radius joinery problem!

[humantorch]

Hello all,
I was hoping for a little guidance in calculating the inner radius for 2 pieces of live edge mesquite I'm attempting to join.
I figured out the outer (or top) piece and I know the second (bottom piece) is NOT the same I just can't seem to figure out what it is.
Any help would be greatly appreciated. [attachthumb=1]

 

[wow]

Can you just use a pattern clamped to the top of the boards just like you have then laid out, then run a router down the joint between the two pieces? Any irregularities will be repeated on both pieces and cancel each other out, making a perfect joint.

 

[Michael Kellough]

It looks like the two pieces are intended to meet and share the same (curved) line.
Therefore, the radius of that shared line must be the same for both pieces.

I assume you want to add a trammel beam to a router to mechanically create the exact same curve for each piece.

Use a straight router bit. Make a test cut by plowing a groove to find and record the exact diameter of the bit. With the trammel beam/router properly configured make the cut on one piece then increase or decrease the length of the trammel beam by the exact diameter of the bit and cut the next piece.

The picture shows that you already got pretty close. If the two pieces are a little miss-matched it's probably because you used the nominal diameter of the bit in setting the trammel beam rather than the true diameter. Or, the adjustment was a little off.

You could also make a curved sanding platen to help correct the existing fit. Make a convex platen by cutting a board with the trammel rig that is about the sandpaper thickness shy of the radius you want. Use a board about 3 feet long so it isn't too hard to handle. For thickness I'd make it substantially less than the real piece. If you try to sand both top and bottom edges at once there is a good chance you'll over-sand one edge. I'd excavate slightly deeper in the center of the concave board so it won't interfere with the fit. Then lightly sand the top and bottom edges individually until it fits well. I might also sand a little through the middle of the convex board too. This is the same principle as back-beveling a scribed edge but you only want to excavate a few thousandths of an inch.

 

[humantorch]

Thanks wow and Michael!
I neglected to mention that the top pice is done. I used a trammel and router to make a template out of mdf and then used a pattern bit to transfer to the slab.
I think I like your idea of using the true diameter of the bit a measurement. Just adjusting the trammel (original radius-bit diameter= new radius). This feels like it will work.
Wow, the reason I don't want to just try running the router down the 2 pieces is because of the density of the material and the chance of error. I feel more comfortable making a template and then testing till it's right. If it were thinner and less hard material I would have attempted just as you suggested.

 

[jonny round boy]

Can you just use a pattern clamped to the top of the boards just like you have then laid out, then run a router down the joint between the two pieces? Any irregularities will be repeated on both pieces and cancel each other out, making a perfect joint.

Wayne,

That method works when doing straight joints, but won't work on a curve. The two radii will never come out the same, they will always differ by the diameter of the router bit.

 

[wow]

Can you just use a pattern clamped to the top of the boards just like you have then laid out, then run a router down the joint between the two pieces? Any irregularities will be repeated on both pieces and cancel each other out, making a perfect joint.

Wayne,

That method works when doing straight joints, but won't work on a curve. The two radii will never come out the same, they will always differ by the diameter of the router bit.

Are you telling me that my spacial reasoning is flawed? It's not, you know. My procedure will work fine...

...you just need to use a small enough bit! 1mm bit = no more than 1mm error.

[embarassed] [big grin]

 

[Sparktrician]

I would think that if you flipped the slabs upside down and attached 3-4 strips across the joint using screws or double-sided tape, then flip the joined slabs right-side up, you could run a Carvex with a fine blade through the joint very carefully and get them quite tolerably close.  Once the strips have been cut, though, you'll have to replace them until you get to the last one.  Disclaimer: I have not tried this, but it seems to have possibilities.  YMMV.

 

[Stoli]

That method works when doing straight joints, but won't work on a curve. The two radii will never come out the same, they will always differ by the diameter of the router bit.

Are you telling me that my spacial reasoning is flawed? It's not, you know. My procedure will work fine...

...you just need to use a small enough bit! 1mm bit = no more than 1mm error.

It won't work as you desire, but this method can be used for a decorative inlay.  Check this link:

Here are some wine bottle and glass holders I made with this technique:

 

[wow]

Here are some wine bottle and glass holders I made with this technique:

I like those!

