A mounting conundrum

Before we can proceed to shape the carcass, the mechanisms must be completely installed and operating properly. With the mechanisms in place, the challenge becomes mounting the main box and tabletop to them. The tabletop is relatively straightforward: center it on the carcass, then mark the top racks. The main box, however, isn’t so simple. Not only does it have to be located left/right, front/back within its opening, but it also needs to be situated in height above the carcass floor. If it’s too high, the bolt on the latch won’t catch; too low, and the bolt will drag on the main box as it opens. To complicate matters, it is impossible to directly mark any of these components. They’re buried deep in the carcass interior!

Being the case, I’ve devised the following plan…

  1. With the tabletop removed, center the main box between the lower racks in the carcass. This dictates the thickness of the shims that will be needed.
  2. Remove the top racks such that the main box can be lifted out versus slid through the front.
  3. Coat the latch bolt with graphite.
  4. Retract the latch bolt by turning the key, and place the main box into the carcass. Align it appropriately, then release the latch bolt. This should imprint a line on the underside of the main box where the latch makes contact.
  5. Align the back of the hole in the metal latch plate with this line.

While I’m certain some level of adjustment will be necessary, this methodology provides a plan to move forward. Such is the nature of product development or, to put it another way, if we knew what we were doing, it wouldn’t be called research!

The shape of things to come

While making the prototype, I’ve had to keep the aprons rough to facilitate installation of the mechanisms. No sense in making them curvy when I’ll need to stabilize them for mortising. This week, however, the project is starting to turn a corner.

To install the lower racks, there need to be defined boundaries. Unlike the upper racks which extend beyond the carcass, the lower racks are contained within the carcass, and are unseen. This means knowing the extents of aprons once they’re shaped. I created a template from photos of the original, and referenced it from the center line on each side block. Repeat the process on the undersides. These lines can now be carried down the verticals on each leg for reference during final shaping. 

Legs! – The making of…

I wanted to back up a moment, and explain how the legs got their start…
As with any cabriole leg, it begins with a pattern. From where this pattern originates is a mystery at least to me. I’ve consulted with several folks, but no one seems to know of a good design reference on the subject. The best I’ve found to date is the information provided in Jeffrey Greene’s American Furniture of the 18th Century. In this case, I am fortunate to have pictures of the Oeben, both from the Getty’s website, as well as from my time with the original.

Front view of the Oeben table

Front view of the Oeben table

With this photo, and knowledge of the leg height, I have everything necessary to re-create the profile. Simply project the image onto a sheet of paper with two marks representing the height of the leg, then trace! The profile is symmetric, being cabriole, so this single profile can used on both sides of the leg blank without concern. Many thanks to the ladies at The Arts Council for allowing me use of their projector for this exercise.

At first, the curvature of the leg is subtle. That’s because the knee is hidden by the ormolu. Same for the “feet” which are covered by sabot. I had to interpolate in these areas which isn’t really a concern since, if I got it wrong, a) I can refine the profile during later stages, and b) it’s going to be hidden by ormolu anyway! The profile can be transferred to template stock, 6 mm thick plywood in this case. I use graphite paper for this…

Leg profile on template stock.

Leg profile on template stock.

Now, it’s a straight-forward exercise in coarse, medium, and fine…Begin sawing out the rough profile, then refine it successively using rasps, scrapers, files, and sandpaper. The resulting profiles should be smooth, continuous curves without noticeable discontinuities. As I work, I’ll run my fingers along the edges while looking away to test. It’s easier to feel imperfections.

Tools used to refine the leg profile

To refine the template, use rasps, scrapers, files, and sandpaper.

With the template “faired”, choose stock for the legs. The legs of the Oeben table are completely veneered, so none of the “core” grain is visible. If, however, you aren’t planning to veneer your piece, consider the grain direction of your blanks carefully. Lines running diagonally through the end grain will yield straight grain down the legs. Lines parallel to the faces in the end grain will produce bull’s eye patterns at the knees. Mill a face, and a perpendicular edge. Don’t bother with the other two sides. They’ll be removed in subsequent steps. Just be certain to use these two reference faces for all of your layout.

