A few weeks ago, I read an entry on the Popular Woodworking Editors’ blog by Bob Lang, where he noted how some woodworkers are more concerned with shavings, the waste material, than they are with the finished product. I tend to be one of those woodworkers. I have made tens of thousands, perhaps hundreds of thousands of shavings, and I still marvel at them. Thick ones, fluffy ones, ones you can read a newspaper through; I like them all, but my favorite are the ones made when planing inlay bandings, which make beautiful geometric shavings.
The living history farm, where I volunteer has an Independence Day celebration. I wanted to demonstrate woodworking, celebrate the holiday, and make a shaving the project and not the waste. To do that, I glued up a stack, of the following:
2 layers of red dyed veneer
1 layer of holly veneer
2 layers of blue dyed veneer
1 layer of 1/16” holly veneer
2 layers of blue dyed veneer
1 layer of holly veneer
2 layers of red dyed veneer.
The stack was glued together with white glue between boards covered in packing tape, using plenty of clamps to ensure a good bond. I used extra glue, so the clamping pressure would impregnate the veneer with glue, giving it a more homogenous nature. I also oriented the veneer so the grain in all the layers were running in the same direction to facilitate planing. When dry, one edge of the stack was planed and glued to a piece of scrap, so the piece can be clamped in the vise and get the maximum yield out of the stack.
Planing the shavings calls for a perfectly sharp blade, a steady hand and a bit of trial and error in setting the chip breaker and cutting depth. If the chip breaker is set too close to the cutting edge, the shaving will be fractured and too delicate. Set too far back and it will not form a nice curl. The depth, needs to strike a balance between being so thin that the layers tend to separate and so thick the shaving won’t withstand being manipulated. The shaving is planed off and worked into a nice spiral shape. The 2" wide stack, yielded over 250 shavings.
Most adults and many children have seen plane shavings, but very few have seen multi-colored ones, which made for a fun day handing them out and demonstrating how they were made.
As noted earlier, the large center “drawer” on the original breakfront is a Butler’s desk. This is an attractive feature, but one the customer wasn’t interested in reproducing. Although it will almost certainly never happen, I decided to make the center pull out section, in such a way, that it could easily be retrofitted to match the original. This added almost no time to the construction.
As built now, the drawer has many of the features of a conventional Butler’s desk. It pulls out a short distance, being stopped by a cleat fastened to its top which strikes the upper drawer blade. The hinged fall front, is supported by quadrant stays and is held in the up position with a Butler’s catch. You can view a 3 part video, documenting the installation of this hardware at the following links.
Where the as built pullout differs from the original, is instead of the typical Butler’s desk interior, it has only four equally sized drawers in two tiers. These drawers are more functional than an actual gallery type interior, which is why the customer chose to go with this configuration.
The carcass of the pullout section has a bottom board of 3/4” thick material; mahogany where visible and transitioning to pine where hidden by the drawers. The top is 1/2” pine with a thin strip of mahogany glued to the front edge. The sides are 3/4” mahogany at the front, and 1/2” pine where hidden by the drawers. The transition from the thicker mahogany to thinner pine sides, creates a lip, The lip allows, if desired, for the fabrication of an insert with pigeonhole, document drawers and a prospect door, matching the original to be slipped in at a later date. The case is joined with through dovetails at top corners and a combination of through and half blind dovetails on the bottom corners; half blind on the visible front and through dovetails from there back.
The drawers in the pullout section are roughly the same size as the drawers in the flanking base units and are veneered and inlaid in the same manner. The central drawer blade is housed in a routed dado in the pine portion of the side. Cherry drawer runners are also housed in those dados, and are held in place with screws driven in over sized holes, to allow for expansion and contraction. The case is not that deep, so just a small amount of play is enough to allow for seasonal changes. With the difference in thickness of only 1/4” between the mahogany and pine sides, a very thin drawer guide is required. Screwing this guide in place would be problematic from the standpoint of a getting the screw heads flush in such a thin piece and allowing for movement. I took a lead from the Seymour’s and cut the pine guide with its grain running the same as the sides and glued the guide in place with hot hide glue. The lower drawer blade is 3/16” thick; just thick enough to clear the knuckles of fall front hinges. This lower drawer blade is fastened in place with flat head screws. The runner and guide for the lower drawers being thin, have the same challenge as the guides for the upper tier drawers and here again I followed the Seymour practice, by cutting the pieces with their grain running in the same direction as the surface to which it is glued.
The 1/2” thick vertical center divider is housed in dados cut in the top and bottom boards and fastened in place with screws. Because the grain in the divider runs the same as the top and bottom boards, no allowance has to be made for expansion and contraction. The divider is notched to accept the drawer blades The runner for the upper tier of drawers on the other hand does present a cross grain situation. Given the thinness of the vertical divider and the need to have runners on both sides, a full depth dado to house those runners was out of the questions, so the dado is only 3/32” deep. Even such a shallow dado, provides more than enough strength to support the drawers. To overcome the cross grain issue, two long wood screw were driven through one runner, the divider and into the opposite runner. Then the runners temporarily removed and the holes in the divider elongated to allow for seasonal movement.
The fall front is constructed exactly the same as the doors in the lower case, i.e. a core of narrow poplar boards with their growth rings arranged in a quartersawn configuration. The core was planed to precisely fit the opening; both the perimeter and to sit flush in the opening. The core was then crossbanded with quarter cut cherry veneer.
Also like the flanking doors, the ribbon stripe mahogany frame around the oval was hammered down first and a template was used to route out the oval shape.
The same template was used to trim the crotch oval to shape. The oval was made from a butt end, book match of crotch mahogany veneer. In the how dumb can a person be column, I reversed the core, veneering the outside as the inside face and making the accurately planed outside face, the inside face. Fortunately, the carefully constructed base unit, meant the opening was so square that only a minor amount of planing was necessary to properly fit the fall front to the opening.
The inlay surrounding the crotch veneer oval presented quite a challenge. Being made from 2 layers of 1/16” satinwood, enclosing a layer of 1/16” black dyed veneer, if glued up as a straight section it wouldn’t be flexible enough to bend to the relativity tight radius at the ends of the oval.
The 1/16” satinwood, was sawn from solid stock and brought to thickness with a drill press mounted thickness sander. The layers were just under a 1/4” wide. This is far wider than necessary or even desirable from the standpoint of inlaying, but the width made them easier to form. The form was made from 1/4” plywood, using the same template used to route for the veneer, but instead of the reduced diameter bit, a 3/16” bit is used. This leaves the form the exact shape and size of the finished veneer oval. I originally thought I could steam the pieces, and bend them around the form.The problem is the three layers proved too cumbersome to manipulate before they cooled. Then I used a clothes iron to supply the heat and a spray bottle for the moisture. This worked perfectly, but when released from the form, the pieces of satinwood tended to spring back. This made trying to glue the three layers together somewhat frustrating. Adding to this frustration was the fact that the total length of the inlay was around 8’ long, requiring each layer to be made from 3 individual sections (9 pieces in all). The joints in the sections were staggered. In the end I bent the layers around the form using moisture and the iron, taping the layers to the form. After everything was in place on the form, a band clamp was applied to really chinch things down tight.
To install the inlay, the groove was made to a depth of about 3/32” with the router against the template. The inlay was released from the form. Hide glue was brushed into the groove and the individual sections were carefully pushed into place, using the iron to liquify the glue and help persuade the satinwood to bend. The seam in the inlay is made at the top center, where it will be hidden by the keyhole escutcheon.
Now, with the inlay in place, I went back and applied a thinned down glue to the inlay. I did this, because I was concerned that the glue might not have gotten between the layers, but the thin glue and heat from the iron ensures good adhesion. With the groove being only 3/32” deep, this left a considerable amount of the inlay proud of the surface. I did not want to go any deeper, because that would have put additional stress on the bit and contributed to a less precise groove. Making two passes is out of the question, because of alignment issues between the bit and guide collar. To remove the bulk of the extra inlay, two strips of wood were taped to the base of the laminate trimmer so it could straddle the inlay and trim it very close to flush. A block plane flushed up the ends of the oval, where the router couldn’t reach. All that was left was some light scraping and sanding.