Did you make them for a restaurant? If so, how did they find out about you?

 

[Sparktrician]

That's a really neat technique. 

[smile]

 

[Stone Message]

Can you just use a pattern clamped to the top of the boards just like you have then laid out, then run a router down the joint between the two pieces? Any irregularities will be repeated on both pieces and cancel each other out, making a perfect joint.

Wayne,

That method works when doing straight joints, but won't work on a curve. The two radii will never come out the same, they will always differ by the diameter of the router bit.

Are you telling me that my spacial reasoning is flawed? It's not, you know. My procedure will work fine...

...you just need to use a small enough bit! 1mm bit = no more than 1mm error.

[embarassed] [big grin]

It will work. You need a pattern piece that has a smooth curve as close as you can get to the line of the two edges as they are now running through the centre of the pattern. After you create the pattern it should look like a slab of MDF with a channel running through near the centre. The pattern has to be longer than the boards you have so that the two halves of the pattern stay together.

You will use the same edge of the grove for all of the cutting and this guarantees the accuracy.

You then use that pattern, first on one piece and then the other, using a flush cutter with the ball bearing running on the pattern piece.

Peter

 

[tiralie]

Any help would be greatly appreciated.

Use the top piece to make a pattern by tracing a line onto a sheet of 1/4" mdf. Cut to the line to save the convex side of the curve and then match that curve to your existing piece, sand or plane to match any small changes in the curve of the top piece. Then use a straight router bit to duplicate that edge on your mating piece.
Tim

 

[jonny round boy]

Can you just use a pattern clamped to the top of the boards just like you have then laid out, then run a router down the joint between the two pieces? Any irregularities will be repeated on both pieces and cancel each other out, making a perfect joint.

Wayne,

That method works when doing straight joints, but won't work on a curve. The two radii will never come out the same, they will always differ by the diameter of the router bit.

Are you telling me that my spacial reasoning is flawed? It's not, you know. My procedure will work fine...

...you just need to use a small enough bit! 1mm bit = no more than 1mm error.

[embarassed] [big grin]

It will work. You need a pattern piece that has a smooth curve as close as you can get to the line of the two edges as they are now running through the centre of the pattern. After you create the pattern it should look like a slab of MDF with a channel running through near the centre. The pattern has to be longer than the boards you have so that the two halves of the pattern stay together.

You will use the same edge of the grove for all of the cutting and this guarantees the accuracy.

You then use that pattern, first on one piece and then the other, using a flush cutter with the ball bearing running on the pattern piece.

Peter

No Peter, it won't work. Not on a curve.

Let's say you make a curved template with a radius of 2000mm. If you follow that template with a bearing-guided trimming bit, you'll get the first piece with a radius of 2000mm. All ok so far.

Now, if you then use that same 2000mm template to cut the opposite side of the joint, it will have a radius of 2000mm plus or minus the diameter of the router bit (depending on whether the edge of the template you use is concave or convex).

Obviously the longer the arc the more that difference comes into play - on a very short arc you might get away with it & be able to clamp it up, but on a fairly long arc like the one the OP pictured you're going to end up with a gap of several mm at each end when the pieces are pushed together.

 

[Stoli]

Did you make them for a restaurant? If so, how did they find out about you?

No, these are just my current stockpile of thank-you gifts. I made most of these for the guys that helped me move my sawstop pcs into a basement shop. The remainder will get used when I bring wine to dinner parties, or as end of semester gifts for my kids teachers, etc.

 

[Stone Message]

Can you just use a pattern clamped to the top of the boards just like you have then laid out, then run a router down the joint between the two pieces? Any irregularities will be repeated on both pieces and cancel each other out, making a perfect joint.

Wayne,

That method works when doing straight joints, but won't work on a curve. The two radii will never come out the same, they will always differ by the diameter of the router bit.

Are you telling me that my spacial reasoning is flawed? It's not, you know. My procedure will work fine...

...you just need to use a small enough bit! 1mm bit = no more than 1mm error.

[embarassed] [big grin]

It will work. You need a pattern piece that has a smooth curve as close as you can get to the line of the two edges as they are now running through the centre of the pattern. After you create the pattern it should look like a slab of MDF with a channel running through near the centre. The pattern has to be longer than the boards you have so that the two halves of the pattern stay together.

