What's The Best Way To Make This Joint?

[Julie Moriarty]

First time making a joint with curved wood. FWW recently had an article that showed how to make a sliding dovetail joint on a curved drawer. But that curve wasn't nearly as pronounced as the one I'm working on.

http://i867.photobucket.com/albums/ab233/jules42651/Woodworking/Kitchen%20Cabinets/bowjoint_zpsc60fe961.jpg

I'll be trimming the bow flush with the sides so I'm thinking the sliding dovetail, if I can do it, might not have enough stock left on the outside of the cut. The smallest dovetail bit I have is 1/4" @ 80. The bow and sides are 1/2" thick. Unless I favored one side, that would only leave me with 1/8" stock on the outer edge. Would that be enough? BTW, the panel will be slid into the frame once the sides are secured to the bow. The panel is 30" wide so the joints would have to be strong enough to handle the possibility of slight racking while the panel is being inserted.

Thru dovetails have been voted out. I can't see dowels working. Only what you see in the pic will be exposed. So I was thinking of corner blocks on the inside as a last resort. I was hoping to make a more professional joint though.

[Jeff Duncan]

Do you have to slide the panel in after? I think I would route a dado in the sides and front and glue them to the panel. I would just use a miter joint where the sides meet the front apron. In terms of clean looks that's probably about the best bet and would also make the panel quite rigid.

JeffD

[Julie Moriarty]

Yes, if I glue up the sides to the bow front I would need to slide the panel in from the back. The problem with gluing the panel to the frame is wood movement. I'm looking at this like building a panel door so I am allowing for wood movement and allowing the panel to float. That means the frame joints have to do all the bonding work.

[David Weaver]

Can you use a template board or card paper to mark the length of the board on a drawer, then dovetail the board flat first and then bend it after cutting the joint? I may not be understanding exactly what the joint is since we can't see it.

I don't know how you folks with power tools do things, but I might be inclined to tack (with non permanent and non penetrating glue) some sort of 1/8" shim board to reduce the depth of the dovetail bit, take the shim board off after cutting and then plane where it was glued to clean the surface.

[Dave Richards]

Julie, could you put a rabbet on the ends of the drawer front and pin the sides with tiny dowels? I just did plans for a small cabinet with a bow front. Five drawers and all joined that way.

If the bottom panel is solid wood the grain should be running across the width of the drawer. the rear edge of the drawer bottom should pass under the drawer back and be attached with a screw or two in slots on the back so the panel can expand rearward as needed. If it is ply, you should be able to glue it in with no problem.

[Richard Coers]

Is that panel solid wood? If it's veneer you don't have to worry about expansion. If it is solid, isn't it going to blow the joint apart as it expands? There is now place for that curve to give. The expansion is going to press out on the curve and crack the corner joint. How about a miter cut and then add splines through the miter of the same wood? How tall is the outside frame and how thick the panel? Room for some corner blocking under the panel?

[Jeff Duncan]

So your saying the center panel is solid wood? If so your definitely going to need room for that wide a panel to move! I assumed it was a veneered panel in which case movement is not a problem.

If it is solid wood that's going to be a bit tricky. I think the sliding dovetail may be your best bet. You could use the router table to do the slots in the sides, but you may need to resort to hand tools, (gasp:eek:), to make the tails.

JeffD

[Bill Wyko]

Can you hide an "L" bracket on the underside?

[Julie Moriarty]

The piece I'm working on is the light box in the center below.
http://i867.photobucket.com/albums/ab233/jules42651/Woodworking/Kitchen%20Cabinets/CooktopCabLites_zps014f7c4c.jpg

The cabinets on either side are pull out racks. The clearance between the edge of the wall cabinet doors and the cabinet itself is 1/2". If I didn't cut the sides of the bow flush I could maybe leave a 1/4"-3/8" protrusion but since the pull out racks have some slop in the runners, I'm worried about the cabinet doors banging into the bow overhang.