A rabbet was run on its inside bottom edge of the fall front. A corresponding rabbet was run on the bottom front edge of the pullout section. The rabbets are as wide as the thickness of the bottom board/fall front and as deep as half their thickness. The rabbets allow the the fall front to lie flush with the bottom board of the pullout, when open and still have a full height “drawer” front. The hinges came from Horton Brasses and are very well made, but were deeply stamped with a company logo; not something I wanted on a period reproduction, so I filed the logo away. To ensure the proper alignment between the pullout and the fall front, they were placed in the opening and a registration mark was made on them from the inside. The fall front is clamped in the down position to the pullout section with the registration marks aligned, and the hinges are scribed in place. It is critical that the hinge pin be centered on the joint between the fall front and the pullout section. To keep my options open, only one screw per leaf is used durning the fitting phase. After the hinges are in place and working properly, the unit is placed in its opening and its fit is noted. The fall front was carefully planed to have about a 1/16” gap around its perimeter. This requires unscrewing the fall front several times and planing until the gaps are perfect. Once I was satisfied with the fit, I used a 3/16” rabbeting bit followed by a flush trim bit to accurately remove 3/16” from the perimeter of the fall front (the same procedure used on the doors). The rabbeted bottom edge didn’t leave enough material to allow me to comfortably use the flush trim bit after the rabbeting bit, so after using the rabbeting bit, I hand planed the rabbet away. The perimeter of the fall front is covered with 3/16” solid mahogany, in the same manner as the doors, with one difference, the side edging has to have a recess cut to fit over the quadrant stays. The same procedure used to install the satinwood perimeter inlay on the door was used on the fall front. I had to do some minor trimming of the rabbet in the pullout section to relieve some binding.
The fall front is right at 48” wide and given the potential weight of the upper cabinet, I became deeply concerned that, the weight would deflect the upper drawer blade and bind the pullout section. As added insurance, I removed the drawer blade, which was made easy due to the use of hide glue, planed 1/8” off its inside face (remember the dovetailed ends were rabbeted) and reinstalled it. A 1/8” thick, by 1” wide strip of mahogany was glued to the bottom front edge of the drawer blade to replace the material planed away, thereby returning the opening to its original size. At the rear of the drawer blade, I screwed on a 3/4” piece of angle iron. I noticed the angle iron was not straight, so like a floor joist, I put the crown up, so it would flatten out under weight.The angle iron added considerable rigidity, but really not as much as I thought, still it was worth it.
The pullout serving slides have a core similar to the drop front, but it is made of pine and I wasn’t as careful when arranging the growth rings. The panel was planed to a thickness, so the baize would sit below the surface of the mahogany edging by 1/32”. The panel was planed to fit the opening and a 1” wide by 1/4” thick tenon was run on each end. Except for the last 2”, the tenon was cut to a 1/2” wide. The front mahogany edging is mitered and rubbed in place. The side pieces have the front end mitered and a groove is run on the edge. The groove has to be placed precisely, so the baize will be 1/32” below the face. At the rear of the side pieces, a notch is formed to accept the tenon on the panel. The side pieces were dry fitted and 3/16” hole drilled through them and the tenon. The hole in the tenon was then elongated to allow for seasonal movement of the panel. The side pieces were only glued for about 4” at the front; the pinned tenon held the rear.
After the glue had cured, the slide was planed to fit the opening, which required only a few light passes. Like the drawers below the slide, a 7/64” wide by 1/16” thick solid satinwood perimeter inlay was applied. On the drawers, I did this in the router table with a 1/8” down spiral, but I was unimpressed with the quality of the cut. On the slides, I used a 7/64” end-mill in a laminate trimmer fitted with a clamped on guide. Running the trimmer with a climb cut, left a perfect chip free rabbet.
The slides were dyed, shellacked and rubbed out and then the baize could be installed.The high quality baize came from Londonderry Brasses. I have tried using spray adhesives to adhere the baize and they work well, but I really prefer the traditional hide glue. Gluing down baize can be frustrating; it is all too easy to have the glue seep through the baize and ruin it.
The mahogany edging is taped off and a glue sizing is applied to the panel. The sizing is allowed to dry over night and then hide glue is applied, being very careful to get full coverage and not allowing it to puddle. The edges and corners require particular attention, to see that they are covered evenly. The the glue covered panel is allowed to sit for about 30 minutes. Rough cut the baize to size and lay it in place. Mist the baize lightly with distilled water and with a clean iron set on medium high, adhere it in place. As with applying the glue, the edges and corners, are where you need to concentrate. You can’t let the iron touch the wood or the finish would be damaged, nor can you let the iron sit in any one place for too long, or you risk getting the glue too liquid and ruining the baize. Also, too much pressure will cause the glue to bleed through the baize. The excess baize was trimmed away with a new blade fitted in a scalpel and a straightedge. The goal, of course is to have a prefect cut, but sometimes a bit of the substrate will show. In those cases, if the amount showing is not more than 1/32” or so, you can heat the baize and push it over to close the gap. Stretching the baize to close a gap wider than that risks having it shrink back later. After all the excess is trimmed away, I like to go back an heat the edges and corners to ensure they are completely adhered.
The next installments will follow the building of the cases for the upper cabinets, and the installments will come much more regularly.
In entry four of this series, I spoke of the difficulties of fitting the lower doors, now it's time to detail the steps taken to accomplish that task. Fitting the doors, was a little like driving from Boston to Philadelphia by way of Los Angeles; you get there, but man does it take a while.
The first step in this “journey”, is to plane the doors to precisely fit the openings. By precisely, I mean without gaps of any sort, but not forced into place. Of course this and every subsequent step in fitting the doors has to be done with the base unit assembled on its leveled platform. Fitted this way, the hinges can be easily laid out, ensuring perfect alignment and leaves a margin of error for the final fitting after the hinges are installed. To my horror, I found the center doors came up with a gap of almost 1/8” between them. This made necessary a few extra steps to correct that mistake, which will be explained later.
The hinges I chose to go with were the precision butt hinges sold by Horton Brasses. These lack the hand cast look of the hinges from Londonderry Brasses, but unlike the Londonderry hinges, each one is exactly the same as the next and are free from any play between the hinge pin and knuckles. That last aspect of their construction is the important one, any slop in the hinge itself, translates into a sagging door and inconsistent fits, or consistent “fits” when installing the doors, depending on how you look at it . When installed, they are not so modern looking as to draw attention to themselves. The hinges are made, so when in the closed position there is a slight gap between the leaves; a desirable feature for inset doors, but not for overlay doors. The hinges were made from such heavy gauge brass, that I couldn’t swage them in the vise, to close the gaps between the leaves. I routinely do this with the hinges from Londonderry Brasses because they have a huge gap between the leaves in the closed position. Not being able to close the gaps, necessitated chiseling a hinge mortice with a tapered bottom. I thought of making a jig to route the hinge mortices. I ended up laying out and chiseling them by hand, but I did use a laminate trimmer to remove the bulk of the waste. The hinges were installed with short screws durning the fitting process, because they will be taken on and off several times. Later during the final assembly, full length screws will be used. Even with that precaution, I tried not to use more than one screw per leaf, so each screw would have its maximum holding power.
Now with the doors hung, they can be planed to have a uniform gap. For doors of this size a clearance of just over 1/16” is appropriate. At this point I found another problem; the flanking doors struck the hinges of the center doors. It seems that although I drew it out and was confident the projection of the center section was such there would be clearance for the flanking doors to swing closed, this wasn’t the case. It missed by about 1/32”, but that’s a lot when your shooting for a gap of just over 1/16”. The reason for this is the hinges were considerably heavier than I expected and I probably should have set the center section forward of the flanking cabinets by 1/8” more. It was a little late to do anything about those factors, so I sawed a small amount off one leaf, which allowed the hinge on the face of the case to be positioned more towards the center.
This of course required re-cutting the hinge mortices in the door. This whole episode pointed out how using one screw per leaf during the fitting process is a good practice. Had I drilled all the holes, my options would have been more limited, or the fix more complicated.
With the hinges moved, everything worked as it should, I was especially happy that the center doors lined up in the same plane where they met. I had expected to have to make minor adjustments with a plane and chisel to the faces of the cabinet sides in order for the doors to meet properly. The doors may have fit properly, but they had the raw poplar core, and the breadboard ends showing. To hide that, 3/16” thick strips of mahogany where glued to the edges of the door. I thought for a long time, on just how to precisely trim 3/16” from the perimeter of the doors. In the end the solution was easy and painfully obvious; use a 3/16” rabbet bit, followed by a flush trim bit.
This would not work where the center doors came together because of the oversized gap there. With the center doors still hung, I used a strip of scrap wood planed to 7/16” thick to scribe a trim to line on each door. The strip was centered on the gap between the two doors and a line scribed along its edges. When trimmed to this line, there will be a 7/16” gap between the doors. When the 3/16” strips are applied, to both doors, this will leave a nice even gap of about 1/16”. The original has an astragal at the joint between the two center doors, but this appears to be a later addition, to compensate for shrinkage. It just looks stuck on. A detail like this isn’t in keeping with the high caliber of the Seymour’s work, but perhaps they made the same mistake I did and corrected it with an astragal (highly unlikely, I know, but it makes me feel better, thinking it could have happened). Another bit of collateral damage from the oversized gap between the center doors, was the mitered frame of ribbon stripe mahogany veneer. Because the amount trimmed off wasn’t even, the miter on the veneer no longer landed at the corner. I had to re-cut one miter and replace the adjacent piece of veneer. The miters are no longer true 45˚ but that is unimportant; what is important that the miters fall on the corner. Those who question hide glues resistance to water, should try to remove some. I think they will find it quite able to stand more water and heat than they thought possible. The last step was to take one pass with a hand plane on each edge of the door to leave a prefect glue surface.