Cabriole leg layout

Trace the template profile on two adjacent faces of the leg blank.

Trace the template onto the blank aligning the “back” along the arris between the two reference faces. The “back” is defined to be the flat portion of the leg block at the top which winds up inside of the piece. You’ll have to flip the template over when switching faces. I like to strike lines defining the top and bottom extents of the template, and carry those around the blank to aid in alignment. Another point I neglected to make in my previous posting on the legs, it’s best to layout, and cut the mortises while the blank is still square, before you commence cutting.

Hopefully, this provides a more complete explanation of the cabriole leg process…

Why Build Prototypes?

You might wonder why go through the effort of constructing a prototype when building a new piece? After all, it requires material, and at least as much time to create it, as the “official” piece. Why not just apply those resources to the finished piece?

So here’s my story…

I embarked on this Oeben re-creation, after much study, by constructing a three-dimensional SketchUp (SU) model. The thought being build it in SU to work out the bugs in the process. This may come as a surprise to you, but ALL of the aprons on this table exhibit a curve to some degree (fortunately they’re not compound!). There are very few flat exterior surfaces on the piece! At the time, I thought the easiest way to handle this would be to begin with thick material, then shape it appropriately. I drew it this way, and didn’t look back. Life was good… Or was it?

Now it comes time to cut wood. So, I choose an appropriately thick piece of stock for the “Main Box” front. The “Main Box” being the “drawer” that extends forward when the top is released and retracted. It contains all of the other “hidden” components. For several weeks, I continue working – yet, something keeps bothering me. Each time I look at the curve for the front, it seems out of place, like it’s inappropriate, and not representative of what I see in the original. Finally, I sit down, and look through the photos. I come across an image taken of the bottom. It’s available on the Getty Museum’s website. I’m showing an outline of its shape below. The two open spots at the bottom are where the legs protrude.


From it, I calculate the extent of the curvature at its most extreme, and learn that my drawing is about 1.5 cm too shallow. Not necessarily a show stopper. But wait – I look closer. Not all of the aprons were created in this fashion! It appears that the board comprising the rear apron was bent. Looking further still, I come across this photo showing the underside of the “Main Box”.


It shows a uniform thickness board bending in a gentle arc composing the front of the “Main Box”. So, it appears that my plan to work a thicker board to shape is incorrect. Now the question becomes how did Oeben’s craftsmen accomplish this?

There are several methods for creating curves in wood. The most basic is to cut perpendicular kerfs along its inside. This leaves voids allowing room for compression. However, these spaces are highly visible, and from the previous picture, not in evidence in this table. A second method involves slicing the board into thin planks along the grain then gluing them back together while clamped to a form. Once the glue dries, the board maintains the shape. I have witnessed instances where this was done and difficult to detect, but I don’t believe that this method was used either. A third procedure involved steaming the board and bending it. While this seems to be plausible, was this historically correct for the period in which the table was made?

Following a brief exchange with Patrick Edwards, this appears not to be the case. According to Patrick:

Oeben did not steam bend furniture.  French curves are sawn from solid wood.  Although Denis Papin (1647-1712) invented the steam pressure cooker, no one thought to use it to bend wood until the 19th century.  Complex two dimensional curves are “sculpted” as you suggest using chisels and scrapers.  Final truing of the surface was with toothing planes.

So, my initial methodology was correct, and the aprons were hewn from thicker stock.

If I hadn’t taken the time to construct the prototype, I may have missed out on this learning opportunity. Not only do prototypes help you work through unforeseen issues that weren’t caught during the modeling phase, and provide valuable practice for building the final version:

If you can’t get it right on the prototype, how do you expect to get it right for the final piece?

The process of creating a prototype gives insight into how the piece was originally made – living archeology, if you’re paying attention!

So, what becomes of the current prototype now that this issue has come to light? After all, good material and effort has already been expended to reach this point. Rather than discard it, I decided to laminate material and “thicken it up”, upon consideration. While this will allow me to obtain the appropriate external shape, the interior will remain as it is presently. It’s a compromise I will live with for now…