View Full Version : Fun with a file

tony ennis
01-03-2009, 02:18 AM
Slow day here today. I decided to do some metalwork. Not that I have a shop. I have a project I want to do that doesn't require much. But is does require bench. I need to make one. I have some Home Despot bar stock. I don't have a scribe. It's on order however. And I have 3 mismatched files. I have a set on order. So I'm pretty limited right now.

I ended up clamping my vise into a workmate. I hacksawed (note to self: get a good hacksaw) two 5" pieces of 1.5" x 3/16" bar stock. The stock has rounded edges. I decided I would flatten and smooth an edge on each such that when put together no light would shine between them. I can do it in wood with a hand plane...

In short, I failed. I could not get them flat. When gently c-clamped edge-to edge, I measured something like an 8 thou gap at one end and maybe a thou at the other. The edges were convex along their length. Of course, filing them at the same time doubles that sort of error.

So, how DO you flatten with a file? After the edges were shiny, I painted them with a magic marker and carefully filed, keeping as much of the file in contact with the work as I could, until the marker was worn away. The marker disappeared but the edges didn't end up flat.

01-03-2009, 02:46 AM
I think you would want to compare each against a reference flat and not with one another. see scraping or "way scraping" (which I have never done)

but to do the best with what you have,
I would file lengthwise and use the flattness of the file to flatten the edges

perhaps clamp the file and carefully run the work over it

01-03-2009, 02:54 AM
Are you familiar with draw filing? An application of a smooth file at an angle across the work piece will produce a very fine, flat finish. You have to be very careful how you work the file, though. Because your filing over a small area, and because of where your hands are placed, and because of your body's motion (or the file's tendency to grab at the corners), you tend to put extra pressure when approaching and leaving the work piece. A convex shape is the result. I do it all the time :D Draw filing will lessen this error, somewhat, but you still need to be aware of why its happening to make sure you don't start doing it again.

If it needs to be very flat, I keep a refrence handy. A 1-2-3 block works nicely. I use it to check the work after every few strokes to make sure I'm not starting to make it convex.

Hope you get your shop put together soon! :)

01-03-2009, 03:06 AM
Yep, that's how I would do it. How far you go depends on how flat you want it. But you either need 3 to cross ref, or a standard for reference. At the simplest "close enough for most work" level, I would use a good scale laid across and look at the points it touches. File them away. And use a small machinist square to keep the flat perpendicular with the sides. If you get close by this method, but not close enough, then a reference flat with transfer media would be my next stop. Just like scraping, but using a flat/fine/smooth single cut file. Also, holding/presentation technique is critical when going for better than typical results. Both "over hand" and draw filing might work well depending on setup and desired result.

For the final bits I often choose a Barrett with very fine pattern or even SC paper (like lapping) in the "over hand" position. Basically, get in a position like you are looking for light penetration, and place the file in position. Grasp at the root of the tang with your "strong hand" finger tips (generally thumb + index, perhaps middle), then cup over with your off hand to steady. No handle is used in this technique. If anything, and you just can't stand a bare tang, wrap with masking tape. Sight along the file (through the peep-holes formed by your fingers), and work in the normal fashion (light cut on push, light lift on return). Head position is such that if you continue the return stroke to extremes, you would poke the tang into your eye on your strong (handed) side. This provides very fine directed control and relative ease in visual confirmation of flat blade presentation. Die maker "file hands" use this technique to split tenths (and polish) on critical bearing surfaces. That's where I learned it while working as a draftsman/designer in an extrusion tooling shop about 20 years ago...

And of course, you could eventually proceed to scraping to go still further. Or you could lap it. But you asked about files, and that's how I would do it...

BTW, once both are flat and have a fine finish, the light blocking takes care of itself. Matching the two together without necessarily being flat is a different matter and can be achieved by getting close and then lapping only.

01-03-2009, 08:53 AM
just keep inspecting it with a square, and remove material where its high. you don't have to do a full on power stroke, you can control where it removes material by where the weight is (on the file). draw filing is also good.

remember when you were a kid and you tried to cut something with your dad's saw? It was horrendous; the saw wobbled and got caught, it seemed like it was binding once or twice per stroke it was more of a fight than pleasurable pass time. Now if you worked at it, you eventually got so you could grab a piece of hardwood, scribe a line and hit it with the saw and it comes out perfectly. How? you learned by concentrating on the motion of the saw to make your arm propel it back and forth in the same plane, not the curve motion the arm naturally takes.

that's the trick. concentrate on the motion of the file and keeping it in the same plane. the eye is very good at judging flat and square so its not a preposterous an idea as the freshly frustrated filer might think. have a little square at hand so you can check many times, that feedback will help you correct, it doesn't take too much practice to get it but it is a skill you have work at a bit.

01-03-2009, 09:09 AM
Assuming that the rounded edge you wish to file flat and square is straight to begin with, sometimes it is helpful to file a slight 45 degree chamfer on the corner edge parallel to the surface being filed. This will serve as a visual reference as you remove the original surface.

01-03-2009, 12:35 PM
First...make sure your file is actually flat...not all of them are.
Make sure you push down on the middle of the file.
I've built a lot of things with files.
You can do really nice work with them if you are patient and learn how.
Good quality, sharp files will make your life easier. (Note..I did NOT say easy)

Your Old Dog
01-03-2009, 12:43 PM
I agree with Torker. Any time spent with a file is time well spent if you want it to be. There is much to learn and the knowledge is a great asset to have in the tool arsenal. A lot of famous knifemakers got that way with doing edge file work on their knives turning them into art.

