Every sliding window hides part of itself. A two-panel slider hides one panel behind the other, which is why it can only ever give you about half of its own width as openable area. A stacking system hides several panels against one jamb. A pocket system does something categorically different: it takes the panel out of the opening altogether and parks it inside the wall.
That one sentence is the whole story, and it is almost never read carefully. Because the word pocket appears in a window catalogue next to words like thermal break and multi-point locking, it gets filed in the buyer's head as a feature of the window. It is not. It is a description of where the panel goes when you are not looking at it, and where it goes is a hole in a wall that somebody else has to design, frame, flash, insulate, drain and eventually open up again.
This article is about that second hole. It is the hole that never appears on a window schedule, rarely appears as a line item in a budget, and is drawn on a sheet that the window supplier in most projects never sees.
A window schedule is a list of openings. Each row carries a reference, a size, a configuration, a glass build-up and a finish. On a pocket project, every one of those rows is a lie by omission, because each of those openings needs a second hole the same size next to it, and that second hole has no row.
The result is a documented-by-nobody gap. The architect draws the cavity on a plan or a detail. The structural engineer sizes a beam whose span the window drawing never mentions. The framer builds a split-stud channel. The waterproofer treats a void that is outside the weather line whenever the window is open. The window supplier ships a unit whose rough opening is one of two published widths and may never be told which one was built.
| Element | Who decides it | Which drawing carries it | Who usually never sees it |
|---|---|---|---|
| Panel, track, rollers, seals | Window or door system supplier | Window schedule, elevation, supplier sections | The framer, until delivery |
| The cavity itself | Architect and framer | Plan, partition or wall detail | The window supplier, in most projects |
| Header and load path | Structural engineer | Structural or framing drawing | Anyone pricing the window only |
| Flashing and drainage of the cavity | Design team and waterproofer | Detail sheet, sometimes a note only | The supplier's warranty logic entirely |
| Access panel for servicing | Architect, in theory | Almost nobody, in practice | The owner, in year eight |
The sentence worth remembering: a pocket window is two openings. One you look through, and one you never see. Only the first one gets a line in the schedule, and only the first one has a price.
Start with the most useful public artefact on this subject: the published rough-framing detail for a large-format pocketing sliding door. On that one drawing, the same unit is dimensioned twice on the horizontal plane — once as a rough opening width not including the pocket, and once as a rough opening width including the pocket.
Sit with that for a moment. A window schedule has one width column. This system has two, and they differ by the width of the panel that disappears. Which one gets written down determines whether the wall gets built long enough.
Three further details on that same published sheet are worth more than any marketing copy about minimalist sightlines:
There is also a vertical trap. Where a flush sill is used, the published detail requires the bottom of the recess in the floor to sit a specified distance below the top of the finished floor, and warns that failing to do so can leave the rough opening undersized or mislocated. The same manufacturer notes that the flush sill option is not recommended over wood floor construction. So the threshold detail that most people specify a pocket for is the one with a published floor-condition caveat attached.
The lesson is not "be careful with dimensions". It is narrower and more useful: the pocket converts a tolerance problem into a sequencing problem. A normal window can be measured after the wall is finished and still be installed correctly. A pocket cannot, because half of it stops existing the moment the wallboard goes on.
The single most commonly missed structural consequence: the header over a pocket opening has to span the visible opening plus the full pocket cavity. Not the opening. The opening and the hole beside it.
A published contractor guide for large-format pocket doors puts it in plain numbers: an eight foot single-pocket door means a header carrying roughly sixteen feet, and a twelve foot double-pocket means each header working over roughly twelve feet. A separate retrofit guide gives the framing rule of thumb that the rough opening width ends up at about double the panel width plus an inch.
| What is being described | The number people use | The number the structure actually sees |
|---|---|---|
| Window schedule width | The clear opening | Roughly double it, plus the pocket framing |
| Header span | Sized to the visible opening | Opening plus cavity — the most commonly under-specified item in the whole assembly |
| Deflection limit | Whatever the code span table gives | Any downward movement at the head pinches the frame and stops the panel moving |
| Fixing into the head | "Fix it to the header" | Published guidance calls for a minimum fastener embedment into the header; the unit is captured at the head and rides on the sill |
On a load-bearing wall the consequence is an engineered beam, and in some jurisdictions a stamped review, because the wider the pocket the larger the member. On a non-load-bearing partition it is still a continuous header with cripple studs above it and below the pocket to keep the wall coherent. Neither of those appears on a window drawing, because a window drawing does not know the wall is hollow.
