How to Adjust Wall Closures in Revit

Open any set of drawings a contractor has actually built from and go to the window details. The plaster return into the reveal is drawn. It has a dimension on it. Somebody made a decision about whether the render stops at the frame, runs past it, or dies into the masonry, and that decision costs money on site.

Now open the Revit model those details came from and look at a plan detail at 1:20 scale. Very often the finish layers run straight past the opening and stop dead at the structural core, because nobody set the wall closure. The detail was drafted over the top in 2D and the model quietly disagrees with the drawing. Every quantity that comes off that model, every enlarged plan, every section cut through a jamb inherits the error.

Wall closures are the setting that fixes this, and they take about four minutes to understand. They are also one of the few Revit behaviours that live in three places at once: the wall type, the window or door family, and the reference planes inside that family. Miss any one of the three and the other two do nothing.

The walkthrough below follows that recording step by step, with links to the exact moment each setting is changed, and then adds the project context a short tutorial does not have room for.

What a wall closure actually is

A wall closure is the rule that tells Revit how the non-structural layers of a compound wall behave when they meet an opening. Left alone, a multi-layer wall in Revit treats an insert as a hole punched clean through every layer. The finishes are cut off flush with the opening face. In reality the plasterer carries the render around into the reveal and returns it against the frame, so the modelled wall and the built wall stop matching at exactly the place the details get drawn.

The video opens by demonstrating the end state rather than the settings. Dragging the window through the wall thickness makes the plaster follow it, on both the outside and the inside face, and the reveal keeps closing correctly at every position. That behaviour is not a modelling trick applied once. It is a set of parameters that, once configured, holds for every instance of that family in every wall of that type.

Step one: the wall needs layers, and the right layers need Wraps ticked

Nothing here works on a single-layer wall. The first thing the video checks is that the wall in question has a real build-up. Selecting the wall and opening its structure shows a stack of layers rather than one generic thickness, which is the precondition for everything that follows.

The part people skip is the next one. Each layer in a compound wall carries a Wraps tick box, and only ticked layers take part in a closure. In the recording the finish is wrapped on the outside face and again on the inside face, which is why both returns behave later on. A layer with Wraps unticked will be cut flush no matter what you do in the family, and this is the single most common reason a correctly configured window still produces a square-cut reveal.

Think about which layers genuinely wrap on a real build-up. The video only wraps the finish layers, inside and out. The rest of this table is normal practice rather than something the recording shows:

Layer	Wraps in practice	Why
External render or plaster	Yes	The plasterer returns it into the reveal against the frame
Insulation	Usually yes	Continuity of insulation at the reveal is a cold-bridge requirement
Structural core (masonry, concrete)	No	The core is the thing the opening is cut through
Cavity or air gap	No	It is a void, and wrapping it produces nonsense geometry
Internal plaster or plasterboard	Yes	The internal reveal is finished and painted

Getting this list right in the wall type is worth more than any downstream setting, because it is the layer definition every instance of that wall inherits.

Wrapping at inserts versus wrapping at ends

The same dialog carries two separate wrapping controls, and they answer two different questions. The video points at both: wrapping at inserts, which is the openings, and wrapping at ends, which is where the wall itself terminates.

They are worth keeping mentally separate because they fail in different ways. Wrapping at inserts is what you notice in an enlarged window detail. Wrapping at ends is what you notice at a party wall junction, a door opening into a shaft, or the end of a wall that stops in the middle of a room, where an unwrapped end shows the raw core in plan. A model can be immaculate at every window and still expose bare blockwork at three wall ends on the same sheet.

At this point the recording makes a recommendation rather than a demonstration: rather than driving the behaviour from the wall type, control it from the family instead. That is the workflow the rest of the video builds on, and it is the one worth learning, because a wall type is shared across a whole project while a window family can carry its own intent.

One honest caveat. The recording never sets the wall type and the family to contradictory values to show which one wins, so this article does not claim a precedence rule. If you need to know what happens when the two disagree in your own template, build the case and look, rather than trusting a blog post. What the video does establish is the working sequence: enable the wraps on the correct layers in the wall type first, then drive the actual closure behaviour from the family.

