jig for milling small pieces with the router

[poto]

A friend of mine commissioned me to make him a set of wooden puzzles (the payment of the commission was a new RTS400!). The puzzles were a Bedlam cube (13 pieces) and a Soma cube (7 pieces). They were to be made out of cocobolo, and based on 3/4 inch cubes. Each piece of each puzzle is different, and most are quite three-dimensional. So I had to figure out a way to make these pieces out of solid pieces of colobolo (I didn't want to just glue a bunch of cubes together), just using my limited set of tools (ATF55, MFT1080, OF1010).

So I came up with this jig. The requirements were that I had to be able to set the router position just once (depth and distance from the guide rail), and then never have to adjust it again. I wanted all the router movement to be along the guide rail. This required a jig that would hold the polygonal pieces in place while they were being routed. But the pieces had to be held and positioned very accurately and repeatably.

I discovered that all the pieces could be milled with the pieces held in one of two positions, that differed by exactly 3/4 inch in distance from the guide rail. So I built a jig that would hold the pieces and accurately register them for milling with the router.

First, the finished Soma cube. (I apologize in advance for the terrible pictures - my camera's flash was dying.)

[attachthumb=1]

The jig sits on the MFT1080, and is held down with Jorgensen clamps:

[attachthumb=2]

Veritas Wonder Dogs help to make micron adjustments perpendicular to the guide rail, and then hold the jig so it doesn't move side to side.

[attachthumb=3]

The jig is in two parts. The main part lies under the guide rail, and is exactly 1.5 inches tall. It has a wide surface to support the guide rail. The second part is an L-shape, with four threaded rods passing through it. The long arm of the L supports the foot from the router. Two pairs of threaded rods hold the pieces tightly to the main jig part during milling.

[attachthumb=4]

And here's what a piece looks like after the router has gone over it:

[attachthumb=5]

Notice that in the last two pictures, the pieces are held in the two positions, differing by exactly 3/4 of an inch perpendicular to the guide rail.

I found that with the jig I could get 1/10 mm (100 micron) accuracy in all three dimensions when milling the pieces. The trickiest part was getting the rectangular blanks to exactly the right size before milling. I did this with the ATF55 and the guide rail, the Wonder Dogs, and a lot of fussing with the blade teeth vs. the marked line.

 

[poto]

In that last picture you can see the guide rail lying on the main part of the jig. Here's what the router looks like as it's preparing to make a pass:

[attachthumb=1]

And here's a milled piece (not the same one as above) after the router's gone by (I vacuumed up the shavings!).

[attachthumb=2]

I'll admit to making lots of mistakes in practice pieces (I have lots of bits left over for making details on boxes and knobs, etc.). I also discovered that the router (with a 3/4 inch straight-sided double flute 1/4 inch shank rabbeting bit) left such a smooth cut that I didn't have to sand. In my first version, the sanding ruined the close tolerances of the pieces, and the puzzle looked a bit sloppy. Subsequent puzzles were much tighter.

This all probably would have been easier if I'd used a softer wood. But the cocobolo was so beautiful, and it makes a wonderful clinking sound as the pieces are put together.

I also inlaid brass rod in one of the puzzles, making brass dots on the pieces. When the puzzle's assembled the correct way, it has secret messages in code in the patterns  ??? :

Poto

 

[Michael Kellough]

That is really cool Poto! I love that kind of process.

I wonder though, since I don't know the forms,
if this wouldn't be a perfect job for the Jointmaker Pro?

 

[poto]

I think that it probably would be perfect for the JointMaker Pro. Unfortunately, mine was laid up in the shop that month. 

The trick would be to set the cut depth very precisely. Which, I gather from Roger's work, is not a big deal.

A router table or a dado blade on a table saw (neither of which I have) might also be good, but you'd need something pretty special to hold the pieces.

BTW, thanks for the compliments! 

 

[Eli]

Very cool Poto. Thanks for the pictures.

 

[johnsonri]

Great jig engineering Poto!! Now that you've done it once, I bet that if you had to do something similiar, you would get to your end-state quicker and with fewer test pieces.

 

[poto]

You're absolutely right on that, Rey. The first cube (Bedlam) took me several weeks, and a mess of mistakes. The second one took two days, with almost no mistakes. Also much higher tolerances. I think if I made them out of cherry or walnut, I could zing them off in a day, no probs.

Eli - thanks for the compliments. It was a fun intellectual problem, which is precisely what my "commissioner" wanted. He's really into puzzles and mathematical problems; turns out the jig was almost as big a problem as the puzzle!

 

[Warrior]

Really nice Poto! I am impressed. I am too clumsy for that.  That's why my puzzles are usually very large. Like door size. ;D

Eiji

 

[poto]

Yes - I guess the tolerances on doors are a little looser than a 1.5 inch cube!

I just discovered that you can flip the jig so it's on the right side of the guide rail (with the holding places facing the splinter guard) and use it to cut tenons.

At any rate, thanks for the compliments! 

 

[Bob Swenson]

You are a clever devil. You do a great Jig
Bob