And the wall being hollow has one more structural consequence that is worth stating bluntly. Public guidance on cavity sliding systems is consistent: never cut a pocket into a wall that carries floor load or acts as a shear or bracing element without engineering it. The bay that hosts the pocket has had its vertical studs removed or split. Whatever lateral resistance that bay contributed has to be provided somewhere else, on somebody else's drawing, before the first panel is hung.
If a pocket project fails, it usually fails here, and it usually fails on a drawing that was signed off months earlier.
A nominal 2x4 stud wall finishes at roughly four and a half inches, about 114 mm. A nominal 2x6 finishes at roughly six and a half inches, about 165 mm. Published contractor guidance for pocket systems says the 2x4 is almost always insufficient, that the practical minimum is around four and a half inches of clear finished cavity, which means building out to a 2x6 or a doubled 2x4, and that panels wider than three feet need at least five inches of clear cavity. Retrofit guidance repeats the same message in a different way: the minimum finished wall thickness is about four to four and a half inches, and a 2x4 wall at three and a half inches often has to be furred out.
Publicly published figures differ by market, and the differences are the point rather than an inconvenience: one heavy-duty North American frame is documented as being designed for a minimum three and a half inch structural wall, while European systems publish finished wall options in 100, 125 and 150 mm assemblies. Public Chinese guidance for domestic cavity sliders suggests a finished wall thickness of at least about 140 mm, with internal clearance of not less than about 50 mm for a leaf of 40 to 50 mm, and roughly 10 mm of clearance left at the jamb. A published European figure for door operating clearance in a pocket is 8 mm. These are product-specific, not universal, and the honest conclusion is that there is no such thing as a standard pocket wall.
Now bring that to our own published numbers. The table below uses the frame wall width we publish for each series — that is the frame's depth into the wall, which is the only depth dimension in our catalogue — and compares it against the two wall thicknesses above. It is deliberately conservative: it adds a nominal clearance allowance, because a cavity needs the frame depth plus whatever operating clearance your pocket system supplier specifies. We do not publish a pocket cavity dimension, and this table is not one.
| Series | Published frame wall width | Fits a finished 2x4 wall (about 114 mm)? | What the wall has to be |
|---|---|---|---|
| AW90 passive window | 90 mm | Yes, with the least margin of any option | The only series in our catalogue that leaves meaningful clearance in a standard light stud wall |
| SL112 / SL167 sliding door | 112 mm single and two track; 167 mm three track | 112 consumes essentially the whole wall; 167 does not fit | 112 needs a built-out wall to leave clearance; 167 needs more than a 2x6 |
| C113 / C116 / C120 casement | 113 / 116 / 119 mm | No practical clearance left | These are outward-opening casement series; the depth is the frame depth, not a cavity dimension |
| TS130 side press window | 130 mm | No | Needs a 2x6 as a starting point, before the pocket system's own clearance |
| SL126 / SL182 sliding door | 126 mm two track; 182 mm three track | No; 182 exceeds even a 2x6 | Note that the deeper frame is the three-track version. Frame depth is a property of the series, not a function of how many tracks it has |
| SL155 sliding door | 155 mm four track | No | Close to the finished thickness of a 2x6 with almost nothing left over |
Two traps live inside that table, and both of them are finish traps rather than window traps.
The first is that finishes thicken the wall without thickening the cavity. Public guidance on cavity systems lists tile, stone, acoustic panels and decorative wall systems specifically: they add to the finished wall thickness and add nothing to the usable pocket clearance. A wall that grew by twenty millimetres of stone is a wall whose pocket just got twenty millimetres tighter.
The second is intrusion. The same guidance warns that screws that are too long, bowed framing, excessive joint compound and uneven boards all intrude into the pocket and mark the panel as it travels. In a normal wall a proud screw is a cosmetic defect. In a pocket wall it is a machine fault, and it is installed by a trade that has no idea the cavity exists.