The mahogany strips, used to cover the edges of the doors, were ripped slightly oversized from a piece of rough sawn 4/4 mahogany. Between each cut, the edge of the board was planed by hand to remove the saw marks. This planed face was placed down on the bed of the surface planer when the strips were planed to thickness and later against the edge of the door. The strips were mitered to fit neatly around the door. In theory, since I’m using hot hide glue, the strips on three edges of the door could be rubbed in place and not need clamps, but I like to use 3M painters tape as clamps, just to be sure. The fourth strip, can’t be rubbed in place so here the tape is essential. If I felt that any strip took too long to get into place, I heated the strip with an iron to liquify the glue.
When the glue had cured the strips were trimmed flush with the router. Acting on a tip I saw on Woodsmith Shop TV, the doors were clamped together face to face with two pieces of scrap between, which left a gap for the flush trim bit, and gave a more stable platform for the router to ride on. In situations like this, don’t let the flush trim bit project below the base anymore than necessary. Every fraction the bit hangs below the router base is just that much more that could dig in, should the router tip. To reduce tearout, the router was run in a climb cut fashion. Dried bits of glue prevented the bit from getting the strips flush in some areas, so I had to use a block plane in those areas.
After chiseling the waste from the hinge mortices where the strips were applied, the doors were hung yet again, and the gaps checked. I was very pleased that only a few plane strokes were needed to have everything prefect, and I was glad to finally be in Philadelphia!
Inlaying the doors was fairly straightforward. The same templates used to size the veneer panels was again laid in place making sure the veneer nails were placed in exactly the same holes as when the veneer was trimmed. This will ensure the proper registration, so the groove for the inlay aligns perfectly with the junction between the veneers. In the case of the center doors, with their quarter circle indents, the groove couldn’t be done in one continuous run. The laminate trimmer must not be allowed to run around the corners, because doing so would leave a small radius on inside corners. The only real difficulties are getting the router bit accurately centered in the guide collar, which in the case of my trimmer requires some fiddling, and seeing what you are doing with the restricted visibility afforded by the laminate trimmer and guide collar.
The inlays surrounding the crotch veneer, is composed of two 1/16” layers of satinwood, with a layer of 1/16” black dyed veneer between. I measured the circumference of the oval and added 2” to arrive at the length to make the inlay. The individual strips of satinwood were sawn from solid stock, and sanded to a precise 1/16” in a drill press mounted thickness sander. The inlay was glued together with white glue, between boards covered in packing tape. I have a fairly large collection of C-clamps, but this 62” long inlay used each one and a few more would have been helpful. The individual strips were sawn off on the band saw using a thin kerf blade to minimize waste. Before the inlay was placed in the groove one end was cut with an acute angle as the first step in forming the scarf joint where the ends meet.
By some standards 3/16” is a narrow banding, but not when it is made of satinwood and has to be bent. I sawed the strips for the ovals to about 1/8” thick. This is far thicker then normal, but getting a inlay wider than it is thick to bend is difficult, because it wants to twist as it is bent. Due to the size of the oval, I didn’t expect too much trouble, but it took considerable effort to convince the inlay to cooperate. Start by brushing the hide glue on the inlay, and then work the banding into place. The starting end had to be held in place with a clamp. Using the iron and the veneer hammer the inlay was slowly coaxed into the groove. Strategically placed clamps helped to hold the inlay in place until the glue grabbed, which wasn’t very long.
Where the ends overlapped, a line was drawn showing the angle cut on the end and a razor saw was used to trim the end. Fortunately scarf joints on both doors fit perfectly, but if they hadn’t some careful trimming with a chisel would have made them fit.
The center doors were a little more problematic. The tight radius of the quarter circles in the corners would be nearly impossible to do with straight sections of inlays, so they were made as detailed here. Also the June 2010 issue of Popular Woodworking features and article on making the bent inlays. Popular Woodworking also has a video of the inlays being installed on the center door, which you can view here
The rings were cut free on the band saw and the waste in groove where the laminate trimmer couldn’t reach, removed.
To ensure the rings fit fully into the corner a slight bevel is cut on the end. Glue is brushed into the groove and the inlay pressed into place. A clamp helps to hold the starting end in place while the ring is worked into the groove.
The end is trimmed with a razor saw, again at a slight angle to ensure it fits fully into the corner. when the glue is dry, the ends of the rings are mitered with a narrow chisel. I would have liked to miter them before installation, but they were so rigid that it proved difficult to accurately determine where to miter the ends. Mitering in place isn’t difficult but I had to be careful not to damage the groove with the chisel. The straight sections were easy. The miters were again cut with a chisel, but this time the back of the chisel was used as a mirror to judge the proper cutting angle. It usually took a couple of trimmings with the chisel before they miters fit properly. I purposely left banding long, both to give me a few tries to get the proper angle on the miter, and when left slightly long the ends will be forced to a tight fit as the banding is pressed into place. This is much like a a sprung joint employed by interior carpenters.
The banding around the perimeter of the doors are solid satinwood 3/16” wide by 1/16” thick. These are laid into rabbets cut with the router, which was again was done with a climb cut. The miters on these pieces have to be sawn, because the chisel always undercuts a bit and this would show on the exposed edges. They are glued in place in much the same way as the mahogany strips surrounding the edges of the doors and when dry were also trimmed flush like before.
The locks can now be installed. As is almost always the case, the locks weren’t exactly what I wanted. A lock with the pin set 1” away from the edge would have been prefect, but the far superior double lever locks with screwed together cases do not come in that size. There, are so called “distance to pin” locks, which are available with a 1” offset, but these are somewhat inferior to the locks mentioned above, and they don’t come in lefts and rights. Even if they had, I would have probably opted for the double lever locks. The reason for this is the doors require the locks to stay shut, so they will be used frequently and the extra durability is important. I did use the distant to pin locks on the drawers, but here the drawers do not have to be locked, so they may never be used. I’m going to provide the customer with both a right and left replacement lock, in case one would fail, down the road. When inlaying the locks you have to keep in mind that left and right configuration, so you don’t end up with a mis-cut mortice. The keyhole escutcheons are the flanged type, so they are easy to inlay .
Inlaying the drawers follows exactly the same procedure used for the doors, the only difference is the inlays are 7/64” wide versus the 3/16” used on the doors. Lacking a 7/64” rabbeting bit (do they even exist?), I used a 1/8” down spiral bit in the router table to form the rabbet for the perimeter inlay. I had hoped the down spiral bit would cut smoothly, and for the most part it did, but I had one small chip taken out of the veneer. I would have been happy if that was the only veneer/router issue I had, but that wasn’t to be. While routing against the template for the inlays in the field of the drawer, the laminate trimmer snagged on a protruding veneer nail and I routed an arc in the drawer front; it was a nicely routed arc, just in the wrong spot. You can see a video whole ugly scene here.
Once the inlays had dried, they need to be made flush. This is probably my least favorite part of the process. The scraper would seem to be the best tool, but it tends to tear the inlays, especially the curved sections and at direction changes. Taking a slicing cut with the chisel will nicely flatten the curved section and the intersections. Take light cuts to determine best cutting direction, because the satinwood is prone to splintering.
The straight sections of the inlays are planed close to flush with the block plane and then the scraper can be employed.
With the grain of the veneer and inlays running in so many directions, a sort of diagonal slicing cut usually produces the best surface. Now with the scraping complete, the doors and drawers are ready to sand and finish.
The next installment will chronicle making the fall front drawer in the center section and the pull out serving slides.
Thanks
Rob Millard
As I said in an earlier entry, I was somewhat shocked that despite careful joinery the cases were quite flimsy. I guess I shouldn’t have been so shocked, because the sides are comparatively thin, as are the drawer blades, meaning the tenons are short and have no shoulders to speak of to resist racking. With that in mind, before any fitting of the drawers or doors could begin, the backs had to be made and installed.
On some of their larger pieces, the Seymour’s used a frame and panel back, so this is what I used. The frame and panel is ideal, because it provides considerable racking strength at a minimum weight. I’m not sure what configuration the back panels took on Seymour pieces, but I chose to go with three horizontal rails; top, bottom and one at the level of the lowest drawer blade. On the center section, the lower panel was divided in two, with a vertical stile.