The Wall Closure parameter inside the window family

This is the control most people never open. Selecting the window and going to Edit Type exposes a type parameter called Wall Closure, and changing it changes the reveal immediately in the model.

The recording steps through the values on camera:

With the parameter set to Both, the finish returns on the outside and the inside, which is the state the video opens in.
Switching it to Neither makes the plaster run straight and die into the wall with no return at all.
Setting Exterior wraps only the outside face, stopping at that edge, and the return still tracks the window as it moves.
The interior side behaves the same way in mirror image.

The result column below is what the video puts on screen. The Typical use column is not in the recording, it is where each setting tends to belong on a real build-up:

Value	Result at the opening	Typical use
Neither	No return on either face, finishes cut flush	Openings finished with a separate lining or surround family
Exterior	Outer finishes return into the reveal only	Rendered outer leaf with a dry-lined internal reveal
Interior	Inner finishes return only	Fair-faced or clad external wall, plastered internally
Both	Returns on both faces	The default for a plastered masonry wall, and the state shown in the video

Because Wall Closure is a type parameter, changing it changes every instance of that type in the project at once. That is a feature, not a limitation: it means the decision is made once, in the family, and every window of that type on every floor obeys it. It also means you cannot vary it window by window without duplicating the type, which is usually the right constraint anyway.

Revit also offers a By host option on this parameter, which hands the decision back to the wall type. The video does not use or demonstrate it, so treat it as a pointer to look up rather than a step in this workflow.

Making the closure follow the window through the wall

Setting the closure is half the job. The behaviour that makes it useful is that the return moves with the window, so pushing a window towards the outer face shortens the external reveal and lengthens the internal one, automatically, without touching the wall.

That behaviour is authored in the family, and the video goes in to show it. Edit Family opens the window, and at ground level there are two reference planes sitting either side of the frame, offset from the outside face. Those planes are what the wall closure snaps to. The geometry of the frame is not what drives the return, the planes are.

Back in the project, the window’s position within the wall thickness is driven by an instance parameter, and it can be typed as a value or nudged with the arrow controls on the fly, window by window. The demonstration file in the video is German-localised, so the parameter name on screen is not the one you will see in an English template, and this article deliberately does not put an English name on it. What matters is the mechanism: instance parameter, so every placed window can sit at its own setback, and the closure follows each one independently.

That combination is the whole point. Type-level decision about whether the finish returns, instance-level control over where it returns to.

The tick box that silently disables all of it

Here is the setting that wastes an afternoon. A reference plane in a family only becomes a closure target if it is explicitly told to be one. Selecting the reference plane exposes a Wall Closure property, and it has to be switched on. The video is blunt about the consequence: when it is not activated, you cannot use that plane in the project.

There is no warning, no error, no clue in the project environment. The Wall Closure type parameter still shows Both, the wall layers still have Wraps ticked, and the reveal still comes out square, because the family has not offered Revit anything to close against.

If you have ever had a family that refuses to wrap while an apparently identical one from a colleague works perfectly, open both and compare this tick box before you compare anything else.

One plane or two, and what each gives you

The recording ends with a genuinely useful experiment rather than a summary. Turning the Wall Closure property off on the lower reference plane, then loading the family back into the project and overwriting the existing version, changes the result visibly: the plaster now runs all the way to the single remaining closure plane.

That is a design lever, not a bug. Two closure planes give you a return that stops at the frame on each side. One closure plane gives you a finish that carries through to a single line, which is how you would model a reveal detailed to run past the frame and be sealed against it. The video notes plainly that if you want that behaviour, you keep only one plane flagged as the wall closure.

Note the load step in that sequence. Loading a revised family into a project where instances already exist prompts you about overwriting the existing version, and the recording takes the overwrite. That is the correct choice when you are iterating on your own family, and it is worth being deliberate about it on a live project, because it changes every placed instance at once.

Where this actually bites on a real project

The rest of this article is professional context, not content from the video.

Wall closures look cosmetic in a 3D view, which is why they survive so long unfixed. They stop being cosmetic the moment the model has to produce something.

