Replace Leaking Boat Windows with Opening Ports

March 2004 -- I was just going to redo the existing window design, but came up with an idea that was just a little more work and would be way cooler. (You'd think I'd have learned about my big ideas by now.)
The idea here is to never have to deal with leaking windows again.
I also want opening ports, for cross ventilation.
And, I want to put it all together at home in the evenings, then install over a couple of weekends.
So -- no major fiberglass work or other things that would take the boat out of sailing shape.

So, a little history: The windows leaked.
I've seen worse leaks, but any leak is a pain.
I've made it though the last few winters wit

Note that in this picture, I'm missing a dog. Typical move on my part.

I lost my dog! (grin)

I was afraid it had fallen into the trash or something, but it turned up behind a cushion. Whew!

I'm looking forward to the winter storms.
I hope we have a really wet, stinky, sloshy winter!
I'm ready to hang out watching movies and eating popcorn while nature rages,
and not be changing buckets.

h strategically placed buckets.
It's time to deal with it.

I took a good look at the ports to check the plan
before starting to build, being pathologically conservative
to give myself room for mistakes.
The key limiting factor, and the one thing that can make the project fail, is the maximum thickness of the cabin house.

With the Tri-Matrix ports, hinges and dogs are attached to the back plate, which fits over the front assembly.
The hinges and dogs are adjusted on the back plate to compensate for varying thickness of the cabin.

To find the Maximum allowable thickness,
I adjusted the hinges and dogs *beyond* the extreme limits.

The previous owner redesigned the windows, and it was a good design. The implementation went wrong, but the design was right.

Steel frames are screwed into the house, and overlap the windows. If done right, leaving room for differing rates of expansion,
it should have worked.

Perhaps when I take it apart I'll determine exactly why it broke.

However, I have one of my *big ideas*, and am going to redesign the whole thing from scratch.

April 2004 --

The first step was to trace the outlines
of the old windows.
I took a long piece of paper from Office Depot and taped it in place, then rubbed along the edges of the frames with the side of a pencil.

The inside tracing is important, since that *roughly* matches the interior shape of
the window.

The holes are punched with a Wad Puncher.
I'm sure that cutting holes in paper
dulled the blades a bit, but what the heck.
I wanted precision holes so that the pattern was exactly right, since I'm outsourcing the fabrication.
Normally one uses this punch to put holes in a wad of canvas or leather, and bang it with a 2# sledge.
(The center pin is spring loaded and
gets out of the way when the punch is pressed down.)
Whatever. It's the only hole punch I have.

The opening ports are the new Tri - Matrix ports from Newfound Metals. They're made of UV stabilized plastic, and have glass windows and heavy duty hardware. They're much lighter than cast metal,
but still very strong.

The new ports arrived via UPS while I was drawing the pattern, so I verified the template and frame pattern.

The final pattern has holes a bit larger than the original pattern, as fudge factor, to leave room for caulk, and to leave room for differing rates of thermal expansion.

I spent a great deal of time pondering this while planning, and decided that if necessary I'd grind it flush with the new panel. However, I've been waiting until this point in the implementation to decide if that's necessary. Naturally, one should avoid grinding.
I had a gut feeling that with this much planning and effort,
it would somehow work out okay.

This gives me some theoretical limitations
to design to.

To the left is the modified plan.
This would be the top edge of the port side.
If this doesn't make sense right now, just stay with me.

This scenario gives me 1 - 1/8" thickness.
Note the 1/8" overhang at the top of the interior port.
I'm going to come back to that later!

I really think it will work. Some surgery with a Handy Dremel Tool or small grinding wheel might be needed along the top edge of the liner, to provide a
good flat surface on the interior.

The dogs have a lock-nut to keep the posts from moving around. That nut takes up space on the post, although it can be removed if necessary.

To the left, they're tightened to maximum adjustment. (Much more than normal.)

It would be foolish to plan the installation based on the maximum adjustment, since that would prevent any post-installation adjustments.

The port gaskets are 1/4" thick solid EPDM (Ethylene Propylene Rubber.)
I can compress it 1/8" when pressing
HARD with my finger.
That's *point* pressure,
and it would be very difficult
to get that level of compression
by pressing down with
the entire window frame.

