Now that the sealant had had adequate cure time, I removed the excess from the exterior of the window, cutting carefully along the edge then removing the sealant and tape.  This posed no issues.  Then, I pared away the excess teak bungs and sanded the trim as needed, and sanded the small section of remaining cabin trunk surrounding the window.  I vacuumed and solvent-washed to clean up the final debris and sanding dust.

Inside the boat, I cleaned off a few spots of cured sealant on the inside of the glass and checked once more the overall appearance of the seal between glass and exterior trim, now fully visible and with the excess sealant and tape removed from the outside.

I installed the interior trim, using the screws that had been holding it at the start of the project, then masked the inside face of the trim, closest to the glass, and also masked the glass itself, following the line if the trim were to extend all the way to the glass.  The tape would make it easier to tool and clean up the sealant more easily and neatly.

I applied a heavy bead of sealant into the groove between the glass and trim, then pressed and tooled it with a finger to finish off the seam.

Finally, I removed the tape to leave a clean line of sealant, and this wrapped up the installation.  I reinstalled the trim I’d removed from the edge between cabin trunk and sidedeck as well.

Back on deck, I lightly sanded the first coat of varnish I’d applied to most areas last time, preparing for a second maintenance coat.  I vacuumed and solvent-washed thoroughly thereafter.  Apparently I forgot to take any pictures from the starboard side, but you get the idea.

Finally, I applied a coat of varnish to the starboard cabin trunk side, to catch this area up with the rest of the boat, since I’d left this unvarnished till the window trim was complete.  Before doing the varnish on the whole side, I’d dabbed some varnish onto the raw teak bungs to get some finish started in those areas first, and ultimately the new window trim would require a few more coats of varnish to build up enough varnish on the new bungs.  This was a good opportunity to also take care of the second coat on the swashboards and tiller.

Fortunately, my early morning check of the window after installation had better results than last time:  all was sound and intact.  I still don’t know what happened to the glass the first time around, but in any event now the window was installed and looking good.

Because it’d been substantially fewer than 24 hours since installation, I chose not to do much more with the window at this point, not wanting to upset the sealant before it had a chance to fully cure, but inside I removed the temporary trim and masking tape to give me a good look at the seal from inside the glass, where the pressed-out sealant was clearly visible around all the edges.

One thing I could do was install bungs in all the screw holes in the exterior trim.  I glued these in place and left them to cure for the day.  The new trim afforded adequately deep bung holes to accept the bungs and glue.

Next, I turned to the brightwork, and after final preparations applied a coat of varnish to the coachroof handrails, hatch trim, forward face of the cabin trunk, and cockpit coamings.  It was handy in this case that the wooden cabin trunk had various places to start and stop varnish, such as the vertical corners and trim, so I could do the work in a more convenient and piecemeal fashion.

With the deck-based trim under varnish, I continued with the loose pieces in the woodshop–the companionway boards and tiller.

Then, I varnished the port cabin trunk and toerail.

I finished up with a coat on the taffrail and starboard toerail; the starboard cabin trunk would wait till the new deadlight trim was ready for varnish and I’d prepared the area.

After a few days away from the shop, I was looking forward to finishing up the deadlight installation (again).  My time away had given the glass shop time to “almost” finish the new pane of glass–when I called first thing in the morning I was told I could probably pick it up in the afternoon.  This gave me time for some unrelated errands, and on my way home I received notice that the glass was done, so I picked up the new pane.

Back at the shop, I started by dry-installing the new window trim, which now had 4 coats of varnish on it.  With the trim clamped in place in the proper position, I installed two temporary screws in each piece to hold it, drilling and tapping the cabin sides, with their epoxy-filled holes, for the #8 machine screws.  I used temporary stainless steel screws with a Philips pan head because they were easier to manipulate than the “real” slotted bronze screws that I’d use for the final installation.

Inside the boat, I double-checked for proper overhang of the trim to provide a resting place for the glass window.

Next, I test-fit the glass to ensure it fit correctly, then, with tape applied to the exterior edges of the pane, taped it thoroughly in place so I could mark and trim the tape where the trim overlapped it.