This is what's up there now. I still have to dye and apply finish to the doors, that's why they aren't on now. But if they were on, you'd see how close they are to the edge of the cabinet.
http://i867.photobucket.com/albums/ab233/jules42651/Woodworking/Kitchen%20Cabinets/cooktop_cabs_zpsaa1b76fa.jpg

With only 1/2" thickness in the wood of the bow and little or no overhang, most typical bowed drawer face joints won't work in this application. And yes Jeff, the panel is solid 1/2" thick sapele. So it has to float.

I may have to either put blocks inside the bow corners or find some decorative hardware that I'd run through the face. Or both.

[Dave Richards]

Julie, I was confused. I thought you were making a drawer but from your drawing it looks like it isn't a drawer at all. It seems to me the right thing for that center section would be to miter the corners and fit some glue blocks inside to glue the curved front and the sides to.

[Shawn Pixley]

I'm thinking that the sliding dovetail isn't the right way to go (however interesting it would be to try). I might try a mortise and tenon (pinned?). Or you could use corner blocks as Dave described. If the joinery isn't exposed there many ways of adding a block to the back of the arc to simplify the joinery technique. The corners could be mitered at the exposed side. You could also use half blind dovetails to connect the two (with or without back blocks).

Just ideas...

[Kevin Jenness]

Your drawing and photo seem to show a curved fascia with returns and no back housing a bottom panel at least 36" wide which you are making of solid wood. Without knowing the species and cut of timber or its current moisture content versus its projected equilibrium MC as installed, I would think you would need at least a 5/16" deep groove in the sides to accommodate seasonal wood movement, leaving not much meat in 1/2" material. Why not keep it simple; miterfold the front and sides, groove them3/16" deep to receive a veneered panel (that you could glue in place) and put in a back member above the bottom and inboard to allow for scribing the bottom to the wall.

[Kevin Jenness]

P.S., to the original question, a sliding dovetail would not work here with this little overhang; you would need at last 1/4" to have any confidence in the joint hanging together through installation. Traditional inset drawers would probably have a half blind dovetail joint with the bow front flattened off at the back corners to meet the sides at 90 degrees. That would not look so hot as an exposed joint in this application, thus my preference for a miter. Accurately cut miter joints can be taped on the outside, glued and folded to make a reasonably strong joint with ease, reinforced with glue blocks if desired. For general information on joinery Ernest Joyce's Encyclopedia of Furnituremaking is invaluable.

I see on looking back to the original post, the bottom is only 30" wide, but that is still a considerable width whose potential movement needs to be anticipated in the construction.It looks like you are using flatsawn mahogany, a relatively stable wood, but I would still allow for 1/8" per foot of seasonal movement in my part of the world (Vermont).

[Julie Moriarty]

I was toying with this after my last post. I made up some corner blocks and found the curve even at the ends pronounced enough to require taking the blocks to the spindle sander. I tried to imagine getting the right degree cut for a miter joint but without some scrap to do test cuts (I don't have anything with that curve) it was too risky. The blocks would certainly work on thicker material but I look at the 1/2" that I have and wonder how well screws would hold. The sides are jatoba and heavy and rock hard. So maybe the screws would hold well.


I see on looking back to the original post, the bottom is only 30" wide, but that is still a considerable width whose potential movement needs to be anticipated in the construction.It looks like you are using flatsawn mahogany, a relatively stable wood, but I would still allow for 1/8" per foot of seasonal movement in my part of the world (Vermont).

The panel is sapele. This is the first time I've worked with it. The sides are jatoba. According to the Wood Handbook (which I did not refer to until now):
Sapele - Shrinkage from green to oven dry moisture content
Radial - 4%
Tangential - 7.4%
Volumetric - 14%

Mahogany's numbers are 3%, 4.1% and 7.8%. Sapele's numbers are comparable to Elm, Maple and Red Oak. The Wood Handbook doesn't have Jatoba listed but tells you to refer to Courbaril. That species' numbers are 4.5%. 8.5% and 12.7%. But the Jatoba length is longitudinal so that's not an issue.

