Sunday, December 20, 2015

A Ming-Inspired Cabinet (20)

Post 20 in a continuing series....


Work continues on the set of beams comprising the cornice of the support stand.

These beams will join to one another with locking miter joints, and I tackled the miter cutting this time using the Wadkin dimension saw. This is a saw which was designed from the get-go to be configured for the convenient cutting of polygon miters, however in working through the process of cutting miters on these cornice pieces, I became aware of some shortcomings in the Wadkin's design.

The miter fence affixed to the sliding table, when set in its normal position, makes use of an etched protractor in the table surface, along with a sprung detent pin in the fence which fits into holes in the table top, a separate hole position being provided for the most common polygons. Of course there is a position provided for a regular 45˚miter cut, and if i were cutting the miter cut right through the stick, then this would suffice. However, when there is a tenon in the middle, one has to cut the miter from each face, and the same setting of miter fence cannot be used for both cuts. Ideally there would be a secondary miter fence, called a 'back miter' fence, fitted to the saw, and the back miter fence would be set to 135˚. The relation between primary miter fence and back miter fence would therefore be 90˚.

I don't have a back miter fence. They're a somewhat rare accessory, and now that I have had a chance to work a few miter cuts on the sliding table of the Wadkin, I am no longer thinking that a back miter fence is something I want to have any more.

Why not? Well, take a look at the set up when there is a back miter fence fitted:

Image is of a restored Wadkin PK saw, part of the Wadkin 'Temple' up in Ontario.

Notice how, in the above arrangement how the back miter fence really doesn't leave a whole lot of table surface for supporting the stick which is to be miter cut? That's a drawback, it seems to me, with the back miter. It would be fine if the sticks you were cutting were only 24" or less in length, but with longer sticks the majority of the material is going to be hanging out in space and that is an invitation for cut out error or mishap.

It seems to me that it would be a better arrangement to put the miter fence and back miter fence assembly together in the middle of the table for most purposes, as that would at least provide equal amounts of table support for cutting a stick against either fence. Why, something like this, an Altendorf Duplex Miter, would possibly be ideal for the same purposes:

Well, I don't have one of those. I'd like to get a look at one and see if it could be fitted to my saw somehow, or loosely copied in some manner.

I realized early on in mulling over options and in the subsequent fiddling about with setting up the saw for the miter cutting, that it would make sense to place the miter fence in the 135˚ position and work my way from there to the 45˚ position. You'll see what I mean as I go through the process.

With the fence in the 135˚ position (there is no table setting position for that angle unfortunately, so it is a matter of setting by eye and tightening down the fixing bolt), I took two pieces of scrap, jointed and planed them, and made a miter cut on each:

The two pieces were then placed against one another to see if a 90˚ angle resulted:

A 90˚ angle could only result if both cuts were exactly 45˚.

After the initial set up, test cut #1 yielded a result very close to the mark, and a clean miter line:

With the square held tightly however to one stick, a slight gap could be detected down at the end of the adjoining piece:

The fence was then to be adjusted and test cuts made again until the result was a 90˚ corner between pieces. This process was accomplished by setting up a stop block against the fence to start:

In the above photo, the block was set and then the fence swung open a little to create a gap against the end of the block...

...then a shim was added between the block and the fence so as to move the fence open slightly:

The initial shim I picked was 0.006":

0.006" proved to be too much, and in the end I found a 0.002" shim bumped the fence over the right amount:

Funny how people say that such tolerances as numbers in the thousandths of an inch are beyond what is needed for woodworking. I find the opposite is true.

With the 135˚ angle set, I used my large Mitutoyo precision square to establish a secondary fence oriented 90˚ to the first:

Around the same time I also took the opportunity to raise the blade and confirm that the saw blade was making the miter cut square to the face of the stick:

This jig referenced the cuts for length using a stop block for each position. I was able to make test cuts, halfway up the face of the miter, then flipping the stick for a second cut, and checking to see if the two cuts were co-planer. It was a matter of adjusting one or both stop blocks until I had the cut length the same on each side of the miter.

With the longer sticks, the process was repeated, however I had to add fence extensions so as to have the stop blocks located far enough to the rear:

The above picture clearly shows how this saw is best suited for cutting short pieces of wood. It's a bit odd though, I mean, it's not as if as if only short pieces of wood might need miter cutting. What were they thinking exactly, design-wise? I must be missing something.

I really need, I think, some sort of outrigger arrangement to carry longer pieces for cross cutting. Wadkin made an accessory for the saw which accomplished that, however I am considering whether I could fabricate something do accomplish the same end result. It will be something to explore when I'm not in the middle of building cabinets.

I didn't find absolute consistency was realized in my approach to cutting miters on the table saw, so I left the miters slightly fat of the lines and will pare them to fit later on. I would rather that the miters came out perfect right off the saw, but not this time. Some of the inaccuracy may accrue to the table itself, which is significantly bowed along its length.

With the miter shoulders cut, I used the bandsaw to remove the cheek portion and then cleaned up a face of the tenon on the router table. The next step was to mortise, and I brought the hollow chisel machine into action:

The mortising chisel is not quite long enough to plunge the full depth of the required cut, but got decently close.

The short ends all roughly mortised:

In the above photo you can also see that the tenon length has been trimmed back as well.

I took the short end pieces and ganged them, then clamped on some paring blocks, one on each end:

I was able to place the blocks with great accuracy using a digital caliper for checking the spacing of the blocks, and that allowed me to put the mortise end walls within a few thou of where they needed to be.