To keep the weight down and save material, I re-sawed the panels from 4/4 rough sawn clear pine, which by time they were planed to thickness, ended up being about 3/8”. The frame was also made from clear pine, planed to 13/16” . Except for the exposed vertical edges of the flanking cabinets, which are rabbeted, the back panels overlap the edges and are flush with the outside edges of the cases. It may be a minor point, but I decided to size the rails and stiles so when viewed from the inside they are of a uniform width. I was coming down with the flu when cutting the stiles and rails, and I think I got one piece right; never work when you’re sick. I thought it would be cost effective, to use the table saw to make the cheek and shoulder cuts on the rails, but it proved no faster than using a mortise gauge and a tenon saw. I did use a router to make the mortices. I did find it a time saver to make the shoulder cuts on the table saw.
The inside faces of the panels were planed with a smooth plane for appearance, and the outside faces were planed with a scrub plane, also for appearance, that is to appear like period pieces. The beveled edges necessary to fit into the grooves, face the back, leaving the flat face on the interior.
I pre-finished the panels before assembly with a couple sprayed on coats of de-waxed dark shellac. Spraying shellac is not my strong suit, but these flat panels were within my capabilities. The dark shellac gives an aged appearance and seals in the pine scent, which in an enclosed space can be overwhelming.
The frames were glued together with hide glue, depressed with urea. I measured and remeasured the diagonals to ensure the panels came out square. After the glue had cured, I resisted the urge to belt sand the joints flush and planed them with a No. 3 smooth plane. With a joints flush, I brushed on a couple of coats of shellac on the stiles and rails. If I had to do it again, I would have masked off the contact points, so I could have glued and nailed the back on.
At the last minute, I decided to add a dust panel between the storage area and the lower drawers. These dust panels were also made as frame and panels. The dust panels were glued to the rear edge of the drawer blades. Since the frame and panels create a cross grain condition they were sized to leave about a 1/4” gap between it and the back, so if there is any shrinkage, the dust panel won’t force the back off or crack the sides. In all honesty, I think this is somewhat unnecessary, but better safe than sorry.
With the dust panels in place the back panels were nailed in on with hand forged nails; it is necessary to pre-drill in the hard birch. Traditional cut nails, are oriented with the long side of the head with the grain. The forged nails I make, tend to rotate as they are driven, so I keep a pair of pliers handy to align the head properly. See here for more information on making forged nails.
Before any fitting of the doors and drawers can begin, it is essential to have a stable, flat and level surface to rest the base unit on. I screwed together a frame from construction lumber. I believe I may have gotten the only two perfectly straight 2x4’s ever sold at Lowe’s, but I expected to have to plane them straight. I bought 12 footers but cut them to fit in the car, so they overlap in center. I hadn’t used a water level in nearly 15 years, but I dug it out for getting the frame perfectly level. I always liked using a water level for any distance longer than my longest level, because of its absolute precision. You’re probably thinking what is the chance, that the place where the breakfront will be used is so perfectly level, well it is slim to none, but I have to have some standard to build to. I’m certain where it will be placed will be more level than my garage floor, which it turns out slopes 1 1/8” in 8’. For repeatability, I marked the floor with a permanent marker, so the frame can be placed in exactly the same spot each time it is needed.
I started with the drawers in the flanking cabinets ( a later installment will discuss the center drawer). I made a DVD of the entire process of making the drawer from stock preparation to installing the hardware, you can see the details here. I also made a two part YouTube video on hammer veneering which you can view at the links below.
The first step in the drawer construction is the installation of the drawer runners. The dust panels act as the runners for the lower drawers. The upper drawer and the pull out serving slides (also detailed in an upcoming installment) need to have their runners installed in such a way that they do not constrain the sides, but are still firmly attached. On the sides that won’t be exposed, this is a simple matter of driving screws through the sides into the runners. The screw at the front is driven into the standard hole, but those in the rear has to have an elongated slot to account for any expansion/contraction.
The exposed sides represent more of challenge, but a clever method shown by Norm Vandal in his book on Queen Anne Furniture, neatly addresses the shrinkage and solid attachment issues. A picture hanging bit, forms a T-shaped slot in the runner and pan head screw driven in the sides, engage the slot, allowing for movement, but firmly attaching the runner. It takes a bit of trial and error with the screws, to get the runners firmly in place, yet still allow the runner to slide in place. As with the dust panel the runners stop short of reaching the back by about a 1/4”. To keep the runner front sliding back, I toe-nailed it to the drawer blade, with a small cut brad.
Period drawers typically have fairly thick fronts, with comparatively thin sides. In the case of these drawers, I made the fronts 15/16” thick and the sides are 11/32” thick.
The fronts are made from cherry, which is a typical substrate material used by the Seymour’s. The sides and bottoms are of clear pine, re-sawn from 4/4 rough stock. I like to make the drawer sides from bland looking lumber, for the reasons of appearance, workability, and stability. Re-sawing thick stock into thin stock can do funny things; nice flat lumber can be turned to spaghetti. Much of this has to do with how straight the grain is, which you can of course see, but unseen forces also play a part. How the lumber was dried can have a significant impact on stability, and I know of no way to “see” it or predict which boards will have problems. In the case of this lumber, it was well sawn and dried, and it remained true after re-sawing. Still, I stickered it for a day just to be sure any distortion took place, before planing to thickness.
When it comes to structural elements, I’m kind of a fanatic about stock preparation. Well flattened and squared stock sets the tone for the rest of the project, making for a hassle free, faster build. I flattened the stock in the same way as detailed here.
Much has been written about dovetails, so it seems almost redundant to add more. I was somewhat influenced by the Frank Klausz video, I say somewhat, because for me cutting the pins first, seemed unnatural. Cutting the tails first reduces the amount of layout and cutting, by allowing stacking alike sides and cutting them at once.
To me, the real heart of dovetailing, isn’t cutting a single dovetail joint well, but cutting multiples in a timely manner. Stock preparation is the start of this, and having a system of marking the pieces, so it is immediately clear which part is which, is the other. After that, it becomes rather tedious repetition. I start by plowing the groove so what was the outside of the tree, is the inside of the drawer. This, I believe is a minor point, but the concept is, if the drawer sides were to cup, the top and bottom would be forced in, at the half pins, where the joint is the weakest.
Despite very careful construction, there are inevitably minor variations in the drawer openings. The actual building of the drawers begin with sizing the drawer front to fit its opening. To keep things straight each drawer is given a number, as is the drawer blade below the opening. The openings were quite square, but to ensure a prefect fit, they were carefully sized with planes. With the bottom edge of the drawer front resting on its drawer blade, one end is planed to fit the vertical side of the cabinet. With that fit, the other vertical end is marked in place with a pencil from the inside, and sawn at the bandsaw, cutting away the line.
When the saw marks are planed away, the resulting side gaps will be very small but uniform. Now the drawer front can be planed for width, which at the same time will remove any minor chipping left from planing the ends. At this point, you will have a drawer front with the very small gaps at the sides and a heavy 1/32” clearance at the top, but no gap at the bottom. For now, you want to maintain consistency with the position of the groove for the drawer bottom between the drawer sides and front.
Sort the sides so you have matched pairs. It is surprisingly easy to make all lefts or all rights. I mark the front ends of the sides with the number of drawer it will become. Now the sides which were ripped about 1/16” over size, can be planed to match the drawer front. I like to clamp the jointer plane upside down in the vise and use this to plane the sides to width. Mounted this way, the planing goes very quickly and accurately.
The drawer back was ripped to width at the same setting as the rest of the stock, and when the grooves were ploughed the blade was raised to cut completely through backs. This leaves the back slightly too wide, but this can be addressed after the glue up. So that the backs perfectly match the fronts, the back is laid inside face to inside face with the front and the bottom edge of the back aligned with the groove in the front. Scribe, around the front, transferring its “shape” to the drawer back. Carefully saw and plane to that scribed line. I like to leave the back, ever so slightly bigger than the front. so after assembly I can plane for a precise fit. To keep track of its place and orientation, the back is numbered on its outside face.
Even masterpieces of period furniture can have dovetails of variable quality, but the Seymour’s were experts at fabricating elegant and precise dovetails, so I carefully laid out the dovetails to match their best work, but before any layout can begin, the inside faces of all the components are planed with a finely set smooth plane. The dovetails were sawn with an inexpensive Stanley dovetail saw.
Stacked, the sawing goes very quickly, and at this stage precision isn’t terribly critical except in one important area; here as durning the entire process, the base line has to be maintained. Lately, I have been using a fret saw to remove most of the waste between the tails, but I’m not sure it’s worth the effort. I came to this after seeing a video on the Popular Woodworking blog, on speed dovetailing. I even tried my hand at speed cutting half blind dovetails, which you can see here. In a way this speed cutting is kind of useless, but it can point out areas where your technique could use improvement at normal speed.