I have an Excel spreadsheet that calculates wood movement for a number of commonly used woods. It doesn't do Sapele but it does do commercial Red Oak. Tangential changes from 7% to 11% MC for 30" wide boards is 5/16". It's pretty much the same for Elm and Maple. I have no idea what kind of MC changes I'll see in our kitchen or if heartwood (which the panel consists of) shrinks & swells less, but tangential changes increase 1/16" for every additional degree of MC over 11%.

It seemed like such a great idea and was going better than I expected. Now it looks like I may have painted myself into a corner. Maybe I can rip the panel in two and add a stile. That would also give me some additional support. I'm just not sure how that look will fly around here though.

[Mel Fulks]

If you have the cut offs from the panel soak them in water and measure the expansion. I don't disagree with the advice you have received in principle,but if you seal all the end grain ,and if the wood is kiln dried ,I don't think there will be a problem. A center frame member doesn't really go with that design.

[Julie Moriarty]

Thanks Mel. While I am aware of wood movement, I usually just follow standard practices, but I've never done anything this wide. I had planned on sealing the piece both inside and outside, to minimize moisture absorption. I just had no idea if that would matter much. We only run the air conditioning when it's really hot. And Chicago can get really humid in the summer. I have about 20 raised panel doors I made that have been in place for several years or more and none of them have failed but the widest panel is probably half the width of the sapele panel for the light box.

After I made the last post I was thinking maybe I would glue strips of jatoba to the inside of the frame and do it in such a way as to make the panel removable. That way I could do the wiring easier as well as maintain the lights or replace them if needed. It would also allow me to anchor the back piece of the frame to the wall studs. So the only issue at the bow joints would be the ability of the joints to hold the bowed front in place. Also it wouldn't affect the aesthetics.

[David C. Roseman]

Julie – Nice project! I don't think you've painted yourself into a corner at all. But I do have some thoughts on how to proceed. Maybe I'm totally missing something, but I don't see how a sliding dovetail joint would work at all to fasten the panel to the sides. Its purpose is to allow panel expansion and contraction along the joint. In the photo, your panel is oriented with the grain parallel to the sides, so its movement will be perpendicular to the sides. BTW, I’m guessing that the panel in the original light box that you're replacing is made from veneered sheet goods. In which case they didn’t face the challenges you have! Perhaps a better design for you would have been to orient the grain of the panel from side to side (then you could use sliding dovetails at the sides and anchor the bow front rigidly to the panel), but no big deal.

Based on your research for wood movement of sapele, I'm afraid the panel will move enough over its 30" width to create problems, unless you either (1) float it in regular dadoes in the sides deep enough to hide the movement, or (2) fix the sides to the panel and have them move with the panel, floating the attachment of the sides to the bow front. The fact that the panel is in a light box (even the new LEDs create heat), plus mounted over a cooking range, will only make it more susceptible to moisture exchange. The ½” thickness of the sides is thick enough to accommodate simple dados 1/4” deep on either side. That should be enough to handle your panel movement if your research is correct. Given that you'd like to mount the box between your cabinets with the sides flush to the cabinets, I think you're right to have the panel float in dadoes in the sides, rather than have the sides float with the panel.

In fact, I think the bigger challenge with your current design is how to fasten the bow front. Regardless of whether the panel floats in dadoed sides or is fastened rigidly to sides that move with the panel, the curve of the bow front is going to impede the movement of the panel unless you use a fairly deep dado joint in the bow, and fasten it to the panel only towards the center. A 1/4" dado sounds like it may be enough. But you could always go to 5/16".

Bet this project turns out beautifully.

David

[Mel Fulks]

With that nice finish you are using the panel should be stained and finished before installing in the frame since even slight shrinking could show a light line. You have little flat sawn wood in that panel,so that helps too. Do the soak test as an EASY way to ease your mind.