Next, I removed the glass and trimmed the tape where I’d scored it, then used the pilot drill to drill the cabin side for all the remaining screw holes before removing the trim; afterwards I used the 8-32 tap to finish off each hole for the screw threads.  Then, I made final preparations for the trim and glass installation, including covering the deck with plastic and masking the inside edge of the trim pieces where they’d meet the glass, since I planned to leave the sealant squeezeout to cure before removing.

Applying a heavy bead of sealant one trim piece at a time, I installed the trim from outside with bronze screws, then cleaned up the sealant around the outer edges.

From inside, I added more sealant to the corner between the trim and the cabin sides–there was already a good bead there from the trim itself–and carefully pressed the glass into place, watching the sealant press out evenly from inside the glass to ensure a good seal.  Then, I taped the glass in place securely to hold it, and added some of the interior trim as a fail-safe to prevent the glass from falling into the cabin should the worst happen.  To help secure the glass further, I rolled up some bundles of tape and used it as wedges between the trim and the glass, and even a couple small and carefully-installed wooden wedges in one area that needed a touch more pressure.

Last time, I’d secured the latch block for the forward hatch to the underside of the lid with epoxy:  the screws that had held it before were pulling out of the wood and wouldn’t hold it securely.  I thought the existing block, with the new and larger gasket, would allow full closure and gasket compression as is, but it needed a stronger mount.  Now, with the epoxy well-cured, I tested the latch and was pleased with the way it compressed the gasket all around.

I continued the varnish work on the new window trim with a quick sanding and third coat.

During the remainder of the day, I finished up the masking for the brightwork maintenance on deck, this time focused on the interior of the poop deck, cockpit, and coachroof, along with above the port eyebrow trim.

I lightly sanded the new epoxy filling the old deadlight screw holes to flatten it and remove any hard spots.

While I had the trim down, and since I’d already drilled small pilot holes at each screw location, I took a moment to drill 3/8″ diameter recesses to accept bungs later.  I did this on the drill press to ensure accuracy and consistent depth.

Now I test-fit the trim again.  I would have liked to drill and tap for some of the screws to help me hold it, but the fresh epoxy would hold threads better if I gave it come cure time, so for now I relied on clamps.  I arranged the trim so it fit in all the corners, and clamped it in place.  But checking the reveal inside, I noticed that the lower forward corner was too low, the result of the bottom trim piece being slightly too long.  This wasn’t a surprise, since I’d suspected all along that the broken-and-repaired old trim that I’d used as a pattern had “grown” a bit in the process, making it slightly longer than original and therefore skewing the forward joint.

I needed to adjust the joint between the forward end of this lower piece and the forward vertical piece, so from outside I adjusted the lower trim till it was at the correct height, then overlapped the two pieces to make some reference marks for a cut.  I started by taking a small amount off each piece of trim, as both required minor shortening, then, on a last test fit, marked the upper trim with a final cut line where the bottom overlapped it.  I cut the upper piece because I didn’t want to take more off the lower piece, which might bring the fastener location too close to the edge; there was more room for a cut on the other piece.  The line in question in the photos below is marked with two V-shaped pencil marks, or what looks like a W.

After this final cut, the pieces all fit correctly in the corners and with proper reveal all around.  To help mark and later recreate the alignment, I used masking tape along the straight outer edges of all the trim, which would make future placement straightforward.

With all the trims fit correctly, and knowing that the outer curves aligned when in place, I removed the trim pieces to the bench where I could clean up the insides of the corner joints as needed and round over the exposed edges, then finally sand the trim clean and smooth.  Then, I applied a sealer coat of varnish; since sealer coats largely soak into the wood and dry quickly thanks to the large amount of solvent in the mix, I hoped to do a second coat later in the day.

With the work on the deadlight done for now, I turned back to the rest of the brightwork on deck, mainly the interior cockpit coamings, handrails, and some other trim that was not reachable from the staging, and finished up the sanding to prepare for brightwork maintenance.  Then, I vacuumed the deck and trim, and solvent-washed everything to prepare for masking.

I spent the remainder of the day masking off everything I could reach from the staging:  toerail and rubrail; taffrail; outer cockpit coamings; windows and ports; and the edge of the cabin trunk above the eyebrow.

