About ten years ago, I was between shops when I got a call from a historic church in Half Moon Bay that had some woodwork in need of replacing. It was an hour and a half away, so I asked them to send me a photo. After all, why drive that far unless I knew it was a worthwhile job? When I saw the photo they sent me, I knew I had to do it!
The Community United Methodist Chapel is the oldest operating church in San Mateo County. Originally built in 1872, the chapel was thrown off its foundation in the 1906 San Francisco earthquake. A year later, the church was rebuilt and extensively renovated. Many of the furnishings in the sanctuary are original; the bell that was installed in the cupola in 1907 remains to this day.
During my inspection, I found that the beautiful old millwork above the entry stained glass was extensively rotted after 130 years of exposure to the ocean air. All that held it together were the remnants of rusted square nails, layers of paint, caulking, and flashing.
. . .
|The job was to remove the old millwork, take it apart, extrapolate the original geometry, get it milled, re-assemble it, and re-install it like new…. Yikes! But, YES! I wanted that job.|
In the end, how did I get the job? No one else was crazy enough to touch it.
My partner Vince and I packed up our demo tools and took along a 4X8 sheet of plywood to capture a preliminary layout. After recording a lot of measurements and reference points (width, height, inside measures, outside measures, rises, and spring points—more on that later) we were ready to remove the original piece.
We had prepared for a long, arduous removal process, but, surprisingly, once we started, the molding almost fell off the wall. Those old hand-beaten square nails were amazing—they didn’t have any corrosion protection, but they were able to hang on for 140 years.
We removed the entire piece in four main sections, then hauled it home, where I could spread it out and draw up the exact layout—each radius, each individual molding, and the substrate foundation.
The profile stack was a total of 6 in. thick and 11 in. wide at the base. The trick was to draw up all the individual profiles and lay them out in proper orientation.
It took about a week for me to lay out the whole composition and test the old parts to be sure they fit. I did a complete full-scale drawing using two 4X8 sheets of 1⁄4-in. MDF, which I divided up the center line. That way, I could give one half of the layout to the mill shop as a template, with paper profiles printed from my computer glued right on the template. The template included computer layouts for each profile along with their radii.
Sometimes I see things differently, and it seems that whoever was connecting up the wires in my head had a sense of humor that day. When I looked at the front of the chapel, what I saw was a pattern of circles (see photo, right). The challenge was finding the center point and radius for each one. After that, it would be a simple matter of identifying the intersection or tangent points. The first circle was easy—the radius was half the width of the door opening, from casing to casing. It seemed to me that all three circles shared the same radius, but I had to confirm that.
I used a very simple technique which isn’t exactly precise, but it’s effective. I later checked my measurements against the existing millwork, just to be certain that each radius was accurate.
The existing moldings were joined at the tangent point, or spring line, where the two radii met in the center of the s-curve. I simply measured the cord length from that point to the top of the upper radius, right beneath the finial.
With a construction calculator, you can enter those dimensions as RISE and RUN, then press the CONVERT key and the ARC (radius) key to solve for the radius.
If you don’t have a construction calculator, use this formula (X = Run; Y = Rise):
[(X ÷ 2)² + Y²] ÷ 2*Y = Radius
Making a full-scale drawing allowed me to check the length of the radius. I was able to hold the pivot of my trammel arm at the tangent point in the molding and swing an arc up and out to the left side of the door, near where I thought the center might be for the upper radius. I swung a second arc centered off the base of the finial. The intersection of those two arcs identified the center point for the upper radii. I repeated the same steps on the right side of the door.
I was able to measure the ends of the original woodwork, which made it easy to measure and draw the individual profiles and steps. Each profile meant stepping the radii back the width of the profile. I drew each individual profile on the rough template. After a little fiddling and adjusting, I managed to come up with a geometrically true template, which was time-consuming, but simple—anything that didn’t fit correctly or looked “kinky” wasn’t a “fair curve,” which meant a small adjustment to smooth the lines out.
I also used the full-scale drawing to help me when it came time to enter all the information into AutoCAD. AutoCAD also gave me one more opportunity to check that each piece landed precisely. Once the drawings were complete, I carried everything—including half of the template, and the computer files to Haas Woodworking.
I’ve worked with Haas for years. They’re a fantastic mill shop. Their history mirrors that of San Francisco. The shop, founded in 1887, is still family-owned and operated; though today, instead of belts and pulleys, you’ll find the shop is full of modern shapers, power feeders, and CNC machines—everything necessary for making those fantastic S-curve profiles, which minimize joinery and simplify construction and installation considerably.
As I said earlier, I checked and double-checked everything throughout the long process of building this casework. Before assembling all the milled redwood onto the plywood base, I joined the top and bottom plywood forms with biscuits and took one of the layers to the jobsite.
