Hull & deck

Fixed windows

Two long fixed windows: four tries, adhesion tests and, at last, a dry cabin.

Vira has two long fixed windows on the cabin sides. They took me two years and four tries, and taught me more about sealants than I ever wanted to know.

The whole job, as I wrote it while it was happening, is in my original diary article: Fixed ports.

Fixed windows, attempt one

Day 12–239 Feb – 23 Apr 2022

The fixed windows had been set straight into fiberglass resin, probably gelcoat, with zero flex. So the original acrylic cracked from expanding and shrinking in the sun. When I got the boat, they were cracked and leaking, both through the cracks and around the edges.

In February 2022 I checked my plywood template. Very close.

The dry fit of the new acrylic was off by about 5 mm on the long corner.

April was a full disaster. Taking the old acrylic out, I chipped the gelcoat all around. There was no flexible sealant at all, just acrylic set in gelcoat.

Then I glued the new ones in.

One spot didn't get enough Sikaflex, so I redid it. Lesson: use more than enough, then clean up the extra.

Attempts two, three, four...

Day 49–10224 Sep 2022 – 6 Aug 2023

Who designed it this way, I don't know. The frame gives only about 10 mm to bond the lens, and the surface is curved, so it's always trying to pull the ends of the lens out.

The first try was 1/4-inch acrylic cut to size, on the famous Sika 295 UV with activator. The sealant cost more than the acrylic. I followed every instruction.

In September 2022 I saw that the starboard lens wasn't sticking. Probably I had forgotten to prime the acrylic. The other side was fine. Half a year later one window dripped: the Sika had let go of the plastic. I thought I'd messed up the prep or the timing, so I cleaned everything, scuffed the acrylic and did both sides again with Sika.

That lasted a year, and then it dripped again, the same way. My theory was that UV going through the acrylic was killing the bond. So in May 2023 I tried the method everybody recommends: VHB tape and 3M 4000 UV. Cleaning off the old polyurethane was a lot of work. It bonds to fiberglass really well, which is exactly what you don't want at that moment.

In August I filled the gap between the windows and the inner plywood. I'd put it off for a while: it's messy work that has to be done carefully, so you don't damage the plywood face. I used epoxy with a bit of chopped glass for strength, 410 microlight so it sands easier, and 404 plus silica to get the right thickness. Filling a gap overhead is tricky. The next step was to sand it without scratching the acrylic.

The VHB didn't survive half a year. I didn't check why, but I'm sure the small bonding area played a part.

Screws, and a test first

Day 11820 Nov – 27 Dec 2023

I was tired of fighting adhesives. The new plan came with a triple guarantee:

  • Screw the lens down, so the sealant never holds the load. A screw every 70 mm, 36 per window. Lots of holes, I know, but no stress on the bond.
  • Use a different sealant on the lip: DOW 795. It holds glass on skyscrapers, so it should be good enough for us. DOW says it works with acrylic and Lexan, and lots of people online recommend it.
  • Butyl tape around the outside.

Each window is 1,025 mm long, and my CNC only does 740. So I cut each one in two positions, 400 mm apart, left side first, then right side. At each position: drill the holes, mill the slot around, then cut it out of the sheet.

This time I wanted one shot that lasts 5–10 years, so I ran an adhesion test first. I cut 15 pieces of FR4 circuit board and sanded them to bare laminate at the contact spot, about one square inch, all with 220 grit.

For butyl tape I tested acrylic sanded at 220 and 600, Bed-It against a white butyl tape from Amazon, cleaned with 50% IPA. All of them came off at about 6 lb. At 600 grit, neither tape stuck to the acrylic at all; nothing was left on it. At 220 the Amazon tape held a little, not great, and Bed-It nothing.

For DOW 795 I tested glossy acrylic, 220 and 600 grit, cleaned with IPA or TotalBoat dewaxer, bare or epoxy-coated. Glossy acrylic held some, about 10 lb, and the silicone peeled off clean. Every other sample, whatever the prep, held until the silicone rubber itself tore in the middle, at 12–15 lb. We have a winner.

So it was DOW 795 only, around the edge and in the gap. In December 2023 the new windows went in, screwed every 75 mm, with the acrylic and gelcoat sanded at 220 and cleaned with 50/50 IPA.

Finally dry

Day 204–20520 Nov 2023 – 29 Jun 2025

July 2025: no leaks for a year and a half. Now I could finish the inside. First I made two templates, with 20 mm and 30 mm corners. The 20 mm one fit best.

The screws still worried me. They sit in fiberglass with silicone, and I don't trust that seal much, so I put rubber washers under them for a compression seal. Maybe I'm wrong, but washers are better than nothing. By the way, in later boats Irwin dropped the fixed windows and used regular ports. I understand them now.

Then came new teak trim, each frame four pieces cut on my CNC in the garage. Only the gap between the glass and the trim is left to paint. Two years, four attempts, and the windows are finally dry.