Templates creating matching curves

[Dan Lee]

I'm working on a design for a sideboard for our dining room. The top I'm considering will have a curve to the front. The interior will be veneered plywood (maybe tamu ash) with a mahogany frame.

Question is how to get the curve on the frame and veneered ply to match exactly?
For curved pieces I normally make a template by laying out the arc, roughing out with a bandsaw then finalize with rasps and sanding. So if I make a template for either the convex or concave piece how can I use that as a guide for creating the matching template. Everything I've thought of will change the arc of the matching piece.

Any other methods?
Thanks
Dan

[David DeCristoforo]

Use a router and straight bit mounted on a trammel to cut the matching curves.

[Jamie Buxton]

One way to do it is with a straight bit equipped with two differents top-mounted bearings. For instance, use a 3/8" diameter bit with a 1/2" shank. Put a 3/4" OD, 1/2" ID bearing on the shank. Run the bearing against your master template to cut one arc. Now replace the 3/4" bearing with a 1 1/2" bearing. Run this larger bearing against the master template, and the 3/8" bit is moved out exactly 3/8". You can get a variety of bearings at many sites which sell bits. www.eagleamerica.com is one source. Sometimes the larger bearings are called rub collars, and you may find them near the 1/2" shaft shaper cutters.

[Doug Shepard]

This came up a couple months ago and I think you could use a shop made router edge guide modification like the OP came up with to trace the first template and offset it by the exact distance of the router bit diameter to make the 2nd template.
http://www.sawmillcreek.org/showthread.php?t=89950
Or you could just get a MicroFence edge guide:D
http://www.microfence.com/

[Dan Lee]

I should have mentioned the arc is 60" wide with a 2" rise and according to the handy online calculator is an 18 foot radius:eek: Seems like getting the required accuracy for the arc of the mating pieces will be difficult using a trammel

Doug
I'll reread that thread but I think he was trying to get accurately offset arcs, which have different radii,here I need the 2 templates to have the same radii one concav and one convex.

Jamie
I need to digest that sounds like it'll work

My brain has problems wrapping it self around this type of problem. Last time I did something similar one piece was fitting into a curved slot so accuracy wasn't critical.

Thanks

[david kramer]

Dan,

I made perfectly matching templates for this toy box: http://www.sawmillcreek.org/showthread.php?p=844928#poststop

These were much more complex curves than you're proposing, so it can be done! I did it using bushings and the compensating offsets technique that others have described here.

It's a bit hard to explain, but after you've done it once it'll make sense. I got the idea from Hylton and Matlack's book "Woodworking with the Router". They describe and illustrate the technique better than I ever could. You might want to check out if your local library has a copy.

David

[Larry Couture]

I'm in the middle of a coffee table design that has the same general curve as your project. There are several ways to make the templates but in essence they all typically require three templates (let's call them "A", "B" and "C"). The method that seems to be the most straight forward is as follows:

1. Make template "A". This is the template for the outside cut of your field - i.e. a convex curve. Any radius, any shape. Some will suggest you shape only half of this template then make a duplicate and join the two to create a symmetrical full template.

2. Make template "B". This template is a duplicate of "A" made with a pattern bit (straight bit with the bearing on top). The next step will reduce the radius of the curve by exactly 1/2", and will serve as the template for template "C" - the concave inside template. To do this, cut a 1/2" rabbet with a bearing guided rabbet bit. Then remove the rabbet ledge using the pattern bit.

3. Make template "C". Set up template "B" as the pattern and cut "C" using the pattern bit. This set up can be a bit tricky, but really only requires that "B" be raised off your bench higher than "C". I use one extra layer of MDF under "C" to keep the bit off the bench and two layers under "B". Clamp it all down well and run the router between "B" and "C". Don't forget that the router direction is dependent on "C", not "B", or the router may get away from you mid-cut.


So, that's how it's done. But here's the kicker...YOU DON"T NEED TO DO ANY OF THIS - for a radius as big as yours. I just did it this weekend and discovered that both "A" and "B" match "C" perfectly. Why? because for a large radius curve of 13+ feet the divergence in arc height is approximately 0.005".

If you want to calculate it yourself, here's the formula:

r=(Me2+1/4Ce2)/2M

where r=radius, M=arc height and C=cord.

For my project the cord is 50" and arc height is 2". My radius is ~13 feet.

A 1/2" reduction in a 13 foot radius changes the arch height by only ~0.005". It's no wonder the difference between the master and the 1/2" reduced outside curve templates BOTH mate perfectly with the inside curve template "C" - they are the same over the length of my table as far as woodworking tolerances are concerned. But they are mathematically different.

The arc height difference is even less with your 60" cord and 2" arc height.

Save you time. You do not have to make template "B".

The above three template method is critical if your radius is much smaller - example 12 INCHES - but pointless when your radius is in feet.

[Charlie Plesums]

I make the inside and outside arcs on a piece of MDF about 8-10 inches wide, so I have room to attach to the work piece. I think the difference in height is far more than .005" but I have been surprised how forgiving the wood is... at 50 inches long you can bend it quite a way to meet, without undue pressure on the glue joint.

http://plesums.com/wood/bedroom/CherryHeadboard4676.jpg

[Dan Lee]

Larry thanks for that description. Hyltons method in the book David referenced uses a bits with oversized bearings and yours uses a rabbiting bit with under sized bearing. Its great to have options:)

I was wondering about the how much offset there would be given the large radius of the arc. Think I'll test fit before running the additional templates