A bay that hosts a pocket stops being a wall bay and becomes a machine room with no door. The consequences last as long as the building does, and they are almost never recorded anywhere the next owner will see them.
| Thing that would normally live in a wall bay | Why it cannot be there any more | What has to happen instead |
|---|---|---|
| Full-depth studs, bracing, shear resistance | The studs are removed or split to form the channel | Bracing is relocated by the engineer; never assume full-depth studs survive in the pocket zone |
| Cables, junction boxes, switches, sockets | Anything in the panel's swept volume is destroyed by the panel | Published guidance puts services outside the pocket zone by a stated margin; plan switches on the solid wall before first fix |
| Plumbing and drains | A leak inside a sealed cavity is the worst possible place to have one | Route water and waste lines away from the pocket wall entirely |
| Ductwork and grilles | The bay is a channel, not a cavity | Relocate before the framing drawing is issued, not after |
| Insulation, fire stopping, acoustic stopping | The panel occupies the bay when it is closed | Accept a thermal weak point and manage it at the perimeter; confirm fire and acoustic requirements with the design team |
| Anything you screw to the face later | Long fixings for cabinets, handrails, shelving, televisions and trim can enter the moving panel zone | Specify reinforcement approved by the pocket supplier, or build a separate service wall |
You are not buying a window. You are converting a length of wall into a machine room with no door, permanently, in a building that will outlive the drawings. The least you owe the next owner is a note on the as-built drawing and, if you can bear it, a discreet mark on the wall itself saying that there is a moving panel behind it.
Here is a design consequence that is obvious once stated and almost never designed for: in a pocket configuration, the open position removes the thing you use to close it.
Accessibility frameworks for sliding doors address clear width, maneuvering space, hardware position, hardware usability and opening force. Public guidance written for cavity systems makes the point directly: hardware has to remain exposed and usable when the panel is fully open, and a panel that disappears completely may look clean while a recessed edge pull is genuinely difficult for some users. Two published remedies exist — stop the panel slightly proud of the pocket, or specify a compatible push-to-open device — and both of them change the clear width, which means accessibility and travel have to be confirmed together before anything is ordered.
The same nominal-versus-clear trap that governs every opening in a building governs this one, and it is worth naming because it catches people who have already learned it elsewhere. A so-called 32 inch entry door does not deliver thirty-two inches of clear opening once the leaf, the stops and the hardware are subtracted. A pocket opening has the same two numbers — the travel and the clear opening — and only one of them is the hole a person walks through.
If the pocket window is also a required emergency opening, this stops being a comfort question. We have written elsewhere about egress sizing rules; the narrower point here is that a required escape opening has to be openable by the person escaping, and a panel parked flush inside a wall with only a recessed pull is a different proposition from a casement with a lever handle you can see from across the room.
On an exterior window or door, the pocket is not a dry internal void. Every time the panel is retracted, the cavity is open to rain, wind, debris and sunlight. That single fact drags the whole of the opening's weatherproofing discipline into a hole that nobody scheduled.
Published installation guidance for pass-through and pocketing units states it as a procedure: treat the pocket cavity as a separate opening and repeat the flashing sequence inside it. Membrane goes on the floor, the walls and the ceiling of the cavity, with all seams lapped to direct water outward, and particular attention goes to the back wall of the pocket, where water will pool if no drainage is provided. The same guidance notes the practical habit of painting the inside of the pocket a dark colour, because the glass in the retracted panel makes the cavity visible whenever the window is in use.
That last detail deserves a sentence of its own. The inside of a pocket is a finished surface that you can see through glass and cannot reach. It needs a finish decision, on a drawing, before the wallboard goes on, and it is a decision that no window supplier will ever make for you.
Drainage is the part that gets missed. The track does not stop at the edge of the opening; it runs into the cavity, so the cavity floor has to drain as well, and the published warnings on these units are explicit that sill drains must not be blocked. There is also a published rating caveat worth knowing: a dealer summary of one manufacturer's tested data states a design pressure class achieved with a raised threshold, and states separately that flush-sill units are not rated. The flush threshold is the detail most people want in a pocket opening, and it is the one the published performance number does not cover.
