Post 58 in an ongoing series describing the design and construction of a kabukimon, a type of Japanese gate. This is a project for the Museum of Fine Arts in Boston.
1. a feeling of vexation, marked by disappointment or humiliation
Things were going well it seemed after I had tenoned some 22 rails and battens for the three doors on this gate. The tenons were cleanly cut and within a tenth of a millimeter of target dimensions and position. All good, or so it seemed.
Yesterday was going to be a day of doing a bit more processing work on the tenons, however after doing a bit of fettlin’, I discovered that the shoulders of the tenons were not exactly 90˚ to the length of the sticks. They weren’t way out, but probably on the order of half a degree I would guess. My initial response was “this sucks!”, and that was immediately followed by, “how did that happen?”.
When I bought the shaper, one of the primary aspects which drove the purchase in the first place was the presence of the tenoning table, a rather costly factory option on the Martin shaper not often seen. When the machine became available, I was lucky enough to be in a position to afford it. What sealed the deal was the excellent condition of the machine, particularly in regards to the tenoning table, which had apparently seen almost no use. “Maybe it was used on one job”.
Everything else Martin I have worked with has been pretty much set and forget as far as angles and settings of movable parts. If the jointer fence is moved out to some angle and then put back to the 90˚ stop, you can be sure it is 90˚. I used to check all the time, as that is what I had learned to do with previous equipment, but after the dozenth check I have learned to accept that the machine is providing repeatable accuracy. And that is why I made the investment in the first place. I hate screwing around with equipment that doesn’t have seem to be able to provide the basic functions with any reliability or accuracy. That describes most woodworking equipment actually. In most cases, a woodworker getting accurate results from his equipment is a woodworker who is nearly continuously fiddling with, and checking, that equipment. Those that do not will find their results are predictably inaccurate, or worse, unpredictably inaccurate.
So, given the mint condition of my shaper’s tenoning table, and previous experience with my jointer, I generally presumed that the settings were as they should be; i.e., on the money. However, the tenons produced were rudely photo-bombing, as it were, that happy vision.
Something was clearly out of whack, and a look at the situation suggested three possible causes:
- the 90˚ position of the tenoning table was off, relative to the main table
- the guide/support rail of the tenoning table was not running parallel to the main table
- the tenoning table’s fence was not parallel to the tenoning table.
I checked (1) by looking at the underside of the tenoning table. There is a hinged stop piece which must be swung out to get the tenoning table to pivot. The table can be angled 45˚ to the front or rear for angled tenon work. That stop looked to have never been moved, and the stops to the side of it which locked in 90˚ were snug and look to have never been disturbed.
So, while (1) remained a possibility, given the likelihood that the table had never been swung out of its 90˚ setting, and that those settings were set by the factory from new, I thought it probable that the table was set as it should be.
On then to check number (2), whether the sliding table ran parallel to the side of the main table. I brought the table forward and slid it over until it pinched a large feeler gauge against the edge of the main table:
0.0250″, for those metric troublemakers out there, is 0.635mm.
Then I ran the sliding table to the back and upon recheck discovered that the 0.025″ feeler gauge was now swimming in the gap:
I stuffed more feeler gauges in until I had found the width of the gap:
A veritable hand full of cards was required:
That lot totaled up to 0.0765″, or 1.94mm:
Well, something is out there, clearly. The sliding table support rail however weighs about 300lbs, and it attaches to a pair of massive steel bars on the side of the machine chassis, and would have been factory set to run parallel, so I’m really surprised to find this issue.
I talked with Joe Calhoun out in Colorado and he was equally surprised to learn how large the discrepancy was. I then obtained a Martin document from Ed Papa on Long Island showing how the factory puts the sliding table onto the shaper, and sure enough, they set it up to run parallel to the main table edge. I have to presume mine was originally parallel too, but it certainly wasn’t any longer.
The assembly is so freaking massive and heavy and has not obvious signs of damage that I am mystified as to how it could be out like that. Be that as it may, the good news is that everything is built properly, which means I can adjust the thing to get it running parallel. That, in a nutshell, is the mark of well made equipment: it can be adjusted and repaired, by design. It means that whoever built it anticipated it would last a while.
Actually, I am sure that I could deal with the whole problem by adjusting the sliding table’s fence, as I think it is a likely culprit here. These Martin sliding tables come with a fence made by Festo, which has a series of stops built in it for Euro window making. Ed Papa said that he found a lot of the Festo aluminum fences were bowed right out of the box. The fence on mine looks decently straight, and it attaches to the sliding table at three points. One of those points, at the front, is fixed,, while the other two are adjustable by way of eccentric cylinders. I stripped the two adjustable mountings apart and cleaned and lubed them, as they were a bit sticky – probably from lack of use.
At the end of the day, I need to adjust the sliding table, and associated parts and it is all adjustable. The fact that it is adjustable means I am not pissed off at the machine itself; rather I am chagrined personally that I didn’t think to check for a square cut when calibrating the tenon cuts. I assumed it would be cutting 90˚ as that is where it was set, and assumptions are a bitch sometimes. I will strip the entire sliding table down and reassemble it, checking with long straightedges and dial indicators as recommended by the factory manual during the rebuild. I’ll deal with this issue later on however, as I just don’t have the time for that fun at present.
For now, I am left with having to make a bunch of fine adjustments to the tenoned ends to bring the shoulders back to 90˚. This effectively added a day of work to my schedule, which was certainly not welcomed. The slight adjustments of the tenon shoulders will mean the doors will narrow by a very slight amount, on the order of 1/32″ (0.8mm), but this is not a significant issue on a nearly 4′ wide door. The lost production time is the issue.
Here I’m trimming the ends on two of the main door battens:
Same process on a different pair of battens:
By the end of today I had all the main door rails and batten tenons adjusted and further processed towards final form:
A view from the end shows the top/bottom rails on the bottom of the pile, with battens above:
The left side pile is for the left door, and the right side pile for the right door.
The top/bottom rails in a closer view, showing the fully-shouldered twin tenons:
There are some mitered returns to trim yet, however I am putting this pile aside for the moment so I can bring the side door rails and battens along to the same stage. That’s tomorrow’s little adventure. Stil feeling more or less on schedule for completing the door fabrication by the end of the month. Might be a day late, might not.
All for now – thanks for visiting the Carpentry Way. Post 59 lay ahead.