However, if I plan on 1/8" maximum,
it's a safe bet that the actual installation
will be much less.

June 2004 -- The tracings lived on the living room wall for about a month while I thought about it.

Then everything got nasty.
This was supposed to be the fun part!
When tracing out the pattern for the new frames,
I discovered that all four of the windows are different.
The top edges of the old frames don't line up,
and all the windows are slightly skewed.
Clearly, it wasn't bad enough for me to notice,
so I'm sure no one else ever did.

After a few trips back and forth to the boat, staring and thinking and measuring, I adjusted the edges to make them the same on both sides, and added 1/4" all around to cover up any damage I might do when removing the old windows.-

The outer frame has a soft closed cell silicon foam gasket that's
3/16" thick and is intended to be highly compressed.

I'm planning on compressing it halfway, and if it goes more then
there will be a little more room elsewhere.

By setting a tape measure halfway up that gasket, you can see that the outer frame (left) is 1.25" away.

I spent a lot of time pondering this
and building fudge factors into my
plan, but the bottom line is that
1 1/4" will work just fine.

So, while chewing on the construction details,
I took a good look at the seals, which NFM says
can allow ports to be installed without caulk.
This is a new concept, and I'm skeptical. In a big way.

There's an easily compressible closed cell silicone gasket
around the rim. It's soft, and should conform to any
irregularities on the outside surface.
I don't know the actual type of foam, and don't know
how it will hold up over the years, but
it seems like it can be compressed pretty flat.

The bolt holes are inside that gasket.

They're constructed with an O Ring around
the external carriage head fitting.
That assembly fits against the plastic port frame.

It's a cool design, and if I had thought of it
I'd be pretty proud of the effort.
Personally, I have some underwater video
and photo experience that makes me
leery around O Rings,
and I also also believe that
planning for problems will avoid problems.

I'm going to install backup leak prevention.
I hate leaks.
It's a personal problem, and I'm working on it.

The existing windows are going to be 'plugged' with 3/4" marine plywood.

I cut the pattern up and took it down to the boat to make sure the plywood plugs will be
cut to the right size.

Unfortunately, the job took almost two months, and when I got the actual frames back the holes didn't match the pattern, and everything was off.
This was a huge disappointment.
I guess the job was too simple, and it was an inadvertent insult.
It was important to *ME*.
I probably spent too much effort in the planning, and turned the execution into simple mechanics, which is insulting to an artist. Oh. Okay. Sorry.

I paid $510, and could have done a better job myself for less than $100, in less than half the time. There's no point making an issue of it, as having the job redone could cost me another two months. So I spent a weekend in the garage redrilling holes and grinding the metal into the correct shape.

I took the frames to a local sandblasting shop, but they had a New Shop Boss
who said that sandblasting 1/16" aluminum would destroy it and bend it all out of shape.

That was a surprise, since the same shop had blasted some 1/32" thick aluminum for me
only two years before, and it came out perfect.

However, one can't argue with an expert
without risking that he'll screw it up
just to prove that he's an expert.
Since it was new metal, I just sanded the frames seriously with 80 grit aluminum oxide.

Then I had them hardcoat anodized,
to seal the metal. It cost $75 for all four frames, which is less than the cost of phosphoric acid +chromic acid + polyurethane primer, and took just a single day.

August 2004 -- I took my frame patterns to a trusted local metal shop in Alameda,
who has done excellent work for me before and has a reputation for perfectionism.
I wanted them to be just exactly right and I figured if I paid an expert
it would be a better job than I could do myself.

There was no need to prime over the anodizing,
but I put an extremely thin layer of zinc chromate spray on it for two reasons:
1) To place another corrosion barrier over the aluminum
2) To put a soft primer between the anodizing and the cover paint.

The second point is key for me. If anything scratches into the finish, I want the *Paint* to give way and
chip off, and leave the anodized undercoat intact.
Zinc chromate is soft, so in a sense the entire visible paint coat is sacrificial.
The Paint is the primary protection.
The Anodizing is the Real Thing.

I'm trying to take steps to avoid having bare aluminum exposed to the salt environment, so three layers of protection should get the job done and give me time to fix chips before the metal is exposed
and starts to corrode.

Then they were rolled and tipped with the same Interlux single-part paint that I used on the other
blue trim on the boat.