Contractor guidance for large-format pocket systems names the thermal consequence without softening it: there is no insulation in the cavity, because the panel occupies the cavity when it is closed, and that makes the bay a thermal weak point in the wall assembly that has to be managed at the perimeter.
This is a wall problem wearing a window costume. The thermal value you were quoted belongs to the window unit, not to the wall it sits in, and no unit value has ever contained the uninsulated bay next to it. Our own published thermal values are system values rather than whole-window certification values — 1.3 for the AW90 passive window, 1.6 for the C120 and 1.7 for the C116 — and no sliding series in our catalogue publishes a thermal value at all, because a sliding sash has to keep a running gap and can only ever be a contact seal.
So the thermal story of a pocket window is: the best glazed unit you can specify, installed in a bay that has no insulation, cannot be ventilated, is open to the weather part of the time, and is next to a warm room the rest of the time. Condensation risk in that cavity should be a design line item, not a surprise.
The brief where this bites hardest is the all-glass room. A pocket is requested for 100 aluminum sun room wholesale projects more often than for any other application, because the whole point of a glass room is that the wall should disappear. But in a glass room there is no wall to donate. The pocket has to be borrowed from a solid return, a service core or an adjacent enclosed space, and if none of those exists, the honest answer is that the room cannot have a pocket at all — which is worth knowing at concept stage rather than at shop drawing stage.
One public guide for cavity sliders puts the weakness plainly: the moving parts live inside the wall, and unlike a hinged leaf you cannot simply lift it off. Good kits are designed for this — the leaf comes out by removing the head jamb or a removable access panel, and the trolleys unclip from the track — while site-fabricated pockets are not, which is why they age badly.
Manufacturers know this. The published pocket detail for one large-format system literally labels an interior pocket trim as removable for pocket access. Other public guidance recommends reserving an access opening of at least about 300 mm. And more than one installer source carries the same blunt warning in different words: once the wallboard is on, the assembly is not reversible, so the panel must be operated and tested thoroughly before the wall is closed.
The design deliverable follows directly, and it is small enough to fit in one line on a drawing: put the access panel on the same sheet as the pocket, give it a size, give it a finish, and check that the tile and joinery schedules do not cover it. A service call in year eight otherwise costs you a wall rather than a part.
Routine care is unglamorous and the same sources agree on it: keep the floor guide and track clean, because dust and hair are the usual cause of a sticky run; vacuum the cavity with a narrow tool; lubricate lightly and with a silicone-based product rather than an oil-based one, which attracts the grit that caused the problem. For the screen hardware that often accompanies these openings, published maintenance guidance adds two rules worth keeping — never let a retractable screen snap back into its cassette, and roll it away in high wind, because the mesh is an insect control device and not a structural one.
Automation is usually treated as an accessory that can be added in a later phase. On a pocket it cannot, and the published numbers say so precisely.
For one large-format pocketing system, the published instructions state that where automation is fitted the pocket width must be increased by twelve inches minimum to allow for motor installation, that a two-direction configuration needs a further six inches minimum on the side opposite the motor to accommodate the return pulley, and that installing the motor requires additional pocket depth.
Every one of those is a framing dimension. And there is a second-order problem sitting behind them: the motor needs a cable, and the pocket bay is the one place in the building where you are not allowed to run a cable. So the route has to be designed into the pocket structure before the wall is closed, and the decision has to be made before the wall exists.
A pocket is one of five common answers to the same brief. It is the best answer for a narrow set of conditions and the worst for most of them, and the differences are all structural rather than aesthetic.
