How to build a router table (only 9 steps)

I needed to round over some corners for a thing I’m working on, but I have no router table.

I needed a router table.

Here are the instructions: https://youtu.be/N7FfROEw3-w

9 Likes

I saw this “trick” a couple of years ago in a YouTube video. He even made and used a fence. It looked a bit on the dangerous side, but it is probably not much more dangerous than a normal router table.

The router table was a success, but the overall task was a failure.

Here’s the idea. Suppose I have an endgrain board that has cracks, and I want to patch them. I could do V-carve inlay with a shape that covers it, but I thought how about just a rounded rectangle pocket cut with a straight bit, and rounded rectangle plugs that are slices from a stick.

If the stick is given a known arc at each corner (with the router table and roundover bit), I could define a matching toolpath for the pocket by measuring the width and thickness. I can dimension the pocket to match the plugs, rather than the other way around. And once the toolpath is defined, I can reuse it several times. Maybe not just for repairing cracks but also for decoration.

Getting a close fit is not trivial, so idea number 2, I can define a toolpath slightly undersize, and make successive passes that shave off a tiny bit, then move out of the way and pause, and then shave off a bit more, repeatedly enlarging the pocket. I can test fit my plug after each cut and stop the job once it fits. If the increments are small, it should be a really good fit.

What I discovered is that the roundover bit does not leave behind a circular arc, or at least not one that is tangent with both surfaces. I’m sure it’s fine for a smooth, comfortable edge on a table or something, but it does not produce the precise cross section I was looking for. Maybe I could have modeled it, but that sorta defeats the purpose of having a simple plug defined by width, length, and corner radius. I also considered trying to refine the shape of the plugs to be a true circular arc at each corner, but I could not think of a method that did not involve some real work.

Next best method might be back to V-carve inlay. Or maybe I can still use a straight bit and incremental pocket embiggening, just cutting the plug on the CNC. I would have to slice the plug first, then cut out the perimeter of the plug, which can be whatever shape, and then incrementally increase the size of the pocket until the plug fits.

I know, I’m over-thinking it. (But isn’t that the fun part?)

Or… Take some sandpaper to the outside edges of your plug… “Sand to fit” is a thing… Probably easier to do that and you’ll get a better edge if you use a good sanding block on the flat edges.

1 Like

Sure, but now you are talking about actual work. :smile:

I went through with a test piece just to see how bad it is, and I can get a nice fit on the four flat sides, but the misshapen corners prevent it from fitting.

I tried making a 3d printed sanding block with a 1/4" radius inside corner to try to fix the corner shape, but it was not going well. The paper wouldn’t stay put and it was uneven along the length of the bar.

Although now that I say that, maybe I could use spray adhesive and slice the bar into plugs first, rather than trying to shape the entire bar length. That might be tolerable, since there is not that much material to remove.

Why perform labor when you can engineer your way into out of trouble?

I think you might be reinventing the bowtie or butterfly inlay?

A butterfly with rounded corners would be easy to cut on the cnc. Even in endgrain. I think it would look great.

You can also do a pretty good job fine tuning a hole with a sharp chisel. There isn’t much art to it.

2 Likes

C’mon Jeffe, art is in the eye (and pocketbook) of the beholder!

beholder

3 Likes

Yes, bowties is what I was aiming for, just with a slightly different shape to enable zero hand tools.

Upon reflection, maybe the bar with not-quite-perfect corners still has a usable methodology.

If I assume the corner is indeed an arc of the nominal radius, but displaced outward and upward by an unknown amount, perhaps I can make a gauge to measure those displacements. Then I can model the shape after all with just a couple more scalars (let’s suppose I’m not willing to do a curve trace).

The shape would be defined by: width, thickness, arc radius, arc displacement error (width direction), and arc displacement error (thickness direction).
This assumes all four edges have the same profile, which seems reasonable.

I will try it tonight.

It takes a lot of skill to achieve this with minimal skill. (My approach to artwork in general, not bowties necessarily.)

A slight bit generous in the corners, so my offsets were evidently a bit off, but the incremental sizing worked nicely for the four sides. The best fit occurred at 0.2 mm offset from the dimensions that were measured.

Getting there.

2 Likes

Here is how I should have estimated the roundover error.

This is indicating about 0.1 axial and 0.3 radial offset, whereas my previous attempt had gone with 0.4 x 0.4 (too much).

At 0.4 x 0.4 this gauge predicts the corner gap I saw with the test cut, which is pretty reassuring.

2 Likes

2D printers are just so accurate :slight_smile:

2 Likes

Seriously, where’s the T-square and architect’s rule? The compass and protractor? No drawer of thin sheet metal templates? Feh… Amateur… Bet his eraser is even pink…

:wink: :japanese_goblin:

With the better model of the corner, the test fit was pretty good:

This was still with the incremental embiggening method, just a better corner shape.

Good enough for the cutting board patches (five total), here is the dry fit:

It’s gluing now and tomorrow we’ll see. I’m pretty sure it will be at least okay, maybe it will be great.

Incidentally, to get a good comfortable fit without having to force it, the maple cutting board went one more pass than the MDF test (meaning an extra 0.05 mm offset). Without a methodology that adapts, that would be a real pain.

2 Likes

I had somewhat high expectations but this is even better than I had hoped.

Here’s before:

And after:

And close up I can’t see a gap, even in the corners:

I would not have thought such a tight fit were possible with a straight bit, because it’s basically like chasing zeros and if you miss even slightly oversize, you’re sunk. I’m super happy with this outcome.

So to summarize
Step 1. Cut a bar with 2 straight sides on table saw and two flat and parallel faces on planer (dimensions need not be precise)
Step 2. Use machine with roundover bit to round all 4 long edges
Step 3. Measure bar width and thickness with calipers and then slice into a bunch of plugs
Step 4. Use OpenSCAD script to generate corner gauge and print it out to measure radial and axial errors in the roundover shape
Step 5. Use OpenSCAD model to generate pocket shape with measured dimensions and rounded corner errors and add several concentric offset curves too
Step 6. Load into EstlCam and create pocket for innermost curve and successively larger contours (optionally, use lead-in to avoid “dent” artifact where plunge occurs)
Step 7. Generate gcode and then manually insert gcode between concentric curves to move the tool out of the way and pause ("@pause" in Octoprint)
Step 8. Run the job and test fit at each pause, and stop when the plug fits
Step 9. Glue plugs in place and surface entire board after glue dries (dovetail is my preferred bit).

There. Nine steps as promised.

5 Likes