Four coats, scuffing seriously between coats
with a green 3M pad.

I tried a light sanding with 220 grit
between coats 2 and 3, but was afraid
that I'd go too far and damage the anodizing,
so switched back to the 3M pad.

The last coat was applied on the dock, since my garage is dusty and I thought the dock would be cleaner.
But the dock was 90 degrees and buggy on that one day, and my garage would have been better.
Oh heck.

The "PO" bedded the old windows with 5200. That stuff is Evil.
Evil. 5200. Evil. 5200. Evil. Evil. 5200. Evil. 5200. Evil. Evil.
3M 5200 should really be a "special order" product.
It has no use in normal boat maintenance.

Using 3M 5200 doesn't mean that something will never leak,
it just means that you can never fix anything without destroying it!

I scraped as much residue away as was possible with a razor blade, and filled the old frame screw holes with epoxy.

Note the old construction. No wonder the windows leaked! The old windows were just barely caulked in, leaving a huge air space between the frames and windows. When the external frames leaked, this space filled with water, and then it was just a matter of time before it worked its way inside.

I'm bedding the frames in 4200.
This is still a serious Polyurethane Adhesive Caulk,
but it's not as destructively permanent as 5200.
Nonetheless, I doubt that these frames will ever be removed without destroying them, just because I'm using so much of caulk and the fiberglass is so uneven and rough. This project is designed and constructed
to last the remaining life of the boat.
(And even *then* I won't use 5200!)
If the frames need to be repainted,
it will have to be done in place, and at that point I'll bet the whole deck will need to be painted.

Note the 1/4" gap between the frames and the house,
where the original windows were glued in place.

When everything was as clean as I could make it,
it was time to mount the new frames.

For the first one, I stuck little bits of 1/16" neoprene foam on the back, to ensure that I wouldn't screw them down too far and squeeze all the new caulk out. It is very soft, and easily compresses to 1/32".
This turned out to be a Bad Idea, and I only did it on one frame.
The caulk is very goopy, and the frames have so much surface area that it's impossible to accidentally squeeze all the caulk out.

I'm very concerned about the thickness of the cabin side, and need to ensure that it doesn't deviate from the plan.

So I clamped down really hard on the top edge where the wood strips are,
to guarantee that I don't accidentally
add an extra measure of caulk.

It needs to be exactly 1/4" of wood and 1/16" of frame at this particular place,
or I'll have problems fitting the ports in.

That would be a bummer.

August 9, 2004 --
Here's the Starboard side.

Right now
they look really Blue.

Frankly, my paint job is just plain lousy. Putting the last coat on at the dock was a big mistake, as it was just too hot and buggy. These frames are barely acceptable, and I'm taking one of the port side frames home to sand it down and repaint.

However, my main concern is that they don't leak and
are structurally sound.

Sidebar -- I gotta tell you this. Do you see the scallops in the picture above? I actually formed a GARROTE out of piano wire and wood blocks, to try and work the old frames off without destroying the gelcoat. I wasted an hour on that and achieved 8 inches, and had to make three new garrotes. 5200 requires a grinder. A heat gun and a piano wire garrote can't overcome this stuff.
DON'T USE IT for normal boat projects.

With this design, the top four through-bolts
for the opening ports will go
though the cabin side at the very top.
I thought that this would be strongest.
It will also keep the bottom of the ports
about 3 inches off the deck.

I took a 15" long, 3.4" wide strip of 1/4" marine plywood, and embedded it in the caulk directly behind the holes for the ports. This will
allow the bolts to be cranked down hard.

The strips of wood are soaked in epoxy,
so if by some bizarre chance any water
gets this far they won't rot.

Cleaning every vestige of 5200 away
took 18 hours. I probably should have just taken a grinder to it. Look at this picture,
and see the legacy of 5200.

When the boat was originally built, the windows were glued in place using a Two Part adhesive called Plexus. It's very hard, structural, and tenacious, and some if it is still there. When the original (two generations ago) windows were removed, it looks like the gelcoat also ripped off with the Plexus, exposing fiberglass.
This made the 5200 even harder to remove.

After trying a number of tools, scrapers, heat guns, and other stuff, I found the fastest way to clean up this particular problem was using this sanding wheel on my Handy Dremel Tool. I went through six of them. It allowed quite a bit of control, but was slow.

