Bathroom door and pan

1Where the post goes

The question that needed a drawing: does the post sit inside the shower wall or outside it? Neither. It sits IN the wall line and becomes part of it. The post replaces the wall panel across the opening, its shower face flush with the wall face, and it is flush on the living-space side too.

The stop is not a strip screwed on afterwards. It is a rebate, a step machined into the post itself, and the door closes onto its shoulder. That is what lets the door sit flush without anything standing proud.

2Parts

3Wood choice, and the way round it

Screw withdrawal per inch of thread is roughly 980 lb in hard maple, 940 in white oak, 540 in poplar, 400 in pine. The concern is not static load, it is fatigue: each shock cycle lets a screw crush its own hole a little, the hole grows, and a loose hinge cracks the microcement around it.

But that only applies if the fixing is a wood screw. Through-bolt the hinge or set threaded inserts and the wood species stops mattering, which makes the poplar already in stock perfectly usable. Oak or maple if screwing, poplar if bolting.

4The floor, and the order it goes down in

Yes, it gets waterproofed before the microcement, and the order is not optional. Microcement is a finish, not a waterproofing layer. On a wall its sealer is enough. On a floor you stand on, the sealer abrades where your feet are, so the waterproofing has to live underneath.

Use the waterproofing that your microcement supplier approves. Mixing a membrane from one brand with microcement from another is where adhesion failures come from, and the failure shows up as delamination six months later. Ask them what their approved wet-floor build-up is and buy that. Same rule that picked your sealer.

5Kerb height, and why

The number is not set by showering, it is set by parking off level. Water left in the pan tilts with the van, and the van sits parked for hours.

But the better control is the fall, not the kerb. If the pan drains completely there is nothing left to spill. Chase flat spots out of the pan and the kerb becomes a backstop rather than the thing you are relying on.

6Pan and drain parts

7The lip, because the jamb cannot get deeper

The wall at the opening is 0.5 in, and drawers run the kitchen face for the full length of it. That combination removes the obvious answer. A door sits flush and still stops only if the stop lives behind it inside the wall, which needs the door thinner than the wall: at 0.5 that leaves 1/8 in of shoulder for a 3/8 door, and 1/8 in is gone the moment a 20 x 76.5 door cups. The fix would be a jamb deeper than the wall, projecting on the kitchen side, and the drawers forbid it.

So the stop rides on the door, and the reason that is not a compromise is geometry. A rebate is a shoulder and can be slipped past. A lip is a face bearing and cannot, at any thickness. Only the lap has to beat the door's cup, not the thickness. So the lip is 0.0625 in of flat bar lapping 0.5 in, the door reads as flush with a fine shadow line, and nothing carries a bead.

It works only because the door opens toward the lipped side. The lip lifts straight off the wall everywhere at once, so nothing binds at the head or near the hinge. Reverse the swing and this detail is impossible and the deep jamb comes back as the only answer.

Make the lip metal, not microcement. Black anodised flat bar, about 1 in wide by 0.0625, on the strike edge and the head, roughly 96 in of it. It takes the impact instead of the finish, protects the door arris, and it is the trim family already set by order 189513. Note that is flat bar, not angle: an angle's face leg starts at the corner so it has nothing to overhang with, which is also why the WGS angle stays allocated to E4.

Put bumpers behind the lip, clear pads or a thin EPDM strip, so it is not aluminium hammering microcement on every washboard hit. It kills the noise and takes up the tolerance too.

Hinge edge stays flush and unlipped, so an ordinary hinge works in the jamb. The 0.0625 in asymmetry between that edge and the strike edge is invisible.

8Hinge

Continuous hinge, 316 stainless, full height. Not butts. This is 30 lb of door, 76.5 in tall, in a vehicle that shakes for a living. On butt hinges the whole weight lands on eight or twelve screws, and every washboard mile works them. A continuous hinge spreads the load over the entire edge, so no single fixing carries the door, and it cannot sag out of a 0.0625 in gap the way three butts can.

