Sunday, October 16, 2011

SAPFM NE Chapter October Meeting: Chairmaking


Two of Matt Wajda and Bruce Eaton's period chairs.

Yesterday was the SAPFM New England chapter October meeting, conducted by Freddy Roman. It was hosted by Matt Wajda (pronounced "Wayda") and Bruce Eaton of Piscataqua Design. Matt and Bruce are furniture makers, both graduates of the North Bennet Street School, and the topic of the day was fine chairmaking. Their shop is in the same old mill building in Rollinsford, NH as Al Breed's shop; they sometime collaborate with Al and other craftsmen on specialized parts of projects.

Unfortunately, Freddy had another commitment, so he couldn't stay, but he had asked me earlier to run the meeting and keep things on schedule. I was happy to help, especially since my current project is a Queen Anne foot stool, with many similarities to chair construction. Once again, this was a spectacular opportunity to spend some time with outstanding craftsmen who are generous with their knowledge.

There were ten SAPFM members in attendance. I had met most of them previously at other SAPFM meetings or through the Guild Of NH Woodworkers Townsend document chest project with Al Breed. Freddy had expected more, but one of the attendees said the good weather was to blame; it was a gorgeous Saturday at the peak of New England leaf-peeping season.


Chapter coordinator Freddy Roman gets the meeting started.

Matt was the primary presenter for the meeting. He's an excellent teacher, very engaging and dynamic. He's done this presentation several times at the North Bennett Street School, and the information flows like a fire hose.

If you'd like to learn how to do this, Matt and Bruce are in the process of putting together a schedule of week-long classes. You can reach them through the contact page on their website to get on their interest list.

Matt said whenever they have a commission for multiple chairs, they always make one or two extra. This allows them to work out details and acts as a backup in case they make a mistake on one. He used two of these as demonstration models, a Queen Anne side chair, and an Arts and Crafts side chair.

As a result, some of the photos below show parts in a more completed state than they would normally be for a given operation, and switching back and forth does throw off the order of operations a bit.


Matt presenting at the workbench. Bruce is seated at left in the Chippendale chair from the first photo. I love how in high-end shops like this, they use spectacular demo pieces for everyday furniture. Of course, no one else wants to touch the pieces for fear of damaging them.

The beautiful curly cherry chest to the right in the photo above is a recently completed commission. Per the modern aesthetic, the back and drawer bottoms are all maple, and all surfaces inside and out are finished to the same high degree. 200 years ago, the secondary wood would have been pine or poplar, and the hidden surfaces would have been unfinished, even on the highest quality work.


Bruce shows the back view of the Chippendale chair.


It's hard to see in the photo, but there's a full-scale drawing of this QA chair taped to the whiteboard, with rough and finished stock lists.

Matt emphasized repeatedly the importance of having a drawing, so you can take actual measurements and angles off it. Just be aware of foreshortening due to angled and curved components in some views. He makes templates of every part directly from the drawing, using flexible 1/8" plywood. He makes shaper or router patterns from those out of MDF.


The ideal board for chairmaking is flatsawn, not quartersawn. That's because the chair sides are laid out on their sides, then rotated after cutting out. The result presents a quartersawn view from the front. If the board was quartersawn, the front view would present a flatsawn face, with distracting cathedral patterns.


When making a single chair, trace the sides from the pattern in a side book-match. Normally, this would be an 8/4 board, he's just using this thin one because it's easy to handle for demonstration.

After marking, saw the legs out rough. If using machines, attach the pattern to them and flush up with shaper or router. If using hand tools, clean them up with spokeshaves and rasps. Either way, use a handplane to flatten the reference flat where the seat mortise will be located.


For cabriole legs, cut them long enough for the knee blocks, then cut those off. Mark all three pieces with upward-facing arrow tick marks on the inside corner so you remember the grain orientation.

Mark each leg and its blocks with a corresponding number of tick marks. Then when you attach the knee blocks later, use the correct blocks for each leg, with the marks to the inside corner. This assures the best possible grain match per leg.


When making multiple chairs, trace out back-to-back pieces for each chair, but flip the pattern end-for-end to mark out the matching pieces.


Disassembling the A and C chair. It's like a big puzzle. It's also completely loose, no glue for now, but all the tenons are snug enough that it holds together.