[David C. Roseman]

Thanks Mel. While I am aware of wood movement, I usually just follow standard practices, but I've never done anything this wide. I had planned on sealing the piece both inside and outside, to minimize moisture absorption. I just had no idea if that would matter much. We only run the air conditioning when it's really hot. And Chicago can get really humid in the summer. I have about 20 raised panel doors I made that have been in place for several years or more and none of them have failed but the widest panel is probably half the width of the sapele panel for the light box.

After I made the last post I was thinking maybe I would glue strips of jatoba to the inside of the frame and do it in such a way as to make the panel removable. That way I could do the wiring easier as well as maintain the lights or replace them if needed. It would also allow me to anchor the back piece of the frame to the wall studs. So the only issue at the bow joints would be the ability of the joints to hold the bowed front in place. Also it wouldn't affect the aesthetics.

Julie, I replied before seeing your most recent post. Sounds like a great idea to make the panel removable. One thing to keep in mind with this, though, is the risk of the panel cupping in that environment if not captured in dadoes. If you sandwich it between fixed trim strips across the front and back, with the bottom strips removable, that should solve that problem, but still let it move.

David

[Julie Moriarty]

Julie – Nice project! I don't think you've painted yourself into a corner at all. But I do have some thoughts on how to proceed. Maybe I'm totally missing something, but I don't see how a sliding dovetail joint would work at all to fasten the panel to the sides. Its purpose is to allow panel expansion and contraction along the joint. In the photo, your panel is oriented with the grain parallel to the sides, so its movement will be perpendicular to the sides. BTW, I’m guessing that the panel in the original light box that you're replacing is made from veneered sheet goods. In which case they didn’t face the challenges you have! Perhaps a better design for you would have been to orient the grain of the panel from side to side (then you could use sliding dovetails at the sides and anchor the bow front rigidly to the panel), but no big deal.

Based on your research for wood movement of sapele, I'm afraid the panel will move enough over its 30" width to create problems, unless you either (1) float it in regular dadoes in the sides deep enough to hide the movement, or (2) fix the sides to the panel and have them move with the panel, floating the attachment of the sides to the bow front. The fact that the panel is in a light box (even the new LEDs create heat), plus mounted over a cooking range, will only make it more susceptible to moisture exchange. The ½” thickness of the sides is thick enough to accommodate simple dados 1/4” deep on either side. That should be enough to handle your panel movement if your research is correct. Given that you'd like to mount the box between your cabinets with the sides flush to the cabinets, I think you're right to have the panel float in dadoes in the sides, rather than have the sides float with the panel.

In fact, I think the bigger challenge with your current design is how to fasten the bow front. Regardless of whether the panel floats in dadoed sides or is fastened rigidly to sides that move with the panel, the curve of the bow front is going to impede the movement of the panel unless you use a fairly deep dado joint in the bow, and fasten it to the panel only towards the center. A 1/4" dado sounds like it may be enough. But you could always go to 5/16".

Bet this project turns out beautifully.

David

Thank you David. I'm hoping it turns out well too. :)

To your comment about the panel bowing if there's nothing to prevent it, I can see that happening. I've dyed and finished countless sample pieces, trying to get the dye formula right, and all have bowed when only one side is finished. They do seem to come back once the other side is finished but the widest sample piece was about 8". 30" is a different story.

I had one idea where I'd somehow secure the bow to the panel and slide it in from the front, into the dadoes. I was wondering if hot glue would hold that and if it there would be enough give to allow for wood movement. I could permanently bond the bow to the panel at one point in the center and use some flexible adhesive at points along either side. Though I have no idea what kind of adhesive that would be. Hot glue came to mind but I have little experience with it.

Another option would be something taken from the bread box joint concept, using slotted pieces attached to the inside of the bow and securing the outer ends of the panel to them. I have no idea if the bow piece will lose any more of it's bow shape so I would want to make sure there's something keeping it from doing so.