I got to work on building new trim pieces for the starboard deadlight.  I found a teak board of appropriate size to work from, and laid out the original trims to ensure there was plenty of room to duplicate.  Then, I resawed the board into a blank about 3/8″ thick before planing it down to a final thickness of 5/16″, which was a bit heavier than the existing trim to better allow for bungs over the screw heads, but not so thick as to stand out from the older trim on the boat.

I traced the old trim pieces on the new board, and cut out the replacement pieces, leaving the line so I could ease them down to final contours as needed.  Then, with each corner aligned and clamped over the end of the bench for access, I shaped the four corners so the pieces would mate correctly, then lightly sanded down to the cutlines along the inside edges.

Using the old trim as a guide, I marked all the screw locations and drilled 1/8″ pilot holes through the new stock, which then allowed me to set up the trim in place on the cabin side with small nails through the holes for alignment, and some clamps holding the trim securely.

While the test-fit was generally a success, it soon became challenging when I attempted to match up the screw holes using the #8 machine screws that would be needed in the final installation.  I started with one piece of trim (the bottom trim) and enlarged the pilot holes slightly with a bit large enough to allow the machine screws to go through, and tried to get all the fasteners installed as a test.  But several of the holes weren’t lining up conveniently and I soon fell back on an idea I’d had earlier, and which I should have just done from the getgo, which was to slightly drill out and fill all the existing holes with a structural epoxy mixture.

I used a 1/4″ bit to enlarge the #8 holes–I would have liked something larger, but the holes were close enough to the edge of the opening that a larger bit would have caused problems there.  I filled the holes using a syringe, and covered each hole with some tape to hold the epoxy in since it was a looser mix than I might have used because of the requirement to get it into a syringe.  Once cured, the filled holes would allow me to simply drill and tap new holes for the fasteners, without the headache of trying to match existing holes to the same extent.

Checking the boat first thing in the morning to see how the sealant was curing and whether I could remove the braces, I was mightily disturbed to discover a pair of cracks in the center of the glass.  These cracks had appeared on their own sometime overnight, and I was baffled by it.  Since the lens and trim installation had gone so well overall, and the end result yesterday had so obviously not included any glass cracking, this was not really the start of the day I’d been hoping for.

All I could think was that there must have been some pre-existing flaw in the edge of the glass that, at some point after installation, had caused the cracks to grow.  This was laminated glass, and anyone with a windshield in their car has probably seen a tiny chip grow into an inspection-failing crack in mere seconds some winter’s morning when the defroster was on, or any other circumstance.  Perhaps this is what happened here; I truly did not know.

And frankly, it didn’t really matter right now.  I allowed myself a few moments of self-pity.

What mattered was that I’d have to start over and remove everything I’d just installed so I could get the glass replaced.  I had the small benefit of the knowledge that the sealant had had little time to cure, and therefore removal right now would be about as easy as it could possibly be, so I got right to it.  To begin, I removed the bracing from inside and installed the two short pieces of interior trim, at the vertical ends, to give something for the window to lean against if it wanted to move during trim removal outside.  I needed the glass still intact so it could be a pattern for the new lens.

Here’s a few more photos of the cracks for posterity, and with a clearer view with the bracing removed.

Next, I removed all the screws from the exterior trim, then removed the trim carefully.  While the sealant might not have been fully cured, it was still plenty stuck, and it was a challenge to remove the poor old trim pieces without damage.  The bottom piece of trim, which had had some damage previously (and been repaired after the last removal) did not survive this second removal in as many weeks, unfortunately, but I was hardly surprised given its condition anyway.  Now I planned to make new trim pieces for the replacement installation.

With the trim off and the glass out, I scraped away the bulk of the remaining sealant, and sanded and solvent-washed the area clean again.  Any final remnants of “old” sealant could be removed a little later, but I got it 95% of the way there.  Then, I left to take the old glass to a glass shop where I could get it replaced.

Preparing to install and rebed the deadlight, I started by masking down a sheet of plastic to cover the deck and environs for protection during the work. Then, I installed the interior trim to give the deadlight frame the bearing surface needed for installation.

Working from outside, I masked the inside edges of the glass lens, then set it in the opening to check the fit and trim the tape along the edge of the trim on the inside.

To help hold the glass in place during bedding and installation, I rigged up some blocking to push the glass tightly into the frame from outside.