| Configuration | What you see when it is open | What the structure has to carry | Servicing | What usually kills it |
|---|---|---|---|---|
| Nothing. The panel has left the opening | Header over opening plus cavity; built-out wall; relocation of bracing and services | Only through a designed access point | Wall thickness and header span, discovered late | |
| Stacking multi-track | A neat stack of panels at one or both ends | Header over the opening only; a standard wall is usually acceptable | Fully accessible | You never quite get the wall to vanish |
| Lift-and-slide | Panels stay in the opening | Opening only; frame depth drives the wall | Accessible | Sealing depends on sash weight; heavy glass changes the specification |
| Bi-folding | A folded bundle at one or both jambs | Opening only; standard framing | Accessible | More frame and less glass when closed; more panels to keep aligned |
| Swing or French | Two leaves standing open | Opening only | Fully accessible | Swing clearance eats the room; and for internal room dividers indoor folding doors are usually the cheaper way to the same result |
Two published comparisons help finish the picture. One contractor guide puts the maximum practical clear opening of a pocket system at about twenty feet against thirty feet and beyond for a surface-mounted multi-rail system, which is counter-intuitive until you remember that the pocket's cavity has to be framed, spanned and flashed. The same guide puts the additional rough-in effort for pocket framing at a range substantially above a surface-mounted slider of the same width, and adds that the gap around the panel in the cavity is a sound path, so a pocket is rarely the right choice where acoustic separation is the priority.
| Brief | Why it fails | Specify this instead |
|---|---|---|
| Retrofit into an existing finished wall | The cavity has to exist before the wall is closed; retrofitting means opening, framing and refinishing both faces | A surface-mounted multi-track slider |
| Load-bearing or braced external wall with no engineering | You are removing vertical structure and lengthening a span | Engineer it properly, or move the opening |
| A narrow return with services in it | The bay has to be completely clear of cables, pipework and ducts | Stacking system, or relocate the services first |
| A room where acoustic separation matters | The gap around the panel in the cavity is a sound path | A compression-sealed opening system |
| An accessible route with a fully flush panel | Fully open removes the grabbable hardware | Stop the panel proud, or specify push-to-open, and recheck clear width |
| A brief with nobody named to own the cavity | The part that fails is the part that has no line item, no drawing owner and no tester | Name the owner in writing before the framing drawing is issued |
A pocket cavity is fundamentally a water-management and structure problem. We publish no water tightness class, no air permeability class and no wind load class for any series in our catalogue, and pretending otherwise would be the single most damaging thing this page could do.
| What we publish | What it actually tells you | Why the gap exists, and what to ask for |
|---|---|---|
| 6060-T66 alloy on every series | A floor, not a differentiator — a row that is identical across the catalogue cannot decide anything | Ask for mill certificates tied to the heat number on the packaging |
| Profile thickness 1.8 to 2.5 mm, locally up to 3.0 mm | Two statements in one sentence: a floor, and a set of positions that are not printed | Ask where the thicker sections are, on a section drawing |
| PA66GF25 multi-chamber thermal break | Material and construction; the strip width is published for some series only | Ask for the strip width in millimetres, because it is a structural number as much as a thermal one |
| Frame wall width from 90 to 182 mm | The frame's depth into the wall. It is not a pocket cavity dimension | Nobody publishes the cavity dimension. Ask the pocket system supplier, and add their operating clearance to our frame depth |
| Sash width and frame width | Visible sightlines, not travel and not panel size | Ask for the panel width in millimetres, because that is the length of wall you are donating |
| Track height 53 and 31; 43 high and 8 low; 45 high ｄｒｏｐ with an optional 85 barrier | Two numbers usually means two tracks, not a tolerance. High versus low is a choice | The track does not stop at the opening. Ask how the cavity floor drains |
| Glass sash thickness 55, 54.5, 43 and 90 mm | A stack-up of the sash, not the glass thickness and not the rebate | Ask for the rebate dimension on a section. Sash thickness is a proxy, not a measurement |
| Load capacity 400, 220 and 200 kg | Per sash, not per unit, and dominated by the glass | Maximum panel size is published against an assumed glass build-up. Change the glass and the maximum changes |
| Thermal values 1.3, 1.6 and 1.7 | System values for three named series, not whole-window certification values, and not values for any sliding series | No sliding series publishes one, because a sliding sash must keep a running gap. Ask for a calculated value against your exact size and glass |
| 240 hour salt spray and 5,000 N shear | Published test results for one named series. They do not transfer to other series, and they are not air, water or wind classes | Ask which series a test belongs to, every time |
| 30,000 cycle hardware figure | A rig number for the hardware, not a service life for a window, and a handle cycle is not a loaded panel cycle | Ask what was cycled, and under what load |