You will use the same edge of the grove for all of the cutting and this guarantees the accuracy.

You then use that pattern, first on one piece and then the other, using a flush cutter with the ball bearing running on the pattern piece.

Peter

No Peter, it won't work. Not on a curve.

Let's say you make a curved template with a radius of 2000mm. If you follow that template with a bearing-guided trimming bit, you'll get the first piece with a radius of 2000mm. All ok so far.

Now, if you then use that same 2000mm template to cut the opposite side of the joint, it will have a radius of 2000mm plus or minus the diameter of the router bit (depending on whether the edge of the template you use is concave or convex).

Obviously the longer the arc the more that difference comes into play - on a very short arc you might get away with it & be able to clamp it up, but on a fairly long arc like the one the OP pictured you're going to end up with a gap of several mm at each end when the pieces are pushed together.

No, you use the same side of the template to cut both pieces of wood. That way each will be an exact copy of the template. The diameter of the bearing guided flush cutter does not come into the equation.

I have drawn a crude drawing.

Peter

[attachimg=1]

 

[Stoli]

No, you use the same side of the template to cut both pieces of wood. That way each will be an exact copy of the template. The diameter of the bearing guided flush cutter does not come into the equation.

I'm missing something.  If you use the same side of the template, you will end up with the exact same curve, I agree with that.  But those curves will not necessarily mate together, will they?

In your example, I think what you are saying is that the cutter will always follow the same edge of the pattern, but the final edge is formed when the router bit cuts closest to the pattern in the upper piece, and furthest from the pattern in the lower piece.  This is exactly the problem: they will have different radii.

In my first pic, I am moving the router bit along the center line.  You can think of either of the outer line as the template in your example.  In the second picture, I have removed the "waste" material and brought the pieces together.  You can see that they do not mate perfectly.

 

[Stone Message]

No, you use the same side of the template to cut both pieces of wood. That way each will be an exact copy of the template. The diameter of the bearing guided flush cutter does not come into the equation.

I'm missing something.  If you use the same side of the template, you will end up with the exact same curve, I agree with that.  But those curves will not necessarily mate together, will they?

In your example, I think what you are saying is that the cutter will always follow the same edge of the pattern, but the final edge is formed when the router bit cuts closest to the pattern in the upper piece, and furthest from the pattern in the lower piece.  This is exactly the problem: they will have different radii.

In my first pic, I am moving the router bit along the center line.  You can think of either of the outer line as the template in your example.  In the second picture, I have removed the "waste" material and brought the pieces together.  You can see that they do not mate perfectly.

No, the radius of the cutter does not come into it at all...

For the first cut the template is fixed to the top board and the red edge of the template is used to guide the cutter.

For the second cut the template is fixed to the bottom board and again the red edge of the template is what the bearing guided cutter follows.

You then have two exact copies of the same shape.

All you have to get used to in the counter intuitive use of the slot in the template - you only use the red (top) edge of the hole in the template as the guide. You make two exact copies of that top edge of the guide.

Long before I started videoing I used to make solid oak edged round and oval tables using the same template for the solid oak and for the veneered ply that formed the top of the table.

OOPS:

Sorry chaps and apologies to those who have been correcting me - I have just remembered when I was trying to find a photograph - I must have used a guide bush with a removeable ring. Otherwise the shapes are slightly different. It may be the medication, could be age but is probably my evening pain killer from Bordeaux!

Peter

 

[Michael Kellough]

Peter thought he must have had guide bushings in mind when he suggested the simple template.

I just spent a bunch of time describing a method that uses only one template and a guide bushing. Like Peter I finally concluded it won't work and deleted it all. It could work with two guide bushings if you found the right ones. It comes down to an offset that exactly matches the bit diameter.

Other that the adjustable trammel beam the only other way is to use two perfectly fitting opposite templates. And the best way to make templates is with the trammel beam so unless you plan to make a bunch more pieces with that radius you might as well proceed to the real thing.

Pushing the two pieces together and running a jigsaw down the line (pretty much the same as the 1 mm router bit idea) is a technique that was used to make double wide dugout canoes with a hand saw in the old days. It work best if you are chasing a straight line. If the joint is curved the fit will always be hollow in the middle. The ends of the concave piece will need to be shaved down.