The third of three installments describing the gate’s installation at site. 1st installment.
I received some additional photos and video from Mike G. so I have a bit more to share about the conclusion of the install process than I would have had otherwise. We left off the previous post at the stage of having put up the main frame assembly on April 2nd:
Uh, ‘high drama’ with the crane over with, the crowd, along with the photographers, dispersed, leaving us to continue with the process. Matt was available until just after 1:30, as he had a meeting to attend, so Mike was my helper for the remainder of the afternoon, along with my ever-supportive wife.
With the notches for the main doors cut on the granite plinths, but the wind precluding the installation of the doors, I decided that the left side paneled section, or waki, should go on next. This section had been pre-built in my shop and was brought to site as a sub-assembly unit. All that remained was to fit it on. The connection of this sub-assembly to the main post was by way of a series of double dovetail sliding keys:
In the above pic I’m fitting the last of the wooden plugs which restrict the ability of the dovetail keys to slide out of their mortises.
There were some unknowns with this process of fitting the waki, however these things I had known about ahead of time. These unknowns concerned the end of the concrete fence and how the wall post would fit against it, and whether my granite component heights were on the mark or not. In the old gate, the wall post was about an inch away from the concrete, and there had been a wooden filler strip jammed in the space. In the new gate, I had made the wall post a larger section and had placed it closer to the concrete. There were some blobs of mortar and roof tile overlaps at the top of the wall however, and these constricted the room available for the post. I had made the wall post’s threaded rod hole elongated so as to allow the post to slide as close to the wall as possible, my plan therefore being to hopefully squeak the flanking post by in front of all those dovetail keys on the main post.
A reader, in a private mail to me, wondered why I was using hammerhead keys in some locations, and dovetail keys in others. Well, the reason I went with dovetail keys in the above context is that they can sit in a shallower mortise and do not project as far out from the post surface as would a hammerhead key. I needed all the room I could get as I anticipated a tight squeeze to get things together. The dovetail keys are weaker locking devices than the hammerhead type, however in this case the issue is less important as the flanking post is vertical and bears upon the granite plinth and sill below. Also, to compensate for the weakness, I had made this connection with 5 pairs of dovetail keys, which was plenty since their task was mostly to hold the flanking post tightly to the main post.
Anyway, we lifted the waki assembly up and put the wall post down atop its threaded rod hold-down, and then swung the flanking post over to the row of dovetail keys on the main post. It didn’t quite clear. However, it only took a little fiddling to rectify the situation. Essentially I loosened the hold-down on the main posts, used a sledgehammer to move the main posts a bit laterally, and then with a little persuasion, the flanking post could be coaxed by the dovetails keys and into place. Then it was lifted up a couple of inches, the dovetail keys engaged in their mortises and ‘clunk’ it started to go together. Then a little further sledge persuasion was needed to drive the connection home:
I was pretty psyched that it was coming together well.
A view of the front of the waki after installation:
(Photo: Mike G.)
I had good fits at the meeting of the flanking post, middle of sill and wall post, though all were meeting different granite elevations.
Once the waki was in place, the main posts could be shifted back over again and cinched down:
That left a space between the wall post and the concrete of around 1/2″.
The other side of the gate had similar framing elements but a side door instead of a fixed paneled section. As before, the dovetail keys and plugs were installed on the main post:
This time the main posts could not be scooted sideways at all, so,with limited space to push the wall post back against the concrete wall, I trimmed back the stub tenons slightly on the header to get it to fit to the wall post and yatoi sen:
The flanking post needed a little adjustment on the bottom to obtain a good fit as the granite was not perfectly flat:
After that was down with a good fit the wall post’s threaded rod tie down was tightened up.
Then it was time to fit the side door, which had previously been fitted to the framing in the shop. It went on with modest coaxing:
The door was then taken off, the hinge pins removed from the frame and some PL300 put into the mortises. Then back together and the hinges aligned more closely:
Then the process of tweaking the fit began:
Just about there:
A fit was achieved and the door lock engaged:
This is pretty much how things ended on April 2nd, framing-wise:
Then we brought the main doors inside the gate and wrapped them with moving blankets and plastic sheeting:
I was pretty beat by the end of the day. I think the cumulative effect of the time crunch on the bubinga tables project from last year, which ran a month later than expected, then the nearly daily work schedule of fabricating the gate over the past 4 months, and the stress of the deadline and the install scene, was starting to affect me. I was so tired that evening that I was okay with going to a ‘fake’ Japanese restaurant (most ‘Japanese’ restaurants in North America are run by Chinese or Korean people, and the food, in my opinion, is almost never as good as a real Japanese restaurant). Sure enough, the “sushi” at Haru on Huntington Avenue was truly bad, among the worst I have had, and I was too tired to really care. We only picked that place as it was the shortest walk from the hotel. I was so fatigued I could barely operate chopsticks – not exaggerating. Got home at 9:00 pm, and promptly fell asleep, only to wake up at 2:00 am. and could not return to sleep. Only in the last couple of days has my sleep pattern begun to return to any semblance of normality….