But throughout this, it didn’t look right.  Because of the way I’d ultimately disassembled and removed the deadlight, I’d assumed that reinstallation and bedding would occur in roughly reverse order, and from the outside of the boat, since the interior trim featured a lip that extended into the opening behind the glass.  But on the outside, this left a wide 1/4″ or greater space between the lens face and the cabin side, and since the exterior trim was flat, this meant that the trim would never bear against the lens and make sealing difficult.

There was just no way the installation was going to work this way, and after looking over the remaining three deadlights, it seemed clear that what I had to do was basically the opposite of what I’d started:  I’d need to install the exterior trim first, then bed and install the window from the inside, creating a tight seal where it mattered at the outer edges of the deadlight.  The excess thickness and space could be better managed from the inside, and better matched the existing installations.

So now I dismantled my setup, and removed the trim from the inside of the opening.  Working from outside, I dry-fit the exterior trim, using at least two screws in each piece driven in all the way to hold it for now, but also partially inserting every other screw to ensure that they all fit correctly and would thread in as needed.  I had to clean out a couple of the screw holes with a tap.

I masked over the edges of the outside face of the lens, then, from inside, placed it in the opening, holding it with some tape and then some braces extending across the cabin to apply light pressure.  Then, from outside, I ran a knife around the edge of the trim to score the tape so I could remove the section beneath the trim later.

Satisfied with the clamping setup and the overall arrangement, I removed the braces and glass, trimmed the masking tape, and then, from outside, removed the trim to prepare for bedding and final installation.  Then, one piece at a time, I applied heavy beads of sealant and fastened the trim down securely all the way around, ensuring good sealant squeeze-out in all areas.  I cleaned up the excess sealant around the outside of the trim.  I used Sikaflex 295UV sealant.

Inside the boat, I applied more heavy beads of sealant to the inside face of the trim and around the opening, then pressed the glass into place, applying hand pressure till I felt that it had evenly pressed into the sealant all around.  I installed some masking tape to hold it, then the braces running across the cabin.

Outside, the squeezeout was excellent all the way around the glass.

To finish up for now, I cleaned up the excess sealant from the glass and pulled the exterior masking tape.  I’d leave the interior bracing in place for as long as needed to ensure the sealant cured sufficiently.

In other works, I reinstalled the forward hatch to check the fit of the new seal and to determine if the clamping block needed adjustment.  The seal seemed to work well, but when I attempted to engage the hold-down clamp, the screws pulled out of the top of the hatch, apparently unwilling to properly dig into the holes that had worked once, during initial installation, but would not grip the screws again–but not before I determined that the hatch had seemed to pull down correctly with the pressure, as the replacement gasket was a bit taller than the old one.  So I planned to glue the block in place to hold it, and make no modifications to the height.  This would happen next time.

The owner was interested in exploring the possibility of a manual windlass, and I spent a little time making a rough, but accurate-enough, paper template of one likely candidate, the Lofrans Royal, to show the overall footprint.  Space on the foredeck was limited, and part of the exercise was to determine how well such an installation would fit.  Given the requirement for the chain lead straight out of the chain gypsy on the starboard side of the “windlass”–the center of that rectangular protrusion there–and the need for a roller at the stem in line with that, placement options were limited to centerline, somewhere aft of the mooring cleat roughly as shown.   Most rollers are about 3″ wide, depending on anchor type (some are narrower), and such a roller would fit between the center of the stem and the bow chock, but not with much clearance on either side, and such an installation would likely impact the utility of the bow chock on whichever side the roller was mounted.  At this stage, I’d not yet determined specifically what sort of roller would be needed here.

Inside the boat, the chain locker, as it were, extended aft to a spot approximately in line with the aft pulpit bases, so in the configuration shown here the windlass chain pipe would be aft of the chain locker.

In short:  a windlass installation would be possible, but with numerous complications and likely need to reconfigure certain things while making some tighter access to existing installations on the foredeck, and provide a means for the rode to enter the chainlocker cleanly, assuming a chain rode leading off the gypsy.  In addition, the gypsy on this particular windlass was chain only and wouldn’t work with rope rode, though the winch drum on the opposite side would.  But that seemed like it would be cumbersome to say the least.  Some windlasses have gypsies that work with combination chain/rope, though the chain needs to be spliced to the rope in a specific way for a clean transition.