The sides are stacked in a stair step fashion, and the waste is chopped out, working half way from each side. To preserve that all important baseline, a knife is used to deepen the line left by the cutting gauge and a small chip is raised with a chisel. Raising that chip, creates a shoulder to rest the chisel against and prevents the wedging action of the chisel from setting it back. I find the fairly inexpensive Japanese dovetail chisels, to be indispensable for this. Their sharply tapered sides can reach into the corners and clean out the waste, without marring the edges of the dovetail. Care has to be taken not to stab the drawer side below when the waste is freed, but I have seen evidence of this happening on period pieces.
Now is where having a clear, almost foolproof system to differentiate the components, pays off. With the drawer front clamped vertically in the vise, the proper side is laid in place and carefully scribed to transfer the tails to the drawer front. I’ve been using a scalpel for this but care has to be taken, because of its highly flexible blade. Here is the one place I can see the advantage of sawing the pins first. The very narrow necks of the pins makes it somewhat difficult to get in there and scribe. The dovetails at the rear are wider and present no problem with scribing.
When sawing this time, not only must the base line be maintained, but the sawing has to be perfect. Not following the angle will result in gaps. When working with certain combinations of woods, you can saw a tiny fraction of an inch outside the scribe lines to ensure a tight fit. In this case with a cherry to pine joint, that amount of offset can be fairly generous, but less so at the half pins, for fear of splitting them off. On many pieces of period furniture, the saw cuts extend well past the baseline on the inside face of the drawer, which eases chopping out the waste. Apparently this was not the case with Seymour furniture, so I stopped the cuts on the base lines. This leaves a lot of material uncut, but using a card scraper as a blunt chisel will carry the cuts down to the base line, making the chiseling much easier.
Chiseling the waste from the drawer front is kind of brutal. Here again the chip is raised to provide a shoulder to work against. From here until you are about 1/16” away from the baseline, the chopping can be quite aggressive. That 1/1/6” is removed with the widest possible chisel, using only hand pressure, placing the chisel carefully in the scribed line. Most of the waste between the pins on the drawer back was sawn away with a coping saw, and the rest chiseled away as before. At this time the drawer front had a small amount planed off its bottom edge, so the gap around the front will be uniform all around its perimeter.
The inside faces of the components were lightly sanded with 320 grit paper to remove handling marks and two coats of de-waxed dark shellac were padded on with a paper towel. I took care to keep the shellac away from the glue surfaces. When dry, the shellac was scuff sanded. The last step before glue up was to cut small chamfers on the inside faces of the tails.
I don’t like to test fit the dovetails before assembly, instead I rely on careful scribing and sawing to ensure a proper fit. I use an interior hollow care door as the work surface when gluing, because they are very flat and it is important to have a flat assembly surface. To allow for more working time the hide glue has urea added. Even with the urea additive, time is of the essence, so have everything needed close at hand. I apply the glue to the front dovetails and tap in a side. You have to drive the side home evenly and stop at the slightest resistance and correct any problems. The back goes in more easily, and then the drawer is positioned on the bench and the glue is applied to the other two joints. Here, since you are working on two joints at once you have to work with particular speed. I like to quickly place a clamp across the front and back to drive the sides fully home, these clamps can be immediately removed. In theory clamping shouldn’t be necessary, but in practice it helps to close any gaps between the front end of the side and the half blind recess in the front. To prevent the clamping force from distorting the thin drawer backs, a piece of scrap is cut to span across the back. I begin the clamping by squaring the drawer and then place the clamps on, keeping the bars parallel with the sides. More often than not, applying the clamping pressure racks the drawer out of square, requiring you to reposition the clamps to force the drawer into square. I left the drawers in the clamps for several hours and out of the clamps over night before planing and fitting.
As the drawer came from the clamps it is just barely able to fit in its opening, so the sides have to be planed. To plane the sides the drawer is hooked over a padded board cantilevered off the bench. Despite not cutting softwoods as cleanly, I like to use the smooth plane with the high angle frog, because it can handle changing grain well. Of course you have to plane in from the ends so at some point you’re going to be planing against the grain and the high angle frog helps here too. The back is more difficult, because it can’t be planed on the hook, due to the dovetails not having any strength in the side to side direction. I have to clamp it in the vise and I use a low angle plane to place as little stress on drawer as possible. It took a few tests to get the drawer to slide properly in its opening, because the back prevented me from seeing where it was rubbing and I had to proceed cautiously. Don't worry if at this stage the drawer seems to bind. When the outside of the drawer and the inside of the case are finished and waxed, the drawer will glide like it is on ball bearings.
The drawer bottom was made from re-sawn pine which was planed to about 7/16” thick. I never use more than two boards in a glue up for drawer bottoms. I picked the boards for appearance, but most important I wanted the growth rings going in the same direction. Having the growth rings so oriented results in a uniform cup, whereas alternating them would have a wave like cross section, which is more difficult to plane. I like to use a rub joint for drawer bottoms, because applying clamps to thin stock, tends to cause it to fold up, plus there is just something kind of magical about creating a strong joint just by rubbing two glue coated edges together. Rub joints do require perfection when it comes to planing. After the glue cured the inside face of the drawer is planed with a smooth plane and the underside is planed with a scrub plane, which leaves behind the period accurate tool marks. It is easier to smooth plane the convex side if the bottom develops a cup.
Lacking any kind of beaded edge and featuring a veneered front, it is essential that the drawer sit perfectly flush in the opening. A minor amount of projection can be planed away, but anything more than a 1/32” will require altering the shape of the drawer. Only one of the four drawers did not sit flush, and it was out by around 1/16”. My standard practice is to lay a drawer that is known to fit properly, on its bottom board and trace around the inside transferring its size and shape onto the drawer bottom. Then I measure over the depth of the groove from that line and saw out the drawer bottom. Working this way ensures the drawer bottom won’t distort a properly fitting drawer. In the case of the ill-fitting drawer, the same procedure is used, only this time the cut lines are skewed to alter the shape of the drawer. The sides and front of the bottom are beveled to fit into the groove. I set the marking gauge to about 1/64th less than the width of the groove and strike a line with it. From there it is a matter of planing to that line using a well sharpened fore plane and skewing it, to get a smooth cut. It usually takes at least 2 test fits. It is this test fitting the presents the greatest danger to the drawer. Getting the drawer bottom skewed or jammed will almost certainly split the drawer sides, which are quite fragile due to the groove.
I once again test the drawer in the opening, but this time with the bottom in place to be sure it still fits flush at the front. If it passes muster then the bottom can be nailed in place. Minor adjustments can be made by placing a block behind the end of the drawer that goes in too far, and firmly striking the protruding end. This will rack the drawer so it will fit flush.
The next step is one that isn’t in line with the best modern practices, that is to apply glue blocks to lock the bottom in place. Locking the bottom in place with glue blocks can lead to the bottom cracking, but it does provide extra bearing surface and adds considerable stiffness to the drawer. I used a sliding T-Bevel to establish the cutting angle, and the glue block was bevel ripped on the band saw, so it stands proud of the drawer side by a strong 1/16”. The saw marks were planed away to provide the best glue surface, and the glue block is checked to see that it fits tightly, to both the beveled bottom and the drawer side. The glue blocks that run along the sides are kerfed at approximately 4” intervals. These kerfs stop just short of cutting completely through the glue block. The front end of the glue block is mitered. With plain hot hide glue, the side glue blocks are rubbed into place. When dry the blocks are planed flush with the sides making the one glue block into a segmented block. The rear of side glue block are sawn off with an angled cut at the rear. The front is handled differently; here the blocks are individual pieces rubbed in place with rather large gaps between blocks. The front blocks are mitered where they meet the side blocks. They too are planed flush.
The last step is to install the drawer stops, which are small blocks glued and screwed to the drawer runner.
The next installment will feature the inlaying of the doors and drawers.
With the lower cases glued together, attention can now turn to the doors. The doors on this breakfront are troublesome. At first glance the lower doors, the focus of this entry, would appear to be easier to make than the upper doors. After examining them more closely, and giving the construction considerable thought, I believe they will certainly equal or even exceed the upper doors in complexity. Obviously the upper doors have several technical challenges, but fitting them to the opening will be considerably easier than the lower doors. There are several reasons for this:
2. The lowers doors fit flush along the vertical edges and the top, so even the slightest
warping or twisting will stick out like the proverbial sore thumb; heaven forbid the
door developing a convex warp on the outside face.
3. The edges are covered with a thin strip of mahogany to conceal their composite construction.
4.The inlaid crotch veneer panels extend very close to the edges of the doors, making even a slight misalignment in centering of those panels, readily apparent (this is really an extension of number 1).
Of these considerations, number 1 and 2 are of the most concern.