While April 2nd had been quite windy, April 3rd featured showers. This put a slight damper on proceedings, and added a bit to my worry as to whether we could complete all the work on schedule, however we were fortunate in not having any significant or extended downpours.
The next task in framing was to put the kasagi beams on – these are the ones with the upswept ends and beveled upper surface that meet the main post and cap all the flanking and wall posts. Before these could be fitted however, there were some covering boards which needed to be fitted in behind the wall posts to conceal the ends of the tiled roof concrete wall. I had the dimensions of the wall on my CAD drawing, however there were unknowns and irregularities in the tile work which I knew ahead of time would have to be dealt with in the field. Though I had the original cover boards from the old gate, these seemed way out in their sizing and shape so I couldn’t trust them as a template. I had glued up some blanks for these cover boards out of some VG 8/4 stock, however I discovered on April 2nd that the tiles extended a bit higher than my glue up. So, first order of business was to slice up a bit more material and glue another board onto the cover pieces to widen them adequately.
After making a cardboard template of the actual wall end and tracing a suitable pattern onto the glued-up panels, I was given access to the wood shop in the MFA where I was able to trim the parts to size. Then I had to cut a large dado on the panel so it would fit around the wall posts. For this task, I had brought along my groover:
(Photo: Mike G.)
It took about half the day to get those squared away, but in the end I thought they came out decently enough:
These cover boards still need a little clean up, and the edges will be painted white. That can wait. They are fastened to the concrete wall with Tap-Con screws, which greatly stiffens up the entire wall post and its associated connections.
While I had been dealing with this matter, Mike and Matt had been fitting the kiosk shoes with stainless rods and domed caps, and securing fasteners. There were some complications with that process, however they were worked it out in the end. The kiosk is ready for install, however we are waiting on the welder and this work is now slated for the week of April 13th.
With the wall cover boards in, the kasagi could at last be fitted. With a little careful trimming of some of the mating surfaces, these went on easily enough, and were a snug fit joinery-wise:
(Photo: Mike G.)
Above, we have the double hammerhead tenons atop the flanking post, along with the yatoi sen, engaging the end of the kasagi with its three stub tenons.
(Photo: Mike G.)
Simultaneously, the other end of the kasagi was engaging with two sliding hammerhead tenons atop the wall post. A sledge was required to drive the parts together, and they went together well.
Next was the hanging of the main doors. Fortunately the wind had taken a break, along with the rain. A friend of Matt’s had shown up, Corky, who also works for Byggmeister, and he also pitched in to help. With four of us, it wasn’t so bad wrangling the 200lb. gate doors around.
Here we have one door hung and the door hook also installed on the back:
(Photo: Mike G.)
The old gate had the rear posts too close to the front, and when the doors were swung open they banged into the rear posts. Something they appear to have overlooked in design. In my design, I pushed the rear posts back a bit, had the foundation granite paving extended as well, and now the doors tuck inside of the rear posts, which widens the view of the garden.
The second door was soon on, and I see that the hanging stiles are a bit tight to one another:
(Photo: Mike G.)
More interference than I was expecting, but all very fixable.
Off came the door and some planing ensued:
(Video: Mike G.)
It took a few rounds of ‘on and off’ with the doors, and a little planing of one of a hinge-side stiles, some minor adjustment at the granite plinth, but it came together eventually.
Last task was to fit the door drawbar, kannuki, which required mounting 4 bronze carriers:
Below the middle carriers you can see the two receivers for the door hooks.
Almost there – checking the sliding action of the drawbar:
The remaining time was spent cleaning up and putting the tools, etc., back in the box truck.
At the end of day 3, the install was 99% complete. Here are some photos – as usual, click to enlarge:
The temporary chain-link fencing at the front makes a full-width view a bit difficult at this time.
Here’s one front side:
And the other:
Rear (garden side) views:
Bidding goodbye to my two helpers, I drove back to Western MA, 2 hours or so, went to the shop where my wife met me and helped me unload the truck and then crawled off home, thoroughly spent. I’m still recovering today, three days later, and will be going to be taking some time off for a bit more rest.
It came out well I thought. Really there were no significant issues to deal with, which was a bit surprising. The MFA was delighted with the gate, my helpers seemed to have had a good time. No one got hurt.
The last few tasks involve mounting the kiosk and fitting the remaining bit of sheet copper to the gate, along with the odd tweak and adjustment here and there. I’m scheduling this work for the week of April 13th. The grand opening party will be on April 24th. It is an invitation-only sort of thing, however after that the garden will be open to the public during business hours, and free to visit. If you’re in the area, I hope you’ll take a look and let me know your impressions.
I’ll take a break from blogging for at least a week. If you’re new here, there are plenty of old posts to check out, and I hope you’ll do so. Until next time then, be safe and enjoy yourself. Thanks for visiting the Carpentry Way. Post 85 to follow.