on the drawback (either from the inevitable slop in
the runners, or slight movement of the part) is enough to lose the precision
of the sled.
...and, if you're like me, don't forget to put the splitter back on when you
take the sled off.
The error in "squareness" is based upon a defect in the angle of the cut
relative to the side of the board that is supposed to be perpendicular to
the cut in order for it to be "square" (as in a right triangle). As I am
getting this, when we take the side of the board parallel to the rail of the
table and the one that the blade engages is taken as the reference line from
which the kerf is supposed to be absolutely perpendicular in order for the
cut to be "square," then the error on the other side of the board (the one
on the rail side) is such that the deviation from perpendicularity is only
0.0015" over a board 5" wide?
Side A of board
____________________________| Vertex B
Side B of board
When sawn on a perfectly aligned machine (one that can exist only in the
imagination) then Vertex A = Vertex B = 90 degrees iff Side A is parallel to
Side B (not demonstrated by the original poster) and the right side of the
board is "square".
But if the alignment is not "perfect", then the angles of Vertices A and B
would not be 90 degrees and the end would not be square. So, is the
gentleman saying that the deviation from perfect square across a 5" distance
between *lines* A and B is 0.0015"? If that is the case, then we have:
Side A of Board
____________________________ Vertex A
_________________________ /__| *Vertex B
Side B of Board C D
Highly exaggerated, of course but *Vertex B (also Point D) would be the
perfectly aligned position. However, due to the error, C becomes the Vertex
B and the distance from C to D is 0.0015". Is that what you understand this
Since the distance A to D is 5" and the distance C to D is 0.0015", we can
find the angle CAD from:
Tan CAD = 0.0015/5 = 0.0003 and angle CAD = Around 2 seconds of arc!
WOW!!!! I doubt that this degree of error would ever be worried about by
any woodworker - if that is what the original poster meant by error over 5".
Hell, I'd bet a damn good roofing square is off by at least that much.
You've got to be a good distance away from Vertex A before such a small
error would be noticed. Consider D to be 5 miles away from Vertex A. Then
C would be out of place by 0.0015 miles and that comes out to be a
misplacement of only 7.92 feet!
So, I ask again, how was this error measured? It seems to me that just
tightening a tape measure could compress wood by that much at the
application point... then you've got the error of the tape measure and the
error of your eye and the error of the changing water content of the board
and some other errors thrown in. I'd say it's close enough for government
work ... IN THE OLD SENSE!!!
Oh, yeah. Note that in the second ASCII Art attempt, the misalignment could
have been rotated the other way but the 2 seconds of arc would be the same
since the ideal vertex would form the midpoint of the base of an isosceles
OK up to your last number. The tan of 0.0015/5 =0.0003 = 62 arc
seconds. Still a very small number. It is fairly reasonable to measure
this . Take a 4" long board and cut it in the middle. Now flip one
piece over and put the two cut edges together. The angle error is
doubled. If you push one piece against a straight edge the other piece
will deviate from the straight edge by 2 X 0.0015/5 X 24" or 0.014".
This is still a small number but it is within reach to measure that
yep, what both Bruce and AgkiS said. That is exactly what I was
As further clarification, My only concern was with side B of the board,
a piece that had been jointed flat only moments before performing these
What I meant.
As I responded in another post, I used a machinist square and feeler
gauge. Actually, the error is slightly under 0.0015, but my feeler gauge
set doesn't have anything smaller. I'm certainly hoping this is
sufficiently good such that only slight planing or sanding will be required
for any boxes built using the Leigh jig.
BTW, the error was in the other direction, but, as you say, the effect is
the same, I just had to apply the mallet in the opposite direction :-)
How: Machinist's square and feeler gauge
Why: Take four pieces that are out of square by 1/32" (or even 1/64" --
normal high-precision wood measurements), assemble into a frame. How
well-crafted will that frame appear? Just because it doesn't make sense to
measure some things to better than 1/32 or 1/64 (or even 1/16) -- there are
other times when getting something cut square (or mitered at 45 degrees) to
high precision is essential. Another example, take 4 pieces of wood and
dovetail all ends to assemble into a box. If all sides of the box are out
of square by 1/32, how well will those dovetails fit?
I grant you that careful measurements are required for
any woodworking but when I hear folks talking about
somthing that's .0005 out of square, that's getting a
I own a LOT of measuring devices and I believe I only
have one ruler that even shows 1/32" marks, which I can
barely see, little less make a correct mark to.
I do have a Incra ruler my wife bought me several years
ago that has 1/64" and I can say without a shadow of
doubt, I can NOT measure anything using that ruler.
I do all the normal joinery including a few dovetails
and as a general rule, most of the joints are tight and
I have seen these silly "measurement" things for several
years here on the rec and I think it really must make
people crazy cause they can't measure and cut a board at
some of these "extreme" tolerances.
I go to GREAT lengths to "not measure" if at all possible.
I use story sticks and templates if at all possible.
Mark & Juanita wrote:
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.