That trip turned out to be a good idea. Because of past earthquakes, and decades of settling, the doorway was 1 1/2 in. off level across the top. Using the plywood base as a template, I was able to align the arch and finial so that it pointed directly at the window above, a step that enabled me to build and install the final composition without any further on-site trimming.
The original three-centered piece was probably assembled with hide glue, but I used modern glues and adhesives —this made a tremendous difference in assembly, and will no doubt add years of longevity to my work. I used urea formaldehyde to glue the two plywood layers together; and I used marine epoxy to bed moldings onto the plywood background, in combination with biscuits for securing the mitered joinery. On top of that, I fastened as many parts as possible with stainless steel screws through the back of the plywood.
First came the plywood base, made from two almost identical layers of marine plywood cut on a CNC machine, which made assembly very easy. The two pieces forming each layer fit together perfectly. The CNC machine even allowed for an offset in the joint between the two layers. Once the base or foundation was complete, I had a reliable form for precise placement of each additional layer.
I started installing the moldings with profile D, because it was flush with the edge of the plywood base. Next I installed profile C. The gap between D and C was intentional; I wanted room for expansion and contraction, plus a little wiggle room so that both moldings could be located precisely on the foundation. To install profile B, I used a spline joint. That piece had to be installed accurately, and I didn’t want any fasteners to ruin the clean face of such a delicate molding. I cut the grooves for the spline joint on my router table using two custom curved fences that matched the radius of each molding. I used bending plywood for the 1/8-in. spline.
In fact, profile B was the “key” to unlocking the entire assembly process. With B set temporarily in place and secured by the spline, I was able to place profile A against B and mark for critical end cuts and intersecting miters. Also, once B was positioned on the spline, I was able to slide profile A up along B, until A met the center line. That’s how I marked and laid out the most critical cuts—like the long acute miter following the center line.
At this point, I had to work off the actual piece as opposed to the MDF layout. First, I marked the top where it crossed the 6 3/4-in. horizontal line at the base of the finial. After cutting that line (a nerve-wracking but necessary step), I made a tick mark at the top and a tick mark at the bottom where the S-curve crossed the center line—the same simple method I used for marking miters on ordinary casing around an arched door. In this case, the tick marks were a bit more daunting: The top tick mark had to perfectly split the 1/2 round at the top of Profile A.
To cut that miter, I flipped the moldings over and drew a straight line from the upper tick mark to the bottom tick mark. I rough-cut near the line on each piece with my bandsaw, then used a straight-edge and a series of flush trim router bits to clean right to the line. Yes, my palms were sweating, and not just because the material was 2 3/4 in. thick.
Details That Matter
|Once all the pieces were cut and everything fit perfectly, I glued up the final moldings. There were no fasteners piercing the face, anywhere. This way, nothing breaks the smooth, continuous flow of the moldings.|
|The original geometry of the piece contained a bad water catch, which led to serious water damage and rot in the original woodwork.|
|I filled the deep pocket with marine epoxy, and shaped it so water would drain out of the area. None of this is visible from the ground.|
|Installation took only two days…|
|and most of the second day was spent on the step-flashing.|
I was in junior high school shop class when I first fell in love with wood—the colors and grains of different woods joined, shaped, and polished.
A mix of teenage-jobs later, and a year and a half as a gravedigger in Monterey, at 20 years young I became a fireman. My schedule left 20 days off a month, so I was really able to get back into making beautiful things from wood.
An early gift was to learn from a Swiss master carver, Fritz Abplanap, who first came to America in the 1920s to carve a convent chapel with full-size saints and angels, all in black walnut. With him, I first experienced grace in the sweep of a gouge.
I discovered stone carving and spent years in a realm of sensuous forms and smooth polished surfaces of alabaster and marble. I carved by hand, hammer on point, tooth chisel, rasps and sanding, and more sanding, polishing and more polishing.
Through a series of dramatic impulses and events in 1977 I was now called Sangeet and spent four years in an Indian ashram, on my path to enlightenment . . . but while meditating on that elusive carrot, I spent a good part of my time doing some wonderful woodworking under very simple hand tool conditions. There is nothing quite like getting teak and rosewood delivered by bullock cart and cutting dovetails by hand.
The ’80s brought several tsunamis of change that swept me through some carving jobs and body surfing in Hawaii, and onto the mainland US for more time in a spiritual community. From there, I took a woodworking job in Munich, Germany, but ultimately washed ashore just north of San Francisco, where I continue to roam to this day.
My time in the San Francisco Bay Area has blessed me with some fantastic projects and mind-bending challenges, working alongside amazing craftsmen. It is here that I fused my love of geometry with woodworking.
From my well-equipped shop in San Rafael, I keep looking for better ways to meet the challenges that come my way. And I am always open to that next wild project! Feel free to email me at email@example.com.