After the end walls were trimmed down, I worked the long grain sides of each mortise, like so:

This set is done:

Next step was to dress the other tenon cheek, which required an offset fence on my router table:

The offset fence allowed me to clean up the tenon cheek right close to the shoulder without fear of gouging into the shoulder with the cutter.

I brought the tenon thickness down to the mark, and checked the tenon/mortise fits of the short pieces to one another like so:

Of course, the short pieces actually connect to long pairs of beams, however those long beams are not through cut out yet to the same stage. Gotta leave something for tomorrow I guess.

Another look at one of these joints - they are only about halfway done at this point:

All for now my friends. On to post 21.


  1. The precision you impart to your joinery is astounding to me. Do you have to go back and fiddle with it again before the final glue up - do the joints move while waiting to be glued up?

    1. Ralph,

      many thanks for your comment.

      I rarely, if ever these days, do a 'final glue up', or any glue up for that matter. I trust the joinery to do the job all by itself, and I design my pieces around the idea that well-cut joinery alone is more than sufficient, and, in nearly all cases provides for later demountability so that repairs can be readily undertaken sometime down the line if need be. I might glue up a panel if I needed a width beyond what the material provided otherwise, however the glued-up panel then becomes an independent component within the structure. Also, longitudinal glue ups like you do edge to edge with boards, are the most stable arrangement between the parts.

      Putting that aside for the moment, part of your question seems to be whether the wood will move after the joints are cut, to the point that the fit can be spoiled or at least will be needing further attention before final assembly.

      The answer to that question is a bit complicated I suppose. The larger the stock is, the less dry it is, and the more prone to movement it might be (either in terms of a species which tends to move a lot or the timber having come from a log with significant growth stresses present), or the propensity of flat grain portions to move more than other areas, then the more that one could have significant issues with joinery abutments shifting after the joinery has been cut. One would be ill-advised, indeed, to take a bunch of green timbers, cut all sorts of joinery in those pieces then store them for a few years before assembly. Probably almost every joint would need significant reworking, and the outcome wouldn't be pretty.

      I cut a scarf joint once in a 4"x10" Douglas fir timber and other than a trial assembly, left the sticks apart for a couple of months afterwards. When I went to fit them, one of the timbers had significantly cupped across its width. A lot of hassle ensued, resulting ultimately in a joint which went together, but it was a good lesson to be sure. If you're working with green timbers, then the joinery should be cut as close to the time the structure is assembled as possible for the best possible fits. I myself do not work with green timbers, as later drying invariably spoils the fit of many of the joints, and in some cases can introduce splits right through connections which can seriously degrade their strength.


    2. With furniture work, the timbers are going to be well seasoned. However, as you joint and plane stock down, you will be able to observe how stable the material might be, and use that information accordingly. If the wood is persistently unstable, it is best not used to make a structure. A waste of labor really. Ideally you obtain a sufficient amount of material that you can carefully select for grain orientation for the parts. The grain orientation on frame members should always be that which inherently minimizes their movement - i.e., quartersawn or rift sawn. If you carefully work the stock down, re-jointing and re-planing to correct of movement, you should end up with stable stock. I've not had any significant problems with such stock afterwards when it comes to joinery, nor have I observed problems in service with those joints.

      One place with dry material where fitting problems could easily crop up, and have cropped up for me in the past, is putting together wide boards, which almost invariably have some flatsawn portions, with joints like dovetails or finger joints. Since the flatsawn portion of each board moves more than the other portions, it would be likely that a couple of wider boards which were fitted with dovetails might not fit each other so well a week or two down the line. For such constructions, it is ta good plan to glue-up the joints as soon after you have cut them as possible, if you cannot otherwise keep the material stored at the same relative humidity in the meantime. Failing that, the pieces should at least be stored in their assembled positions.

      I'll be doing some carcase dovetails on these cabinets, and anticipate dealing with the aforementioned issue, however one huge advantage I do have with this stock of bubinga is that the wide stock is actually quartersawn, and thus won't likely move too much. It takes a bit of the pressure off.


  2. Awesome Chris! Your work is ever inspiring.

    Hoping I can convince you to spoiler some of it....wondering what do you have in mind for the attachment between the support stand and the case bottom?

    1. Brian,

      as a maker yourself, you have discerned one of the technical issues with this design of cabinet, namely the attachment between the stand and the carcase, a case of frame and panel versus solid plank, and their different fundamental grain orientations. Well spotted!

      No spoilers to be provided, but the details will be revealed in the near future.


  3. Hi Chris, As usual, amazing.

    Would you be able to achieve the same miter cuts with a radial arm saw or a sliding miter saw?

    1. Barak,

      thank you for your comment. Sure, a radial or SCM saw would also work for taking those cuts. I could have also used my shaper's sliding table and the larger tenoning head, or used handsaws, cross-cutting and ripping.

      I've got various options available to me, and I elected to go the route of using the sliding table saw this time. I've tried most every method to cut the miters on this type of joint, and I tend to think that the sliding table saw holds the most promise for achieving results of the highest accuracy. However my saw is not up to that yet, as I have learned.

      My hope with the saw was that I could achieve the miter cuts I wanted right off the saw without further attention, however this goal was not achieved. I might have tried the miter work with shaper instead, however I need to pull the sliding table off of it and completely reset and align it, which is a full day job I've been putting off.

      Each time I cut a joint I am thinking of ways I might cut it better next time. Some of those explorations yield better results, while some do not.



All comments are moderated:

Anonymous comments will be deleted

Spam will be deleted

Comments containing links unrelated to blog content will be deleted