It self-jigs, which matters here: there is no aligning of separate leaves, so the door hangs true first time in an opening that may not be square. And it stiffens the hinge edge, which is worth having on a panel this tall.

The screws are the weak point, not the hinge. A hinge leaf screwed into the edge of 0.5 in plywood has almost nothing to bite. The door needs a solid hardwood lipping on the hinge edge, glued on before the microcement goes anywhere near it, and the hinge screws go into that. The frame leaf goes into the post, which is solid, so that side is fine.

316 rather than 304, because this knuckle sits in the wet zone and will hold water in it. Not zinc, not plated: salt air is the threat model for the whole build.

Blank, no holes. Drill to suit the lipping and the post rather than taking whatever the factory spacing lands on. And 0.06 leaf minimum: 0.037 marine stock exists and is sold at Lowe's, and it is a locker-lid part, not a door part.

Take 1.5 in open width, not 2. Open width splits evenly between the leaves, so 1.5 gives 0.75 per leaf, which is exactly the lipping width. A 2 in hinge wants 1 in of bearing on a 0.75 in lipping and you would be screwing into fresh air.

On length, do not chase full height. Stock is 72 in and the door is 76.5. That is 94 percent of the leaf, and a continuous hinge is not there to reach the ends, it is there so no single screw carries the door. Run it from the top and finish short of the kerb, where a gap is wanted anyway. The 96 in is only worth hunting if it is genuinely 316.

9The frame: species, section, and one trap

Quartersawn white oak, with sapele as the alternative. Crosscut Hardwoods, 3065 NW Front Ave, 503-224-9663, stocks sapele in both flatsawn and quartersawn and is a few minutes from PDTM.

Not plywood, and the reason is specific. The hinge screws go into the narrow 0.5 in reveal edge of the post. Solid hardwood holds an edge screw. Plywood edge does not, and that is the one fixing that cannot be allowed to creep on washboard.

White oak for screw holding and toughness, and it is closed pore and rot resistant, which is cheap insurance next to a shower. Sapele if you would rather have easier machining and a little less movement, though its interlocked grain tears out if the blades are not sharp.

Quartersawn is not a preference here. It roughly halves seasonal width movement, and that is the difference between microcement over the post surviving and crazing.

Rectangular, and flatter than the word post suggests. It is a board standing on edge in the wall plane, not a timber.

Ease every arris that meets microcement. A sharp wood corner under a brittle coating is a crack initiator.

The trap: microcement over solid wood will craze. Wood moves seasonally, microcement does not, and the post face sits in the finished wall plane. Three ways out, in the order I would take them. Cap the post's shower face with a thin ply or wedi strip so the coating sees a stable substrate. Or run fibreglass mesh over the wood to panel junction, the way the shower spec already does at the wall and floor corner. Or stop the microcement at a black anodised reveal, which is the E4 move and an existing part of the trim vocabulary.

10Latch, and the two positions

Your instinct is right that this is two different jobs and one catch cannot do both. It is actually three.

One change to what you described: the driving lock has to work from the kitchen side, not the bathroom side. If it can only be set from inside, you either lock yourself in or you leave the door open to reach it. The bathroom is empty when you drive, so the lock belongs where you are.

The good news is that the two-position behaviour you want is real hardware, not a compromise. Marine flush latches are made with detents: a light position that just holds the door, and a full throw that positively locks it. Flush mount means nothing projects, so the drawers still clear. Look at Perko and Southco flush cam latches, and check the door thickness range, because most are quoted from 1/4 in and 0.5 in is at the thin end.

The catch bolt goes through the door EDGE into the post, so it does not conflict with the lip, which lives on the face. Keep every part of all three mechanisms on the kitchen side or inside the door. Nothing goes on the shower face.

Privacy and transit stay separate parts on purpose. A single lock doing both means the one you use naked at 2am is the same one holding the door on a mountain road, and they want different actions and different failure modes.

11Still open