To insure stability, the doors begin as strips of poplar roughly 7/8” square in cross section, which are then glued together so the growth rings are arranged in a quarter sawn configuration. This is admittedly a time consuming method, with the many joints to plane, but the result is worth it. The strips are laid out on the bench with the growth rings positioned correctly, so you can determine which edges need planed. The edges are planed with the jack and jointer plane. Instead of clamping and un-clamping each piece in the vise, a block is clamped in the vise and used as a stop to plane against. I’m not too particular about getting the joints prefect; as long as the joints can be closed with hand pressure alone and stay flat on the bench, that is good enough. My fondness of hot hide glue is tested when making panels like this, because the many joints and the quick gel time, even with the addition of urea, means only 4-5 strips can be glued together at one time.
My surface planer can’t handle stock wider than 12”, so that limits the width of the blanks. The center doors are just over 24” wide, meaning they have to be made up in 3 sections, while the end doors only need 2 sections. The sections are carefully flattened with portable power plane, straightedge and winding sticks, and then run through the planer. The now the sections can be joined. Unlike the individual strips, great care has to be taken joining these sections, because the panel must be flat and the finished thickness uniform. Keeping the panel flat is simply a matter of planing both mating edges at the same time, so any deviation from square is canceled out. With the sections being so carefully flattened, getting the joints flush along their length was easy.
After the glue had dried, the faces were planed with a smooth plane.
The original doors were made with breadboard ends, but in general breadboard ends make me nervous, especially under veneer. Still, they do add rigidity, and as noted warping would be disastrous. Given the built in stability of the panels, being essentially made from quatersawn stock, and the veneer cross-banding that will follow, the breadboard ends present little risk. To size the door panels, I used the same techniques as for the sides of the base unit, that is using the router and a flush trim bit to size and square the door blanks.
I used the shaper to run the tongue the door and the groove on the breadboard, which was made from birch. Just to be sure there wasn’t any mis-alignment, the breadboards were planed slightly oversized, and then planed flush after the glue dried.
The veneering as always, is my favorite part. The process began by preparing the substrate and the crotch mahogany. Both were coated with thinned down hot hide glue. On the veneer, this glue acts to both soften the veneer and to stabilize it, by completely saturating the veneer with glue. This glue saturation won’t eliminate the fine crazing, typical of aged crotch veneer, but it will lessen it. With the application of the glue, in a minute or so, even an extremely distorted sheet of veneer will become as pliable as a wet noodle. The sheets of veneer are wrapped in plastic food wrap and pressed between boards. Even though considerable moisture will be introduced when hammer veneering, I still like to let the veneer dry before applying it. To accomplish this, the veneer must be taken out of the plastic wrap occasionally over the period of about a week. While the crotch veneer is in the clamps, I worked on the substrate. The hide glue size applied earlier, will lessen the absorption of the glue used while hammer veneering, and this allows for a thinner glue, which in turn flows easier, meaning less force has to be applied to the hammer.
Just to be sure no grit or bristles from the brush were imbedded in the size, the substrate is sanded lightly, otherwise these could telegraph through the veneer. A layer of quartersawn cherry veneer with its grain running at right angles to substrate is hammered down. When the face veneer follows, the doors will essentially be made from lumber core plywood. This is hardly a period detail, but since it can’t be seen, and adds greatly to the stability why not do it? The cross-banding, of course won’t be seen, but I still took care to make tight seams between the individual sheets. When the glue had dried, the veneer was carefully examined for bubbles, by running fingernails over the surface; loose places will have a hollow sound.Those loose places are easily repairs by wetting the spot and heating with the iron and pressing with veneer hammer. I followed this with a light sanding, to remove any slight variations in the veneer thickness at the seams, which if left, could telegraph through the face veneers. Like the raw substrate, the cross-banding was coated with the hide glue size.
The face veneering on the center doors, began with hammer down the crotch veneer. Because of the hide glue sizing, this hammer veneering is not as mess free as I’d like. In fact the base of the iron became so throughly coated in glue, I had to stop several times to remove the glue with vinegar. Steaming hot vinegar is a smell you won’t soon forget, but it is the best way to clean off the glue. Despite the glue sizing taming the crotch veneer, it will still have a few problem areas, but at this time I couldn’t fool around getting it perfect, because moisture and the “pull” of the glue will start to warp the door. With that in mind, once the face veneer was hammer down, work began on the back. As noted before, the original has breadboard ends and I replicated that look in veneer on the inside face of the door. With the backs veneered the doors were now stable, which allowed me to stop working at such an accelerated pace.
While the glue on the face still fairly soft, I attached a template made from 1/4” plywood to route the crotch panels to shape. These templates are held in place with veneer nails placed in an asymmetrical patten, so the template can later be repositioned exactly to route the groove for the bandings. The cutting is done with a .040” end-mill installed in a laminate trimmer fitted with a guide bushing. The waste veneer was warmed with the iron and pealed off. Now the ribbon stripe cross-banding could be hammered down. There is one important, but often overlooked consideration when working with cross-banding and that is the orientation of the veneer. Highly reflective woods like ribbon stripe mahogany, must be installed with a consistent orientation, or the variations will adversely affect the finished appearance. In this regard the ribbon strip veneer is much like carpet, which depending on the viewing direction the intensity of the color changes. To keep that orientation, I marked the veneer with arrows at intervals along its length, with a white crayon.
The central doors have a detail that I don’t believe I have seen before, and that is quarter circles of figured mahogany veneer with the grain running at 45 degrees, inlaid at the corners. In another oddity, this detail is not outlined on the straight sides with an inlay. This requires careful positioning of the quarter circles in relation to the bandings which outline crotch veneer panel. I used a cutting gauge to accurately slice the cross-banding veneer in the corners to make way for the quarter circles, and then they were hammered down.
The flanking doors were handled differently, in that the ribbon stripe cross-banding was hammered down first. The cross-banding was not allowed to extend any farther into the center of the door than necessary because, the door can’t be veneered all at once, and covering the whole face, at this stage would have introduced an irreversible warp. Once the glue had cured sufficiently a template was used as before, and the crotch panel hammered in place. Now the back was veneered to mimic a panel with breadboard ends. The next installment will discuss making the back panels and the drawers.
Building this breakfront reminds me of working with sheet goods, because the panels are somewhat larger than I’m used to working with. Because I so rarely work with large panels, I lack some tools that would ease the process, namely a table saw with sufficient rip capacity, or a tracked portable saw. Of course the squaring up process begins with establishing one straight edge, with the jointer plane. Then the panels were laid out for length with a rule, square and knife and cut close to the layout lines on the bandsaw. Click on any photo to see larger version.
A miter plane makes short accurate work of squaring up the ends, and wetting the endgrain results in effortless, glass smooth cuts. As long as the wet shavings are removed from the plane, the plane will suffer no ill effects from doing this. Now the panel can be sawn and planed to width; leaving that until now, allows you to plane the endgrain and not have to be concerned with tearout, as the plane exits the wood. Here is where a table saw would help, because it is somewhat difficult to get a perfectly parallel edge. To aid in establishing an accurate width, I clamp together two pieces of scrap and use it as a gauge to check for parallel.
Since, all of the pieces have the same height; I began with a side panel to the center section, so I could use it for a “template” in conjunction with the router fitted with a flush trim bit, to quickly and accurately size all the other pieces. The center section side panel was chosen, because it is the widest piece and can therefore be used to size the other, narrower panels. In the same vein, the other panels were sized with the router, after first planing one to size and using it as a template for the others. After sizing the inside panels to the flanking sections, they had a mahogany strip applied to the front edge, with a rubbed joint. This will form the notch for the doors.
The depth of this notch has to be checked with the door blank and the scraps of the veneer that will eventually cover the door.(the next blog entry will detail the door construction). The other panels have their notches roughed out on the bandsaw and trimmed to exact size with, the router and flush trim bit.
Not having access to the original, I had to guess at the joinery. The joinery choices were also influenced by the changes made to facilitate shipping and installation, but none of these changes affected the outward appearance of the breakfront. From the photos, it does not appear that the drawer blades or the vertical panels are veneered, so this means the drawer blades are not joined to the sides with dovetails, but with a mortise and tenon.
Some of the tenons can not be through tenons; those would be on the solid mahogany end panels and front most tenon on the drawer blade of the center section. All of the other tenons can protrude though the sides for maximum strength. The mortises were started with a forstner bit, stopping just short of going all the way through, for the blind mortises, and into scrap wood for the through mortises. Squaring up the mortises is tedious, especially the through ones, but it has to be done accurately. For maximum accuracy, I ran the layout lines on both sides, so I could square up the through mortises from both sides to prevent breakout. This is where I made my first mistake because I neglected the fact that the panels have to be made in mirror image pairs, and one of the center panels, I scribed the lines on what would be the exposed face. Luckily, I have developed the habit of lightly scribing my lines and going back and deepening them, so it was easy to plane away the errant scribe lines. The front mortise on bottom drawer blades, are moved towards the back, to accommodate the notch for the door. I thought about using the table saw fitted with a dado blade to form the tenons on the drawer blades, but I ended up using the bandsaw fitted with a rip fence. The waste between the tenons was removed with a scroll saw.
