These are the templates I used. All are used with a router and template cutting bit. To the left is the template used to make the dados in the bottom, back, and top. These dados receive the spacers that run front to back. The center template is the "master" curve. I used this to make the bottom of the mantel, as well as the template for the top. It will eventually be used to make the bending form for the finished front piece. The rightmost template is for making the top. I'll scribe this template to fit the wall, then shape the top with it.
You can see the quilted maple veneer that will eventually cover the front of the mantel peeking out from under the table.
Here you can see the "skeleton" of the mantel. The back, spacers, and ledger strip are 3/4-inch birch plywood. The bottom is solid red maple, and the front is 1/4-inch birch plywood. Its held to the wall with three "Red Head" masonry anchors, one of which is visible in the closest cavity.
I put the plywood front on to provide a glue surface for the finished front, which will be a bent lamination of hard maple with a quilted maple veneer.
This shows the skeleton from the front. Its easy to imagine what it will look like when its completed! Just add the wenge top, and the quilted maple front.
Here's the setup I used to resaw the maple for the bowed front. I found that I needed the featherboard on the outfeed side to keep the board against the resaw fence.
That's an 8-foot long, 8.5-inch wide piece of 5/4 hard maple. I sliced off slabs just under 0.20 inches thick. Those went through the drum sander to take off the saw marks. Final thickness is around 0.13 inches. I was surprised at how much had to be removed. The saw blade did bow a little, and this was easily seen as I passed each piece through the drum sander. The saw marks disappeared at the edges first. I think if I had been more fussy about aligning the fence to the blade's drift angle, I would have been able to eliminate the bow. But there was no real need to do that, I have the laminations I need, at the right thickness.
Here you can see the resaw fence, and a cut in progress.
The resaw fence. The face is maple, the other parts are plywood. Using a solid hardwood for the face makes it easy to square the face to the base using a jointer.
Here are the completed laminations. The total thickness is about 0.55 inches. The quilted maple veneer I acquired to put on the face is 0.025 inches thick. I'm not sure how much additional thickness, if any, will be added by the urea-formaldehyde glue when the assembly is complete.
Here is the wenge top, temporarily mounted to the skeleton. I'll eventually add "tails" on the side, to bring the wenge all the way back to the wall (like the template). A piece of maple will cover the side of the skeleton, and go all the way back to the wall as well. I'll also be giving the face of the top an oval profile.
This is the router and bit I used to shape the wenge. That brown wenge dust went EVERYWHERE! I need to set up some sort of dust collection for this kind of work. The plastic shroud available from Bosch is not tall enough to work with this long cutter. I could have done this on the router table. But the piece was big, and I felt that hand-holding the router was better than using the router table.
Here is the form I'll be using to make the laminated bowed front, and to glue the veneer to it. This picture shows the last step of making the form, laminating two layers of 1/4-inch plywood and one layer of masonite to the top of the form using the vacuum press. I used Unibond 800 urea formaldehyde glue.
The form is built of 3/4-inch birch plywood. It is 8 feet long, and just over 10 inches wide. Its built pretty much the same way as the mantel skeleton, with the internal dividers spaced every 8 inches. I found that this isn't quite enough internal structure, though. Under the force of 0.7 atmosphere of pressure provided by the press, the three layers of the top are actually deflecting down just a bit, forming a very shallow bowl in each cavity formed by the sides and dividers. That's not good! If the laminated maple follows this shape, and I put a gloss finish on it, you'd likely be able to see this distortion. But the form will still be very useable. Instead of using it inside the press, I'll just put the maple laminations by themselves into the vacuum bag, and bend that assembly over the form. Next time I build a form, I'll use more structure.
This is the vacuum pump. Its based on information provided by, and parts purchased from Joe Woodworker. The sub-reservoir and vacuum reservoir are made from lengths of 4-inch schedule 40 PVC. The pump assembly is housed in a 19-inch square box made of 3/4-inch birch plywood. The box is on casters.
The vacuum bag is made of 30 mil vinyl, also purchased from Joe Woodworker. It is 26 inches wide and 10 feet long. To seal the open end of the bag, I applied some self-adhesive foam weatherstrip on one side, and clamped the opening between a board and my assembly table. The seal appears to be almost perfect.
Here you can see the maple laminations in the vacuum press. This is an "outside" pressing. The laminations, and a length of plastic netting, are inside the bag. That is then bent over the form, and vacuum applied to the bag. I did it this way because the form is flawed. I didn't want the laminations to conform to the flaws in the form. Using this method causes vaccum to press the laminations together from both sides, while the actual bending is done without the assistance of the vacuum.
The plastic netting serves as a top platen, and helps to make sure the air is evacurated evenly across the top of the laminations. I didn't bother with any type of platen under the laminations, as the assembly is held tightly to the form with the clamps at the ends.
After about 14 hours in the press, I removed the vacuum and removed the assembly from the bag. After another few hours, I scraped the glue squeeze-out off of one edge, and ran that edge over the jointer to straighten it out. I then cut the assembly to width on the table saw.
You can see the glue lines in this photo. This was Unibond 800 with the light colored catalyst. Its a fair match with the maple, but not perfect. It won't be a problem for this project, as the edge won't be prominent once the mantel is assembled and mounted in place.
Here you can see the laminated maple sitting on top of the form. There is a small amount of spring back. It takes very little pressure to negate, so I'm pretty sure the assembly will glue to the skeleton without difficulty.
What's more troubling, though, is that the assembly is cupped slightly upwards. The individual laminations were flat. This must have been caused by the moisture in the glue. Because I didn't clamp the glue-up with the form inside the press, there was nothing to restrain the wood from cupping. I think I can live with it, but I will have to modify the front of the skeleton, I think, to make sure I have a good glue bond to the back side of this, which is now slightly convex.
Now the quilted maple veneer is in place, and in the press. This time, I put the form inside the bag. As the laminations are glued together, I'm not worried about the flaws in the form.
Once again, the plastic netting is used as the top platen, to make sure the vacuum pressure is consistent across the entire piece.
Here's a closer view of the assembly in the press. You can easily see the figure of the veneer through the plastic netting. You can also see where I used a short piece of veneer tape to cover a small knot. I don't want any of the glue to come through here.
Another view. It took all 5 clamps and a decent amount of weatherstripping inside the bag to get a good seal. Even then, leaks in the system, most likely at the quick-disconnect, resulted in the pump cycling a little. It ran for about a minute every couple of hours.
Two interesting things happened in gluing the veneer to the front in the press. First, there was quite a bit of glue bleed-through. Where the veneer is darker, it must be very porous, and the Unibond came through there. I was able to remove it with a card scraper. I then scraped and sanded to 220 grit using a random-orbit sander. The glue is a good color match to the darker parts of the figure, and I'm going to put a clear finish on. Hopefully, there won't be any significant negative consequence from the bleed-through.
The other interesting thing is, the cupping is gone! Perhaps the laminations dried out, and returned to their previous flat shape. I'm sure some of this was due to being pressed flat against the form for 15 hours. Once its glued to the skeleton, I hope it stays flat.
Now I'm gluing the laminated and veneered front to the skeleton. I'm using a piece of 1/4-inch plywood to protect the veneer from the clamps.
The front sits proud of the skeleton both above and below. The wenge top, shown here upside down, has a groove in it that will fit over the front, thus hiding the joint.