Matt said the tenons should fit so they require hand pressure or a light tap with the mallet to go together, but no so tight that they have to be forced, risking a split. A lot of steps require partial dry assembly and disassembly.


Showing the relationship between thickness and mortise and tenon width. The tenon is 1/3 the stock thickness, with the mortise a matching distance from a reference surface. In front of the workbench you can see three additional chairbacks of more complex styles.


The "H" assembly of the QA chair: sides and back rail. Notice the pine block slipped into the right-hand mortise. This is a spacer used to protect the mortise until the seat side rails are assembled to the back, so it doesn't get crushed by clamping. That also maintains its moisture level.


Checking the alignment of the slat mortises in the crest rail and back rail.


The first of two jigs Matt showed us. This is the "stick" for getting the tenon right on the back splat or slats.

Make tenons on two scrap blocks, then seat them in the upper and lower mortises so the joints are closed properly. Glue a stick across the span of space between them. The stick jig captures the exact as-built relationship between the tenon angles on each end and the distance between them, for use later. Each chair design requires its own stick (if you aren't sure of consistency between individual chairs of the same design, make a stick per chair).


The second jig. After scribing the top length of the QA chair sides from some reference point such as the back rail, clamp the bottom rail of the jig across them without forcing anything into place.

Line up the jig top surface with the scribe lines. Then use the surface as a flat reference face for paring down the tenon shoulders precisely. This ensures they are exactly in plane.


From here, pare the tenon cheeks 90 degrees to the surface with a shoulder plane, keeping it on line straight across. A straightedge laid on the jig should be flat across the tenons.


Checking the crest rail against the resulting tenons (normally the crest rail would not be shaped yet).

Scribe the positions for the mortises and cut them. Note that the tenon sides are angled in (as are the mortises), because when the horns are cut off the crest rail, you don't want to be cutting into the joint.


Examining the mortises inside the QA crest rail. You can start shaping the rail once they are done.


Here's where the stick comes in. Lay it on the side of the splat blank and trace the tenon angles and shoulders exactly. Then connect up these ends with the curved template. Saw out the splat and pare to the lines.

It's worth doing the paring in multiple steps, test fitting the joints between splat and shoe, and splat and crest rail, as you go. A poor fit here will really show.


The flared points on the top of the splat have very delicate short grain and will chip off easily until assembled to the crest rail. Leave extra ears on them, then trim to the pencil lines and crest rail after assembly and glue up.


The splat is chamfered to give it a graceful lighter, thinner appearance. If it was really as thin as it appeared, it wouldn't be strong enough.


The splat in the shoe to check the joint.


Assembling the crest rail onto the tenons.


To glue on the QA crest rail, clamp a cross block to the legs, then run bar clamps from the block up to the horns on the rail. That's their purpose. Trim them off once the glue has dried. Then you can do final shaping.


Reassembling the A and C puzzle.


Fit the seat rails into their mortises; their tenons are angled to form the trapezoid. Then fit the front legs, but leave out the front rail.


With the joints all closed properly inside and outside, measure the precise front rail distance. Now cut the front rail and its tenons.


With the front rail in place, showing where the glue blocks go inside the corners.

The front corners require a "Philadelphia block" spacer on one side of each leg to bring the surface out to the adjacent face. The main glue block then spans the spacer and leg.


Closeup of the seat joint at the back. These are the only through-tenons in the chair (some designs don't have any). Pull the side rails out, saw down the tenons to make splits for wedges, then reassemble the seat and drive wedges in.


This diagram shows that the split and wedged tenons turn into dovetails, for a good strong joint.

Most chair failures occur at the joint of seat to back, especially those using dowel joinery. In contrast, Matt calls this method of joinery "the 200-year joint".

This was an extremely informative session, full of details. While a chair is a complex piece, subject to heavy stresses in use, Matt broke it down to a reasonable set of operations. I'll definitely be building one of these Queen Anne chairs. Our house is going to have a very strange selection of furniture as I build one of these, one of those, as the whim strikes me.

1 comment:

  1. wow! this is an excellent summary! lots of good tips here. so grateful for you taking the time to narrate this so clearly. bravo, keep it up!

    adam

    ReplyDelete