The bottom boards are joined to the sides with through dovetails. The solid mahogany sides will be exposed, so the bottom board has a rabbet run on the outside end, which will place the dovetails safetly behind the base molding. I intended to glue a piece of mahogany back into the rabbet at the front edge where it is exposed when the door is open, but I think a stopped rabbet would have been better. None of these dovetails will be visible on the finished piece, so I just laid them out by eye.
I’m a dedicated tails first dovetailer, but with large panels like these that does present a problem, the transferring of the tails to the pin board. To overcome that difficulty, a straightedge board is clamped to the tail board, on the base line of the dovetails, and then the tail board can be clamped in place against this straightedge. Another advantage of this method, is the ease with which the inevitable cupping the in panels can be accounted for while scribing.
The waste between the tails was removed in the traditional method, but the tough nature of the birch would have taken considerable chiseling to remove the waste between the pins, so I rough cut away most of the waste with a coping saw and used (wait for it) the router and flush trim bit running against a clamped on straightedge; I’ve never had so much use for the flush trim bit as I have had on this job.
The top of the sections have three dovetailed spreaders. Like the bottom boards, these spreaders are rabbeted on the exposed ends, so the dovetails will be hidden by the top. I used pine for all of the spreaders except for the front spreader of the center section, which is birch. Birch was used, because the center “drawer” was changed from being a
At this point, the mahogany panels were smooth planed and the birch panels were smoothed with a card scraper. Dados for the drawer runners and the shelves were run with the router against a clamped on guide. The shelves in the center section are going to be 1” thick because it will hold a considerable quantity of china, while those in the flanking sections with their shorter span are only ¾” thick. To provide for some adjustability, three sets of dados were run for the shelves.
To remove any evidence of machine work, I used a router plane and side rabbet to clean up the dados. The dados for the drawer runners were also cut with a router, but I didn’t have an 11/16” bit, so I cut them with a 5/8” bit and then slipped a 1/16” veneer shim between the router base and the guide to widen it. This worked very well, but I had to learn the hard way that after inserting the shim, I had to feed the router opposite direction. The dados for the runners did not need to be cleaned up.
Next, came a step I labored over. I had made a piece for this same customer some years ago, and I don’t think he was happy with the visible and much lighter secondary wood, so I chose to color the birch to mimic mahogany. This was almost certainly not done on the original, but coloring birch to resemble mahogany does have a historical precedent.
Finishing, started with a careful sanding of the panels using a random orbit sander. The scraped surfaces with with 220 and 320 grit paper, while those that were planed, were very lightly hand sanded with just the 320 grit paper. The grain was raised with distilled water, and the contrast between those surfaces that were scraped and carefully sanded and those that were planed, couldn’t be more stark. The scraped/sanded surfaces suffered from extensive raised grain, while those that were planed, exhibited little if any raised grain. This only reinforced what I have always known, scraping is a rather crude way to prepare a surface, but often the only way, and it is no substitute for sanding. A planed surface is the ultimate, in both texture and flatness. I don’t believe scraping, sanding or planing, individually or in combination has any effect on the ultimate clarity of the finished surface.
With the surface prepared, the panels were dyed with a red mahogany aniline dye. To limit cupping the unfinished sides were misted with water. When the dye had dried, a couple coats of blond shellac were brushed on allowed to dry and lightly scuff sanded. I had originally intended to use Minwax Red Mahogany stain as a glaze to blend the colors of the mahogany, and the birch with its high contrast between the sapwood and heartwood, but that proved to be too red, so Bartley’s Dark Brown Mahogany gel stain was used instead. After the gel stain had dried, a few more coats of blond shellac was brushed on and when ready, rubbed out. I had worried that having one side finished and the other unfinished would lead to the panels warping, but that proved unfounded.
The glue up was pretty straightforward, using hot hide glue, depressed with urea. I took considerable care to square the case while clamping, but when the clamps came off, I was surprised at how easy it was to rack the sections out of square, so that time was effectively wasted.
Now with the drawings and cut list complete, its time for the construction to begin. The Mussey book, gives a description of the woods used in the breakfront, but not what those woods are used for. I decided to use cherry for the drawer fronts, birch for the interior parts of the lower case, and pine for all other secondary components.
Click on any photo to see larger version
Due to space limitations, both for the raw lumber and the scraps left over, I only ordered about 2/3 of the lumber. This way I can place another order later and not have as much left over lumber, which can quickly take over a shop and is just added expense. As usual Groff and Groff came though for me. I needed 4/4 20” wide planks of mahogany, and when I first called them, they only had 5/4 32” wide boards. I was not crazy about the idea of chopping a 32” plank down to around 19” and planning off 5/8”, but you do what you have to do. This bothers me not only from the standpoint of wasting a fairly rare board, but from a financial point too; 4/4 20” boards are about $9-10 a BF and 5/4 32” ones are $15.00 and up. The next morning they called back and told me they had some nice 4/4 mahogany that would finish out to 19” wide, and wanted to know if that would work. That phone call saved me several hundred dollars, and only reinforced my impression of them as a first rate dealer.
I started by constructing the base units. Taking care during the stock preparation phase will pay off at every future step. I don’t have the space for a decent size jointer and to be honest, the big ones scare me to death, so I will never own one. Instead I use a portable power plane, the modern version of a scrub plane, to flatten one face. But, before any tools are used, I thoroughly wire brush the faces of the lumber to remove any grit, which will prolong the life of the planer knives. Mahogany is typically well sawn and due to its inherent stability remains very flat and true when dried. Birch on the other hand is not as stable, yet this lot was very well sawn and quite true, which made it easy to flatten. The first step is to eyeball the board for any gross humps or cupping, and remove these. Then, with an accurate straightedge, (I use one made from a piece of straight grain cherry), check for straightness and plane accordingly. At this point, the board may very well look flat, but could still have a twist in it, so winding sticks are used to check for this. I have a pair of winding sticks I made from mahogany with holly and ebony inlays, but I prefer two framing squares and an off cut piece of melamine; the melamine placed behind the far square, helps to highlight the deviation or lack thereof, in the top edges of the framing squares. You can view a two part video of using a power plane to prepare stock here.
With one face flat, the stock can be sent through the thickness planer. Scaling the photo, showed the sides measured 11/16” thick, so for uniformity I planed the interior partitions to this thickness plus 1/32" for later clean up.
The mahogany plank for the sides, wouldn’t fit through my planer, so I flattened one face, with the power plane, taking off approximately half of the total amount needed to bring it to its final thickness; in theory, this keeps the moisture content equal and the plank flat. When used carefully, the power plane is capable of extreme accuracy. I hold it very lightly and kind of glide it over the surface. I also work with, across and diagonal to the grain to keep things flat. The progress is check frequently with a straightedge, winding sticks and a rule. With a face flat, a line is struck from that face, with a marking gauge, set 1/16” thicker than the finished dimension. That last 1/16” is removed with various hand planes. Starting with a fore plane, the mahogany is planed across the grain; even the most stubborn wood, planes like a dream across the grain, just be sure to have sufficient width that the inevitable splintering will allow the plank to finish out to the desired dimension. Again frequent checking is required with the straightedge and winding sticks. I like to start with the face that was planed to establish the rough thickness. This way the scribe line to establish the final thickness is on the opposite face and the two can’t be confused. After planing across the grain, I go with the grain using a 30” wooden jointer plane, which is light and very accurate. With an accurately flattened face, it is just a matter of scribing that other line and planing to thickness, following the same procedures for the first face.
The birch has to be glued up to achieve the necessary width. I had asked for 9”-10” wide boards, so I would only need two pieces for each panel. Groff and Groff did their best to fill that request, but most of the boards were in the 8” range, necessitating a 3 piece glue up on several of the panels. The vertical sides for the central base section have a strip of mahogany glued to the front edge because that section protrudes pasts the flanking end sections.
Typically I make edge joints with only hand planes, but lately I have been using the shaper to roughly straighten the edges, by offsetting in the infeed and outfeed fences (think of a jointer laid on its side). This has proven to be very accurate, despite the relatively short fences on the shaper.
I had not used birch in quite a while, and I can’t say it was a happy reunion; birch is notorious for its hardness and revering grain. With more cooperative wood, the edge joints are made by clamping face to face, the boards to be joined and planing both at the same time, which causes any deviation from a true square edge, to cancel out and results in a flat panel. The cantankerous nature of the birch made this a bit difficult, so I planed each board separately, which called for more attention to keep the edges square. The wooden jointer, despite its higher center of gravity, is much easier to “keep” square, and add to that, its 50 º bed angle made short work of jointing the edges. The edge joints were planed with a very slight hollow or spring joint, and joined with hot hide glue, which was retarded with urea. With accurate stock preparation, the glue up will go easily, but some misalignment will exist no matter how careful you are. I like to work with the boards on the floor and use my knee and hands to make slight adjustments to get the faces flush. A few minutes of careful work, will save a lot of effort when it comes time to plane the panels after glue up.
