Three tools and three cuts

Tue May 24 2011 20:24:58 GMT-0500 (CDT)

I'm sure you've heard the aphorism "measure twice, cut once."

I have trouble deciding whether that's just bad advice, or good advice presented badly. Yeah, there's a grain of truth in it -- a lot of work is lost to measurement errors -- but the 'cut once' part bugs me.

If I had to restate what I consider the acceptable part of that theory, it would go something like:

  • Make sure your marks are in the right places.
  • Then make sure you cut to the right mark.

Explaining why I dislike 'cut once' takes longer.

The basics of cutting tools:

First of all, let's talk about cutting tools in general. They all work on the same principle: converting some part of a block of material into chips that aren't connected to the block any more.

There's no such thing as a perfect cut though. There's always some error, and the surface will always be 'rough' at some scale. For reference, if you scaled the Earth down to the size of a cue ball -- ocean trenches, mountains and all -- it would be significantly smoother than an actual cue ball.

Just for fun, these images show what 'razor sharp' looks like at 10x and 200x magnification.

Every cutting tool has a range of sizes where it cuts best. You wouldn't want to chop down a tree with a straight razor or shave with a chainsaw, though I'm sure someone has tried to do each as a stunt.

As a rule of thumb, you can assume that a tool's best working range spans two orders of magnitude.. a chainsaw is good for cutting chunks between 1" and 100" off a tree trunk[1], a straight razor works well cutting hairs between .5" and .005". If you try to cut too small, you run into the limits of the cutting edge. If you try to cut too large, you run into the limits of the tool's strength.

[1] - Yeah, I know I'm fudging the concept of a 'cut' here.. the actual amount of material removed will be the width of the blade (probably around half an inch) regardless of the size of the pieces separated. There's still a range of sizes that are easiest to handle, and it's interesting how often those fall in the 'two orders of magnitude' range.

The basics of cutting:

Even if you're working within a tool's preferred range, the results you get depend on the part of the range where you're working. Cuts from the heavy end take more force, and will show greater evidence of the tool flexing. Cuts from the fine end take less force, but show more evidence of interaction between the cutting edge and the material.

For convenience, it's easiest to group a tool's behavior into three kinds of cuts:

  • Rough cuts -- anything heavier than 25% of the tool's range.
  • Cleanup cuts -- from 5% to 25%
  • Finish cuts -- anything finer than 5%

To drive that home with an example, the workhorse cutter on my lathe has a range from .001" to .1". Any cut that removes more than .025" of material from the workpiece is a rough cut. Any cut that removes .005" to .025" of material is a cleanup cut. Any cut that removes less than .005" of material is a finish cut.

To get the best results, you perform those three cuts in that order, allowing an error margin as big as the next finer cut.

Again, an example:

If I'm roughing with the tool described above, I want to shoot for a cut surface roughly .025" larger than the final dimension. I don't want to get closer than that though. I want to hog material out fast and ugly (there's no point in worrying about the surface finish of material you'll cut away in another minute or two), and that's easiest when I stay far enough away that even the ugliest cut can't damage the finished surface.

Once I'm within .025", I start caring about surface quality. Like I said above, a finish cut shows the effects of interaction between the tool and the workpiece, so major variations in the amount of material being removed will have a major effect on the cut. That's where the cleanup cut comes in. The goal here is to get a uniform surface so the finish cuts can run as smoothly as possible. This time I want to stop roughly .005" above the final dimension.

Now come the finishing cuts. This is where I start worrying about surface quality and dimensions. The good news is that I can creep up on both with sequence of shallow cuts that are too light to do significant damage.

Putting it all together:

So here's the secret of getting good results efficiently: use three tools, chosen so that a finish cut on one is a rough cut on the next.

For me, one such set of tools runs:

  • Chop saw
  • Lathe
  • Diamond sharpening stone

On the chop saw, I leave an allowance of .05" to .075". On the lathe, I stop about .002" above the final dimension. Then I take between .0005" and .0001" per pass on the stone until I've reached my final dimension.

I could take the sequence farther using a range of abrasives from pumice to washed emery, but my finest measuring gauges have increments of .0005", and below that you have to pay as much attention to the temperature of the workpiece as to your tools.

So, yeah.. to heck with cutting once. You'll get a lousy surface and only hit your target dimension by luck. Make a series of progressive cuts with the appropriate allowances. You'll get great results, and can be comfortably sloppy every step of the way. It's just that 'comfortably sloppy' on a diamond stone offers precision that would be miraculous on a chop saw.