The frames are screwed down with #6 oval head screws.
Really, the caulk is holding this on, but the screws kept it in place while the caulk cured. Using an oval head screw made it easy to perfectly center the washer.

By my calculations, this sheet of aluminum will expand/contract 1/16" over a temperature range of 30F-150F. (It *will* get that hot in the blazing sun.)
With this setup, the frames can move under the screws without touching the screws, so the anodized protection won't get scratched off and start corrosion. Of course, the boat itself expands and contracts, so the relative expansion should be a bit less.

August 14, 2004 --
Before caulking in the frames,
I propped the 3/4" thick epoxy-soaked plywood in place, and traced out the port holes.

Then, after the frames had cured enough to be stable, the interior plywood was set in place.
I smeared a very, very thin layer of caulk to glue them to the frames, and packed the edges with caulk to hold them in position and provide structural integrity.
Now I *know* it'll never leak.
It will also never come apart again, so it's really important that
the leak prevention works.
I *really* am building a permanent solution. With all these 1/4" thick gaskets of caulk. I believe it's safe.

After that second round of caulking had skinned over and it was safe to work nearby, the interior panels were attached and the port holes traced and cut.

I have some touch-up painting to do on the liner, to cover the spots where masking tape ripped paint off the headliner, and to cover the spots where the old curtain rods were attached.

I'm not going to permanently affix the interior panel to the construction. It will be held in place by a couple of screws, the bolts for the ports, and caulk for cosmetic fill.

If anything leaks, it should work down to the lowest point under the port spigots, and drip down inside the boat rather than pooling up at the lowest point and starting rot.

By the way, after cutting the holes in the 3/4" plywood, I sealed the edges with West System.

The cabin side is
exactly 1 1/8" thick.

I love it when a plan comes together!

The test port will be installed in the head, since I'm still in the process of painting and fixing up the liner there.

If I screw up and do something ugly, it's best to be in the head
because I can shut the door
and pretend it's not there.

Soon, I'll paint the head
bright white just like the panel.

Because this is the first fit, I took the window off the frame so it would be easier to work with, and because the posts sticking out the back would be a problem until I drilled holes to accommodate them.

The goal for the first port is to try and install it without grinding, and see if everything will magically work out.

I bought new 6mm screws (all the fittings are metric) that were 35mm long, but forgot to bring them to the boat, so I just used 40mm bolts with a bunch of washers. This is just for trial fitting and planning -- I'll use the right screws for the final installation.

With all of the flexible adjustment options on the ports, I don't think the back frame has to be perfectly parallel with the front in order for it to be leak free, so I'm going to see if I can make it overlap the ridge on the liner.
The top is 1 1/4" thick without grinding.

I bought this new screw driver just for this project. I should have bought one years ago! It has an adjustable torque setting, so you can dial in the amount of torque applied to a screw or nut, and just let it work until the clutch starts clicking. All the screws are then at the same torque.

Here's a picture of the outside. This is going to look nice!

With the bolts cranked down tight, the soft silicone gasket compresses almost flush,
so I can use up some of that fudge factor!

With the bolts tight against the 1 1/4" thick cabin, the adjustment nut for the hinges protrudes about 1/32".

I had thought that the limiting factor would be the spring on the bolt. However, you can tweak that spring and make it ride over the bolt without any problems. To the right, my thumb is resting against the gasket, which shows about 3/16" of room between the gasket and the spring.
This became very important later.

During the first test mounting, the port's frame flexed to conform to the irregular surface. That didn't seem right.
The interior panel needs to provide a nice, strong, flush surface.

JUST TO TEST THE PLAN, I stuffed some washers of various widths under the inside panel, so that it would angle up nicely and be flush with the ridge in the liner. If this works, I'll come up with a permanent way to shim the panel in place.

I took two 7/16" drill bits and stuck them through the mounting holes, and used them to position the drilling template inside the boat.
Dowels will work, too, but
I have extra drill bits.

The template has guides for the four holes.

I stuck a dab of tape around a 3/8" drill bit, so that I wouldn't go too far and drill through the entire cabin.
That would be a bummer.