Planing the glued up panels is similar to the mahogany plank, but the panel should already be quite flat and true, so the straightedge, winding sticks and marking gauge won’t be needed. Start by planing across the grain, which is the only viable option with the difficult birch. With a very sharp plane and a light cut an excellent surface can be achieved by going across the grain. That surface is quickly and easily cleaned up with a scraper plane.
The next installment will detail the dimensioning and joining the panels to make the base sections.
Over the next few weeks, I will document the entire process of building a Federal period breakfront, from interpreting photos to make working drawings, through to the finishing. I will include all of the successes and mistakes (hopefully there won’t be many of the latter).
This entry will begin with a bit on the history of the piece, what changes are being made, and preparing the drawings.
The breakfront is in the collection of the United States Department of State, and is on display in the Diplomatic Reception Rooms. It is a massive piece, measuring 89 ¼” high and 96 ½” wide. Several sources attribute the breakfront to the Salem or coastal Massachusetts area, but Robert Mussey (see below for a list of the titles referenced in this article) in his excellent book on the Seymour’s attributes it to them. Whoever made it, it is an impressive piece; appearing quite delicate despite its massive scale.
As with any project, this one begins with assembling the reference materials. In this case there are several fairly good photos of the breakfront, in various books. The Mussey book has the best, in that it is taken straight on , is in color, and has a lengthy description of the piece. Yet, if using only this photo, you may think the center section is considerably taller and deeper than the flanking sections. The photo in Treasures of State is of somewhat poor quality but does show the relationship between the center and end sections, and also has a description of the piece. The photo in the Sack book is much clearer, and very accurately shows the relationship, between the center and end sections. A photo in Becoming a Nation is so small as to be of no real value.
Even a careful study of detailed photos, may not result in a true reproduction, because of course a photo will only get you so far, but often it is all that is available. When setting out to reproduce a piece from a photo, I always think of the Iwo Jima Memorial in the Arlington National Cemetery.
There is one error in it, which escaped the attention of Felix de Weldon and others involved in making the statues. By the time work started on the monument in 1951, M-1 carbines had been retrofitted with bayonet lugs, and these were used as a reference, during the sculpting. The original movie of the flag raising, shot by Bill Genaust, clearly shows the carbines carried by Ira Hayes and Michael Strank, lacked a bayonet lug. I bet the rear sights are wrong too, but enough of the obscure history. So, when I find a missed detail in my work, I feel better, knowing an important national monument isn’t perfect either.
The customer decided not to reproduce the Butler’s drawer. I can understand why, since it has limited usefulness in a modern household. I would have liked to make it, but it would have added considerably to the construction time. The original was built in only two sections, but I’m going to make it in 6 major sections; moving and shipping this behemoth is going to be difficult, not to mention expensive, enough without having it being any more unwieldy than necessary. This approach will cause its own set of issues, but those outweigh the alternative.
To prepare the drawings, I used a dial caliper reading to .001” to measure a part of the photo, for which there is a known dimension. Take that reading and divide it by the known dimension. This will result in a factor that can then be used, to determine any other dimension. For example, the photo of the breakfront in the Mussey book, measures 5.712” high, which when divided by 89.25(the actual height of the breakfront) is .064. Now this factor of .064 can be used to determine other dimensions on the breakfront. As an example, if some part in the photo measures 1.167”, its actual dimension is 18 ¼” (1.167÷.064 = 18.234 which is rounded up to 18 ¼”). Depending on the quality of the lens and how carefully the camera was leveled, there may be some distortion in the photo, so it is a good practice to determine a factor for both the vertical and horizontal dimensions. In this case they were very close, with the vertical factor being .064 and the horizontal .062. If the photo had not been taken straight on, the same method would have been used, but the horizontal dimensions would have been less accurate. In this case a certain amount of interpretation, would have been necessary, based on your sense of proportion; in other words drawing it out full size or to scale and seeing if things look right.
I make a quick sketch to set out the large overall dimensions. Depending on the complexity of the piece, this drawing may be enough to go ahead and start construction. The many components of this breakfront, requires more details and this is best accomplished with a full size drawing. The drawing doesn’t have to be a masterpiece; just good enough to get the details on paper, and have a document to take dimensions from. Here again, the dial caliper is used to pick out the fine dimensions. To achieve maximum accuracy I use a magnifying glass made for stamp and coin collectors, to insure the particular element is being measure precisely; even a small error in measuring will result in a large error when scaled up.
I make the drawings on brown kraft paper, which erases cleanly and holds up well during construction. I think the real benefit of the drawing is not so much the dimensions, which are of course important, but how the process of drawing them, fixes the interaction of the various parts in your mind, and allows you to “see” things you might not have otherwise. In the case of this project, I found the setback of the flanking cases, was somewhat dictated by how much the hinges for the full overlay doors project past the edge of the door. In a very real way, when the drawings are complete, you have already built the piece.
Despite, complicating the drawings to a degree, I wanted to have the plan show both the upper and lower case superimposed over one another. This means there are quite a few overlapping lines, so to keep things clear the lower case was drawn with a standard pencil and the upper case with a blue pencil. Thus drawn, there can be no discrepancies in the way the upper and lower cases align.
Now with the drawings in hand, I can move on to a creating a cut list; a necessary task, which I nonetheless despise.
In the next installment, I will tackle the stock selection and preparation for the lower carcass.
For me, the furniture of John and Thomas Seymour is the pinnacle of the Federal period so I was very happy to receive a commission to make a side table based on two examples of their work (see The Fine Points of Furniture by Albert Sack, page 270 and a work table in the White House Collection). I had made a similar table some years ago, which happened to one of the first pieces I felt satisfied with the way it turned out, but there were several areas I hoped to do better. One of those, were the curved inlays on the “turrets”. On that first table, a combination of heat, water and kerfs made with a razor saw, got it to bend, to what is a pretty tight radius. Kerfing sounds crude and it is, but when done with a razor saw and with careful placement of the kerfs, the results aren’t all that distracting. (Click on any photo to see a larger sample).
Wanting to avoid that crudeness, this time the inlay was made in a circular form. The inlay started as a 2” wide stack of 1/16” maple and black dyed veneers. With a straight inlay, the stack could made from more layers, but here for the ease of bending, it is only 12 layers thick. The finished width of the inlay has to be a precise 3/16”, so a dial caliper was used to measure the veneers that will form the “bread”, which will sandwich the strips that will be cut from the stack. With their thickness thus determined the strips are cut from the stack with a thin kerf bandsaw blade and an auxiliary table plywood table is fitted as a zero clearance throat, to keep splintering to a minimum. I use a scrap piece of wood of similar density as the stack to make trial cuts, until the proper thickness is arrived at. With consistent hand placement and pressure the bandsaw is capable of extreme accuracy, but I still check each strip to ensure they are all within .005” of the desired thickness.
On the lathe I turned a form that has a diameter about 1/8” less than that of the finished turret; this was to allow for some spring back as it would be better to have to open the inlay up, than to have to force it to a tighter radius. The strips cut from the stack were pre-bent around the form, which is a nice way of saying I broke them with my thumbs. Here again this sounds crude, but the finished inlay didn’t suffer in appearance. Glue was spread on one piece of the maple veneer, and the strips are quickly put into place, before the moisture in the glue hopelessly curls the veneer. Just as quickly this assembly is wrapped around the form with the strips facing the form, making sure the strips maintain contact with one another. Closely placed hose clamps apply pressure while the glue dries. After the assembly has dried for about 30 minutes, the next layer of the maple veneer is glued in place. The outer black dyed veneer is 1/16” thick and needs a little help to bend to the required radius; so it was lightly dampened and an iron supplied some heat to slowly relax it around the form. When formed, it too is glued in place with the hose clamps.
With the glue dry, it is time to saw off the individual inlays on the bandsaw. To give a solid bearing on the table, a flat was planed on the form. Because the round form is somewhat difficult to control, the inlays are cut thicker than required, to allow some margin of error.
A router bit run in a climb cut fashion, forms the rabbet for the inlay. Of course some hand work is required in the corners.With the rabbet cleaned up, I eyeballed a line that would miter the straight and curved sections together. The inlay is held in place and a razor blade laid on the line, cuts the miter. Hot hide glue is used to glue the inlay in place, using painter’s masking tape as clamps.
When full dry the inlays are flushed up, with a block plane and a scraper. The scraper has to be perfectly sharpened and used with caution to avoid tearout, always working in from the edge.
Recent Comments