Paul Alciatore
01-03-2009, 04:00 PM
Some thoughts on this:

The main thing I would suggest is try holding the work piece by hand and move it, not the file. It is very hard to move a file in a completely straight line. If the work is held fast in a vise, then it will not move. But when you move the file over it, the file will change orientation by a few degrees as you stroke it. Therefore, it will cut on the near side part of the time and on the far side the other part. This will leave the center as the high point.

Hold the file in your left hand if you are right handed. You can brace the far end on the bench or WorkMate if you like, but basically the file stays still. Hold the work in your right hand (left if left handed) and hold it near the center of mass if possible. Allow the work to orient itself with reference to the file's surface and take light to medium strokes. While working on a single side, rotate the work in your hand by 180 degrees every 5 or 10 strokes to even out any tendency to cut more on the leading or training edge. Check for flat and square frequently and adjust your technique as needed. Observe any tendency for a particular area to file down faster and rotate other areas to that position to even things out. This applies to relatively small parts.

I have observed that every "flat" file I have is slightly concave on one side and slightly convex on the other. Place a straight edge on yours and determing which is which so you can be aware of this while filing. I use the concave side for rough filing and the convex side for the final strokes to eliminate any rounding of the edges as much as possible.

On larger pieces where it is impractical to move them over the file, it is possible to hold the file in a manner where it will follow the existing surface instead of constantly creating a new one. The handle is held with a light grip to transmit the motion of the stroke only and pressure is applied to the center part of the file to hold it firmly down on the work and orient it. This central area where pressure is applied is ALWAYS kept over the work and a bit away from the edge so the file is oriented by the work. Use the convex side of the file so more stock will be removed at the area where you are applying pressure. Check for where the high spots are and carefully file them more than the lower spots to reduce them more. The use of the convex side helps this process. Test frequently.

When working with a file, you should clean the file frequently. I keep a small steel brush (toothbrush style) for this purpose and I brush side ways, along the grooves to remove the chips as they accumulate. I also find that it helps to apply a little oil to the file. This helps prevent the chips from jamming in the grooves and improves the surface finish of the work. Any light oil is good.

Draw filing is great for getting a better surface finish but it may be more difficult to control the flatness. I only use it for a few finishing strokes after the basic size and shape have been established or where precision are not necessary. I think it should be reserved for use after you learn the basics.

01-03-2009, 04:31 PM
To really get the feeling of the motion your hands should travel. Get a piece of flat bar somewhat the size of a file, open the vise jaws to about 3" and push the flat bar across the jaws maintaining contact with both jaws. It feels like the hands are moving in an arc away from you when in fact they are moving in a flat plane. Once you get the feeling in your hands and arms of the motion required then try filing, you will find you are a lot more successful at maintaining a flat surface. That is the method I used with my students in high school, seemed to work for them. Then draw file. Keep the file clean by pushing a piece of sharp edged flat bar across the file in line with the teeth. Peter

01-03-2009, 06:39 PM
This is kind of kin to how the first lathe was made. Starting with nothing and ending up with something perfectly flat, or straight, or perfectly timed in relation to distance travelled in response to something turned, as in a threaded rod- these accuracies are built up by using innovative means to manipulate the materials. Marking and scraping 3 angles to end up with all three perfectly angled in an example.

So what's an innovative way to come up with a perfectly straight line of reference- how about stretching a string? A very taut string will be straight, providing it isn't influenced by gravity, side forces, and irregularities in the string itself. A guitar string would probably be a good thing to start with, and a simple form could be used to tension it onto. A piece of 2x4 and a few screws would work fine.

Stand this upright, and the tensioned string will form a straight line. Any workpiece held barely touching this string will show its uneveness in the light passing between the string and the workpiece.

In practice, a tightly tensioned wire like this will work well as a reference in any position, not just upright, because the deflection caused by gravity will be much lower than the deviation from straightness that you would be trying to detect in your workpiece. So such a 'straightedge' could be a totally portable hand-held tool and perfectly suitable to detect deviations from 'perfect' of a thou or less over any reasonable distance. Leave the workpiece in the vice and alternately file and test until the deviations are within your required limits. Getting something straight within a thou or two over its length is not difficult, just takes a bit of time and care. Reference materials are not hard to find- feeler gauges, paper, cigarette papers, foil- you don't even need a caliper or indicator. By the way, I don't recall when I ever saw a file that was perfectly straight or flat, but they still can be used to create a very accurate edge.

An easy jig that can be used to help with the filing is simply a second surface set up behind the vice that one end of the file can slide on. This is made to be the same height as the edge of the workpiece clamped in the vice, and made parallel to the workpiece (conversely, the workpiece is carefully oriented to be parallel to this second surface before it's clamped tight in the vice). In use, the file is thus kept flat, and the edge being filed is kept flat without rounding.

You might think of a square as having straight edges, or so-called 'straight edges' as having actually straight edges, but it isn't often so in the real world unless the device has been prepared and tested to be within whatever the tolerances are spec'd at. This of course is one of the first things to consider- what are your required tolerances-

01-03-2009, 08:55 PM
I recently finished a bed clamp out of HRS with a hacksaw, files, and a belt/disc combo sander. After reading this thread I may have to revisit where it meets the bed.


The reason there's no screw in it is because I was all jazzed when I found it fit good enough to hang all by itself .... :D

With a screw in it clamps hard enough to slip the clutch, that's the object of the exercise, but I wonder if I marked it up what I would find?


tony ennis
01-05-2009, 12:31 PM

1. My files should arrive today :-) I padded the order to get free shipping (their scheme worked beautifully) with some inside and outside calipers, extra file handles, a scribe, some blue dye stuff, and a little rule.

2. It took me a day to realize why my right calf is sore.