The NFM instructions say to use the 3/8" pilot hole
to guide a 5/8" counter bore bit.
I don't see a need for a 5/8" hole around these bolts.
The bolts are actually 15/32" external diameter, and
if it's a closer fit the hole will provide additional strength and avoid straining the adjustment nut on the frame.

With a 7/16" hole, the bolt won't fit easily,
but can be screwed into the wood like a self-tapping bolt.
That's great for the dogs at the bottom,
since it will make them super secure.
If this were fiberglass, I'd make the holes bigger,
but my design has wood at that particular spot.

The holes at the top need to be big enough for the hinge bolts to slide in and out without turning. 15/32" is the minimum size for that, but I only have a 1/2" bit so that's what I'm using.

The hole just happens to be right on the edge of the liner ridge, so I used my Handy Dremel Tool to grind up a bit and make a flush place for the 1/2" drill bit.

Here it is, bolted in place for the second time.

With the interior panel gradually shimmed out so it's flush at the top,
and at the bottom, there's a nice flat surface to crank the frame down against.

The port was installed at near maximum adjustment. This is too tight, as you can see to the left. I found that it was easier to adjust too far initially, then loosen them to fit.

The proper adjustment
is to loosen the top bolts so that
there's about 1/4" clearance
at the bottom before encountering resistance from the gasket.

The cool thing about this picture is that it shows some real solid laminate along that ridge, so if I have to grind into it for a specific port, the odds of grinding through it and creating a structural repair job are slim.

However, it looks like I won't need to touch the grinder. Hooray.

Here's the hole, drilled out to 1/2" wide and 1" deep. The end of the drill bit just kissed the little piece of plywood that's stuck in the caulk on the other side, so I know that I have 1/4" safety margin there.

You can see here that there's about 3/16" of room to tighten up the hinge, if for some reason the port gasket starts to leak and it needs to be pressed harder against the frame.

That should be more than enough room for future adjustments.

At the bottom, the dogs have even more room. The bolt is tightly threaded into an inch of plywood (in addition to the nut that's built into the frame) and isn't going to move no matter how much the boat shakes, rattles and rolls.

Everything is going to work out fine!
The external soft silicone gasket compressed much more than expected, and the EDPM window gasket barely compresses at all. That gave me extra room, so 1 1/4" is no problem.

So after a full day of tinkering, the first port is trial fitted, and I have a plan in place that I'm confident will work.

I need to come up with a secure, structural way to shim the starboard panels out about
3/32" at the top and flush at the bottom.

Next weekend I'll trial fit the other seven,
then take them back off,
seal with butyl and
screw them down for the last time.

But for now, I'm going to spend the rest of the day admiring this port.

After this much work,
the satisfaction is profound.

I think it's beautiful.

The spring on the hinge keeps the port open without needing to clip it up, although there's a little steel pin on the bottom of the window so it can be secured open if desired.

Note the angled lip on the inside edge of the ports. This guides the water into the spigots.

It's great! After the water drained away, I opened the port and the only water that came inside was a few drops that clung to the gasket.

There's a little pool in the groove at the bottom of the lip, but that's only because the ports are new and surface tension is keeping it in place.

I closed the port and dumped a cup of water on it, just to watch the spigots work.

This isn't for leaks,
it's testing
the installation angle.

The recommended maximum installation angle for these ports is 22 degrees, and my house is at 20.

It took a single day to dry fit the remaining seven ports.

The gaskets and O-Rings in the
Tri-Matrix ports will keep them from leaking, while I think about
the final installation.

The ports aren't perfectly aligned, so there's a bit of adjustment needed.
There's enough wiggle room around all the holes to adjust them so they're aligned properly.

I really hope that
painting the frames Blue
wasn't a bad idea!

On the port side, the little ridge in the headliner was exactly 1/4",
and no shimming or messing about with the interior panel was required.

I should have made the interior panel from 1/4" marine ply on the port side, and from 3/8" marine ply on the starboard side.
I'll bet that whoever initially put together the mold for the fiberglass liner on this boat never thought that someone would notice
that the two sides were different.

The liner ridge also gets wider towards the ends.
On the starboard side, there are spots
where it exceeds 3/8".

I'm going to live with the installation dry fit for a while, and just look at it and think about it.
Perhaps grinding a bit with my Handy Dremel Tool would be a good idea in some places.