| Air, water and wind performance classes | Nothing. We publish none of them for any series | A defect is a deviation from a declared threshold. With no declared class, "the pocket leaks" is not a measurement, it is an argument |
| Ask this | Why the answer changes your wall | An answer that sounds reasonable and tells you nothing |
|---|---|---|
| Overall unit width and panel width, both in millimetres | The panel width is the length of wall you are donating | "Standard size" |
| Both rough opening dimensions, with and without the pocket | Only one of them is the wall you build | "We will confirm on site" |
| Finished pocket depth, and whether it includes the sheathing | Your board and finish choices are already inside a published number | "It fits a standard wall" |
| Header span over opening plus cavity, and who sizes it | The most commonly under-specified member in the assembly | "The installer will handle it" |
| Access panel: where, how big, and what finish | Year eight either costs a part or costs a wall | "It will never need servicing" |
| Flashing sequence inside the cavity, and where the cavity drains | The cavity is outside the weather line whenever the panel is open | "We use standard waterproofing" |
| What happens about insulation in the pocket bay | The panel parks there, so insulation cannot | "The wall is insulated" |
| If motorised: extra pocket width, extra pocket depth, and the cable route | All three are framing dimensions decided before the wall exists | "We can add the motor later" |
| Clear opening, and whether the hardware is still usable fully open | The open position can remove the thing you close it with | "It opens fully" |
| Test-before-close protocol, and who signs it | Once the board is on, the assembly is not reversible | "The site team will check it" |
Technically yes, practically it is a small renovation rather than a window replacement. The cavity has to be built before the wall face is closed, so retrofitting means opening the wall, removing or splitting studs, fitting the pocket structure, setting the unit, testing it, and refinishing both faces. Public retrofit guidance treats this as a structural project first: establish whether the wall is load-bearing or part of a bracing system, whether the services can be relocated, and whether a header of the required span can be installed. If the answer to any of those is no, a surface-mounted multi-track slider gets you most of the openness for a fraction of the disruption.
There is no universal number, which is itself the answer. Published figures range from a North American heavy-duty frame documented for a minimum three and a half inch structural wall, through European systems publishing finished wall options of 100, 125 and 150 mm, to public guidance suggesting at least about 140 mm finished with not less than about 50 mm of internal clearance for a 40 to 50 mm leaf. Contractor guidance for large-format exterior pockets says a 2x4 is almost always insufficient and that panels wider than three feet need at least five inches of clear cavity. Because our published frame depths run from 90 mm to 182 mm, the honest method is: take our frame depth, add the operating clearance your pocket system supplier specifies, add your finishes, and compare the total to the wall you have drawn.
In terms of openable area, yes, and substantially: a standard two-panel slider can only ever clear about half its span, while a pocket clears essentially the whole of it. But ventilation rules in many jurisdictions measure openable area in a specific way, and some measure the area of the opening rather than the area of the sash. Check which one your code uses before you count the pocket as a compliance win, because the two are not interchangeable.
This is one of the least discussed and most frequently botched parts of the whole assembly. The screen has to cover the opening, but the panel leaves the opening, which means the screen cannot travel with it and cannot park with it. Published pocket systems list screens as an option, and retractable screens are the usual answer, but a retractable screen has a cassette, and a cassette needs a home on one jamb — a jamb that is already busy with the pocket mouth. Decide the screen at the same time as the pocket, not after, and expect it to change the jamb detail.
Almost always the least useful question, and the reason is structural rather than evasive. The unit has a track and a drainage path. The cavity has a flashing sequence, a membrane and an outlet. Neither party has designed the other's part, and the leak is frequently at the junction where the track enters the cavity. The useful question is not "whose fault" but "who is the first responder" — one named party who attends, diagnoses and makes good, and settles internally afterwards. Since we publish no water tightness class for any series, the water performance of your assembly has to be written down as a requirement by your design team before it can be claimed against anybody.
Yes, in a narrow set of circumstances: new build or a genuinely openable wall, an engineered header, a wall thick enough to hold the cavity and the finishes, a services-free bay, a designed access point, and a named owner for the cavity on the drawing. No, in most other circumstances, and the giveaway is always the same — if nobody has been named to own the hole the panel goes into, the pocket will fail, and it will fail after the wall is closed.