As for the LEDs, my brain hurts from the hunt to find dimmable LED lights. I have a Lutron Maestro dimmer installed in the wall right now. I love them! The best dimmer I've ever installed, and I've installed a LOT! The problem comes when trying to find dimmer, driver and light to all be compatible. Lutron makes a driver compatible with their Maestro dimmers. The LED manufacturers I've contacted all say you have to use their driver or the warranty is void. Some have an approved list of dimmers but none have tested the Maestro. And I don't want to install a different dimmer. This is becoming the irresistible force meeting the immoveable object.

I'll bet when this light box project is complete I will have better than 50 hours into it, when you consider all the time I've spent trying to make the image in my mind come alive.

[Chris Fournier]

Floating tenon.

[Kevin Jenness]

Apologies for repetition, but you are making it too difficult. Sometimes solid wood is not the best choice for a particular design. If you are wedded to that particular piece of sapele, resaw it and lay it up(both faces) on some mdf and wrap your grooved, mitered frame around it. Sapele veneer is readily available. For quick wood movement estimates for most species, there is a timber movement calculator (the "shrinkulator") at Woodweb.com.

[Jim Matthews]

Is this assembly hollow?

It cries out for a torsion box with an MDF substrate and the Sapele applied as a veneer.
My experience with Sapele is that it wants to split along the alternating grain lines, if it's "ribbon" cut.

If the grain orientation is as shown, it should move more front to back than side to side,
much like the bottom of a drawer.

If you mount the show face at the bottom to a kerfed MDF substrate it might take up the movement without
breaking the glue bond. Some of the modern contact adhesives are tolerant of movement like this.

Thinner veneers should exert less force.

I'm not sure it's necessary - if you leave glue the front face to the curve and apply a batten at the back to hide a gap,
it should have enough room to "breathe".

I'm unclear about how much heat will be trapped by the cans, and how much that would affect the panel.

[Dave Richards]

It's probably too late but your questions have been rolling around in my head so I thought I'd offer this.

http://farm9.staticflickr.com/8253/8704938516_93cd9acfe3_z.jpg (http://www.flickr.com/photos/8294157@N08/8704938516/)

Corners mitered with blocks on the inside as I suggested before.

If you're concerned about the panel sagging, you could use some stringers on the top (long brown things) that are glued to the center inch or so of the panel and hook over some cleats glued to the sides (yellow blocks on the sides).

The whole thing could be mounted so it slides out like a drawer if you need access to the inside to replace lights or whatever. Add connectors so you can unplug the lights from the power and remove the whole box.

[Julie Moriarty]

I can't create a veneer out of the glued up panel. My bandsaw will only resaw to 10". So any thoughts of creating a veneer out of the existing panel aren't possible with the tools I have. The panel is taken from the same stock I used to make the panels of the cabinet doors. It blends in nicely with the doors that will be installed on the surrounding cabinets. To me, veneers lack depth. Sapele is one of those woods that has interesting depth. I think chatoyance might be the better word. I haven't done much veneer work but I can tell when something is solid wood or veneer. I'd rather not go down the veneer path right now fearing I might be disappointed with the contrast between the adjacent door panels and the light box panel.

Thank you all for your help, suggestions and information. I feel armed now with sufficient knowledge that I can make this work, if that's possible with a solid wood panel. The removable panel appeals to me for light installation and maintenance purposes, and should solve the wood movement issues. I just have to find the right lights. After being unable to match dimmer, driver and fixture in the low voltage lighting arena, I was thinking I'd just pick up some 120v pucks from HD, cut oversized holes in the panel, dye or paint the inside of the circle ebony and mount the lights with about 1/2" space around each light. That would address the heat issue and allow for air movement. Plus I would have full dimming capacity. LEDs typically don't dim below 20%, some only to 50%. Incandescent dim below 1%. I like lighting controls to run the range from functional illumination to just adding ambiance.

When it's done, I'll post a picture here and tell you how it went. It's like you guys have been in the shop with me brainstorming this, so I owe you. Thank you again!