Also, I didn't do a great job matching the interior panel with the ridge for the entire length.
That's why we have caulk!

Anyway, everything's in a leak-free state,
it's August and it won't rain for months, so there's time to think about everything and come up with a plan for the final, finishing effort.

September 2004 -- Time to finish.

My Handy Dremel Tool made short work of the rough edges around the old curtain rod holes.

That 4200 is pretty solid stuff
once it cures, so I used it
to shim out the interior panels.
There is still an appropriately thick washer
next to the bolts, glued in place with 4200.
I used just enough caulk to make it solid around the frame, and only used it at the top.
If there's a leak, I want it to come out the bottom of the interior panel, not get stuck inside where it will start rot without anyone knowing.

The one area where
the interior ridge was too thick
was ground down a bit.

This turned out to be a good idea, as I
would have had huge problems during the final fitting.
(When I actually did it, I had my left hand
holding a shop-vac hose next to the wheel.)

I just traced the edge, and ground down with
the Handy Dremel Tool, then
smoothed the edges by hand before painting.

Before the final fitting,
all the remaining raw edges were painted with West System epoxy to seal the wood,
paying special attention to the bolt holes.

Given everything I'm doing, I seriously doubt that anything will ever leak, but there's no point taking chances. If the marine plywood ever starts to rot, it will be impossible to replace without destroying everything. That's why the ports are designed to channel leaks inside the cabin, so they'll be obvious and
can be addressed immediately.

I'm using Butyl Tape from McMaster-Carr to seal the windows. Butyl is like Silly Putty with an attitude.

The ports are designed with gaskets and O-Rings to prevent leaks, and are not supposed to need caulk. I'm not taking any chances. One speck of dirt can destroy an O-Ring seal.

The ports didn't leak during the short time I had them dry fitted, and the folks at Newfound Metals did extensive testing.
But no one has left these out on a boat in the sun for ten years.
I'm not willing to wait ten years to find out.
This installation will have multiple layers of leak protection,
and I'm never going to deal with it again.

Butyl doesn't stick to skin,
but it sure likes rubber gloves!

After about 2 minutes,
I just started working it
with my bare hands.

Butyl never sets up hard, but remains goopy and sticky forever.
It has no structural properties,
but does a great job as a water barrier.

By the fifth port I had it figured out.

The butyl should be just as thick as it should be when the ports are finished -- just thick enough so that it will adhere to both surfaces and leave a bit of a gasket.

It's tightly pressed against the
sides of the carriage nuts.

By pressing it up against the sides a bit,
it also blocks light from coming through the white plastic. That's purely cosmetic,
but the ports look funny on a bright day when the outside edges of the interior are glowing.
Pushing a thin layer of butyl up the
side of the frame prevents that.

Here's one of the later ports,
which didn't require clamping.

Note that when this picture was taken,
the frame is slightly bowed
due to the thick layer of butyl.

I went back a few hours later to clamp down on it, but the butyl had gooshed into place and the frame was perfectly flat.

Where the butyl was too thick, I helped it goosh out of the way
by putting a large clamp around the frame.
The wood blocks prevented marking the stainless steel.

This also allowed the bolts to be cranked down hard
without using the bolts to compress the frames.
It helped avoid galling the stainless steel fasteners,
and allowed the bolts to be cranked down hard
without deforming the stainless steel strips on the frame.

One of the ports has a little teeny flaw.

It came like this, I didn't do it.

It looks like the glass was compressed too much during manufacturing. I don't know if this is bad,
or just cosmetic, so it bears watching.

It was clear that four of the ports came from one production run, and four came from a later run, and the manufacturing line had been learning along the way. There's nothing wrong with any of the ports, but four of them fit together a little better.

During the final fitting, I did some tweaking with my Handy Dremel Tool to get the windows perfectly aligned. But when I cranked down on them, they went back to where they want to be. It's not really noticeable.
I figure that however the boat's trimmed, one of the windows will be perfectly level.

Over the galley, to port.

That one window
behind the cabinet was
a real bear to bolt in.
I have it dogged down
very tight, and
don't anticipate ever opening it.
I might think of some way to replace the dogs , and just
leave it permanently shut.