Seems to me that 8 passes is a few more than necessary. A couple of
additional tips might reduce the number of passes.

First though, saws come in all conditions, from pretty good teeth
to weirdly shaped teeth. A saw that is in really bad shape may
take 8 passes.

paul womack wrote:
> At the earliest stage (topping and jointing) you might also want to
> check (with a straight edge) the overall shape of the edge, and adjust
> it to either

A quick look down the teeth from either end after jointing will
reveal problems.

> I found that judging "removing half the flat" on each of the 2 passes
> very difficult.

The goal is to have equal sized teeth - equal sized gullets. If you are
starting with a regularly toothed saw, removing half the flat works. I
check the gullets on either side of the current gullet on the first
pass. If this gullet is smaller, I remove more than half the flats. If
the gullet on the left is smaller, I work the right tooth more than the
left tooth. If this is the largest of the three gullets, I remove as
little as possible while still reaching the tip of both teeth. Coming
back, I normally remove the rest of the flats unless I see a gullet size
problem that I missed the first direction that requires that I leave
some flat to allow enlarging a neighbouring gullet.

> Observing exactly when the flat disappears can be tricky. A particular
> problem is that as the tiny last vestige of a flat disappears, a tiny
> filing burr *APPEARS*.

This happens on the tooth pointing away from you as you file - a wire
edge appears that reflects the light. I brush the back of that tooth to
remove this wire edge. Of course, this wire edge can be sharp.

> If your saw vice is not perfectly even, some parts of you saw will be
> better held than others.

Masking tape, a couple of layers, improves the grip on the saw and
reduces the vibration slightly.

> I have settled on a procedure of applying the set to the saw just
> before the final 2 passes (of my multi-pass approach).

Most authorities recommend setting before filing (Pete does it after). I
joint, then set, then file. I set fairly hard (described as "crushing
the grain of the steel, so there can be no recoil" in Charles L.
Johnson's saw manual). It seems wrong to do this after filing.

> the problem was the set - when the set has been applied the original
> flats are no longer co-planar, which make seeing them difficult.

A really bad saw might benefit from two complete filings (joint, set,
file) rather than a larger number of filing passes. For a pretty good
saw (resharpening) this should not be a problem since almost all of the
teeth should be almost correctly set.

One small additional point on filing -- the direction of work. Charles
Johnson is very specific: file so the feather edge is in the direction
of set on the front of the tooth. The Disston Lumberman's manual is also
quite specific - both sides toe to heel with the tip of the file
pointing toward the heel, arguing that this avoids the feather edge on
the crucial front of the tooth.

Has anyone noticed a specific benefit of either strategy?

Brent