[TheForge] Effective filing WAS: Re: [kl] Re: Sen

[Andrew Vida]

I've copied theforge, the blacksmiths, and artmetal lists as this is 
relevant to each.

Cameron Potter wrote:
> Hi all,
> 
> I have done a fair amount of drawfiling (but end up with very large 
> scratches which take forever to get out).

	This usually occurs because you are pushing the tool beyond its cutting 
capacity.  If your file is in good condition, is of good quality, is 
clean, and is being used correctly, this should virtually never happen. 
  Think of the cutting design of a typical file: in microcosm you have a 
cutting edge preceded and followed by a gullet and another cutting edge. 
  The gullet is there not only to lend form to the edge, but to convey 
chips away from the work.

         .    .    .
         |\   |\   |\
         | \  | \  | \<---tooth
         |  \ |  \ |  \
     ____|   \|   \|   \____

	If your gullet depth is, say, 0.005" then it is futile to attempt to 
press the file any deeper into the metal than that.  In fact, the 
cutting capacity of the file will be far less because at 0.005" in this 
case, the gullet would be filled completely and there would be no place 
for the chips to go assuming that you even has the strength to mash the 
tool that deeply into a piece of steel, and without breaking it no less. 
  I have no idea what a typical maximum depth would be, but I would be 
almost willing to bet that the value would be no more than 30% and 
probably considerably less, or 0.0015 +/- in this case.  This will vary 
with not only the pitch of the file, but the overall geometry of the 
cutting edges.

	When you press beyond the file's capacity, the gullets load up and 
metal chips may even crown the cutting edges in places, lifting the file 
off the work in spots.  Edges that make no contact with the work cannot 
do any cutting. I suspect that there might be some microscopic welding 
going on between chips and cutting edges, which may explain why at times 
chips can be so difficult to clean from the file.

There are several things that may contribute to surface marring.  One 
will be the uneven orientation of the file to the work as it lifts in 
places.  Another will be the rubbing of the embedded chips against the 
work, which may be exacerbated by work hardening of the chip from both 
cutting and rubbing, depending on the material being filed.

There are techniques for mitigating this problem, such as loading the 
file with chalk on a regular basis to help prevent chips from loading 
up.  This is a good practice, but the best is to learn how far the tool 
can be pushed before it begins to retain chips, and then make it a habit 
of working it well below that level.  There is no set formula for this 
that I know of.  The only way to learn is to gain the proper feel, which 
in turn can only be acquired by actual filing.  When a file is cutting 
optimally, one can tell through the feedback to the hands.  The tool 
takes on a certain feel that is unmistakable, neither skating nor 
dragging belaboredly over the work.  It feels "just right" and it will 
be immediately noticed that metal is removed at an impressive and steady 
rate while maintaining precise control.  Furthermore, the moment a file 
begins to load up, one is able to immediately tell because the feel of 
the stroke changes significantly.  Most of the time one can just tell 
when the surface of your work is galling.

There was a time when I was a kick-ass filer and could easily file to 
tolerances of 0.001" without even thinking about it, and even finer when 
I started to concentrate.  Also, when properly used, files can produce 
an almost arbitrarily fine finish on work.  I have seen and even done 
work that was so well executed that you would not be able to readily 
tell that a file was used instead of, say, a very fine abrasives.  A #8 
file will literally polish a surface bright as if with a burnisher.  The 
human hand is the most amazing tool we possess in terms of our 
metalworking.  It is unmatched by any other.

If you have good files, treat them like gold.  Keep them clean, dry, and 
store them separately so they don't bang against each other.  My fine 
metal files are never used for ferrous work.

 > A sen sounds like a perfect
> solution, however, I can't seem to get a clean idea - even from Don 
> Fogg's site.  I just can't see how it would work...

	Scraping, essentially.  Have you ever used a scraper to level a 
surface?  The action is very similar to that though it can be less 
subtle when roughing.
> 
> Is it simply a convex edged plane that you can push or pull?

	Essentially, yes.  The chord depth of the convex edge should be very 
shallow, measured in thousandths of an inch.  If your sen's edges are, 
say, 2" wide, I don't think I'd give it more than perhaps 0.010" to 
0.020" of a belly. What you are trying to do is have a geometry where 
the cutting force can be precisely targeted to a very specific area.  A 
straight edge will not do this as any deviation from perfect compliance 
to a flat surface will result in cutting occurring at either of the 
edges of the work and not in the middle.  With an ever so slight 
convexity, one can "aim" the cutting action with extreme precision over 
any region of the work surface and this aim can be maintained by both 
visual and tactile feedback.  The reduced contact area will also 
concentrate energy in a way that will make for efficient cutting with 
little or no rubbing, which produces no removal of metal.

Also, convexity can be built into the bottom of the sen such that you 
can hold it dead flat against the work.  I've not tried this 
configuration yet, but it seems worth a try.  I don't know if Japanese 
smiths ever did this.

Another possibility (and again for all I know this is common practice) 
would be to have more than one sen, one having convex and relatively 
aggressive edges and another with dead straight ones that will bite less 
and produce very precise results.  Just a thought.

 > How do you
> stop it from digging in, if it doesn't have a tendency to dig in, how do 
> you get it to bite...

	Of the sen I've made, this hasn't been much of a problem.  The set 
angle for cutting should be very shallow.  If it is not, you may not 
have good cutting geometry in the tool and should consider making 
alterations.  It is also possible your steel is too soft.  I was 
surprised to read on Don's site that he grinds a rather acute angle on 
his sen.  I've made mine anywhere from about 80* down to 60, but not 
less than that.  I always thought less would develop more tendency to 
dig ruts, but that's probably more a function of my lack of skill 
compared to someone like Don who does this for a living.

Don also mentions cutting in both directions, which makes sense to me, 
but AFAIK (and I may be wrong about it) Japanese smiths cut only on the 
push stroke, the opposite of the way in which a westerner uses a draw knife.

Also, if you start digging a rut, which will happen if you hold the sen 
at too severe an angle to the work and/or press too hard, stop 
IMMEDIATELY.  Don't even let a second stroke trace the same path because 
you will probably do nothing better than make the rut that much deeper. 
   When you become an accomplished shaver, you may be able to correct 
small digs by altering the pressure, but until them try this:  The 
remedy is to either take a file to the area and smooth it (the easier 
method, which may have a drawback or two, depending on how good you are 
with a file) or alter the angle at which the sen traverses the work. 
Usually you will hold the sen square to the line of cutting action, 
which is directly away or towards one's body.  If a rut is dug into the 
steel directly across the work, alter the angle if the sen such that one 
hand is further forward than the other.  The deviation doesn't have to 
be extreme; say, 20*, give or take.  Then carefully work the area until 
the rut is removed and resume work at the more comfortable posture.


         |          |
         |          |
         |  WORK    |
         |          |
    _____|__________|_____

           SEN

    _______________________
         |          |
         |          |
         |          |

	Normal orientation


	
         |          |
         |          |
         |          |
         |          |
         |----------|<-- rut
         |          |
         |          |
         |          |
         |          |

            Oops...

         |          |
         |          |
         |          |./        ./
         |         ./        ./
         |-------./        ./
         |     ./        ./
         |   ./   SEN  ./
         | ./        ./       ^
         ./        ./         |
       ./        ./ |  Cutting direction
     ./        ./   |
             ./     |
           ./       |
         ./         |
       ./|          |

	Draw at angle to avoid
	compounding the error.