Enlarged plans and details. A 1:20 or 1:10 plan through a jamb shows the closure at full size. If the model does not close, either the detailer draws over the top in 2D or the sheet ships with a wrong reveal. The first creates a permanent maintenance liability, because the 2D linework no longer updates when the window moves. If you build details from the model, the model to detail sheet workflow is only as reliable as the geometry underneath it.

Quantities and finish areas. Plaster and render take off by area. A wall whose finish layers are cut flush at every opening under-measures the reveal returns across the whole building. On a facade with three hundred windows that is a real number, and it is the kind of error that never gets caught because nothing looks broken.

Thermal and cold-bridge reporting. Where insulation is one of the wrapping layers, the reveal is precisely where a linear thermal bridge is assessed. An unclosed reveal is a model that says the insulation stops at the opening, which is both wrong and, in some markets, the geometry an energy consultant will export and use.

Coordination with the facade package. A specialist window subcontractor reads the reveal depth off your model or your details. Ambiguity there produces an RFI, and RFIs about reveals arrive late, because nobody looks at them until the frames are being ordered.

Standards and team conventions

Also not shown in the video.

Wall closures belong in the template, not in the hands of whoever places the first window. A few conventions that hold up:

Decide the Wraps pattern per wall type once, in the office template, and document it alongside the layer functions. Anyone building a new wall type copies an approved one rather than defining layers from scratch. The project template and office BIM standards guide covers where this sits in a wider template.
Author every opening family with closure reference planes flagged, even when the current project does not need the return. A family that lacks the flag cannot be fixed from the project, only from the family, which means it fails at the worst possible time.
Pick a house default for the Wall Closure parameter and state it. Both is a sensible default for plastered masonry. Whatever you pick, the point is that a reviewer knows what a deviation looks like.
Treat a localised family as a translation risk. If your library came from another language version of Revit, the parameter names carried over with it. Rename them on the way into your library rather than leaving a mixed-language parameter list for the next person.
Record the reveal intent in the family description, so the person choosing between two similar window types can see which one is built for a returned reveal and which one is not.

A QA check that takes two minutes

Not from the video, but this is how you catch it before a sheet ships.

Cut a plan region or a section at 1:20 through one window of each wall type. If the finish stops flat at the opening, something in the chain is off.
Open the wall type and confirm Wraps is ticked on the layers that should return, and unticked on cavities and the structural core.
Select the window, Edit Type, and read the Wall Closure value. If it is Neither and the reveal is meant to close, that is your answer.
If the first two are correct and it still fails, open the family and check the Wall Closure property on the reference planes. This is almost always where it is.
Drag a window through the wall thickness and watch. A correctly configured family keeps the reveal closed at every position. One that only looks right at its current setback is not actually driven by closure planes.

Common mistakes

Fixing it in 2D. Drawing the return with detail lines on the sheet is faster today and wrong forever. The lines do not move when the window moves and they do not appear in any other view.

Treating it as a wall problem. The instinct is to edit the wall when the reveal looks wrong. In this workflow the wall only supplies the layers and their wrap flags. The behaviour lives in the family.

Modelling the reveal as geometry. Adding a solid to the family to fake the plaster return produces something that does not follow the wall build-up, does not schedule as finish, and breaks the moment the family is placed in a wall with a different layer stack.

Editing the wrong Wall Closure. The parameter on the family type and the tick box on a reference plane share a name and do completely different jobs. One decides which faces return, the other decides whether there is anything to return to.

Assuming a family library is configured. Manufacturer-supplied and downloaded families very often have no closure planes flagged. Test one before you standardise on it.

Where to go next

If you want the layer underneath this, the useful next step is family authoring itself: how reference planes are set up, how they are constrained, and why a plane rather than a face is the thing you drive geometry from. The Revit family creation fundamentals guide covers that theory, and the parametric window family walkthrough builds a window from an empty template so you can see where the closure planes would go.

Then set it up properly once, in your template, so that the next three hundred windows close their reveals without anyone thinking about it. That is the whole return on a four-minute setting.

If you want the full BIM workflow rather than a single setting, the Archgyan course library covers Revit and coordination the way firms actually run them.