The basic idea, which came from a thread on a machining forum which I can no longer seem to find, is this: Put a single-point cutter in a fly cutter. Angle the head of the mill such that the cutter points straight down. Cut a tooth, then feed by the pitch of the rack. The forum-poster thought it would create a slightly incorrect tooth, because of the angle of the fly cutter.
Turns out, though, if the mill is angled between 90 degrees and 90 minus the pressure angle of the gear, the teeth actually come out perfectly shaped. Here's why:
This is what a single point cutter looks like when it has been swept around the axis of the mill spindle. In this case, the mill is angled to 65 degrees, and the cutter is for a 20 degree pressure angle rack.
The tooth profile that would be cut in this configuration is shown below. Since the mill is angled to 25 degrees from horizontal, but the pressure angle is 20 degrees, the cutter removes an extra 5 degrees from one side of each tooth. Simply rotating the head of the mill to between 70 degrees and 90 degrees means that the plane the tip of the cutter moves in doesn't actually interfere with the shape of the tooth being cut. The result is a normal tooth shape, despite the funny arc the cutter moves in.
Here's what it looks like in practice:
I machined a .1" pitch rack out of some aluminum to make sure the process worked:
On to the actual rack, which is approximately a .5 module metric pitch. The cutter was ground from a small brazed carbide lathe tool. The rack was machined in three passes, to avoid loading the tool too much. This took a very, very long time.
But I'm quite pleased with the results. The rack feels like it meshes quite nicely with a (presumably) .5 module gear pulled off an old stepper motor.
Coming soon, making shiny hand wheels, and machining the spindle and headstock.