Guides
Cabinet Construction & Joinery
The engineering underneath the finish. What a cabinet is made of, how its joints carry load, how solid wood moves against them, how the finish actually hardens, and the tolerances that decide whether any of it looks right once it is on a wall.
Quick Answer
What actually distinguishes a well-built cabinet?
A well-built cabinet is distinguished by four decisions made before anything is visible: a substrate stiff enough for the span and load it carries, joinery that puts adhesive on long grain rather than end grain, a design that lets solid wood move seasonally instead of restraining it, and a finish that hardens by chemical crosslinking rather than by solvent evaporation. Hardware and installation tolerances then determine whether that construction stays true over decades.
Each of those is measurable against a published standard. ANSI/KCMA A161.1 loads shelves at 15 pounds per square foot for seven days. ANSI/AWI 0620 holds a Premium-grade installation to a 0.016-inch gap at wood-to-wood field joints. The Forest Products Laboratory puts tangential shrinkage in white oak at roughly twice its radial shrinkage, which is why a quartersawn panel moves about half as much as a flat-sawn one. Appearance is the last thing decided and the first thing noticed, but it is not what determines whether a cabinet is still working in thirty years.
Executive Summary
The Six Systems Inside a Cabinet
Substrate, joinery, movement, finish, tolerance, and hardware are not six separate subjects. They are six views of the same object, and a decision in any one of them constrains the other five.
A kitchen cabinet is a small engineered structure that happens to be beautiful. It carries static loads on its shelves, live loads every time a drawer is thrown shut, torsional loads when it is shimmed on an uneven floor, and a cantilever load through its mounting rail if it hangs on a wall. It is built from a hygroscopic material that changes dimension with the weather, joined with an adhesive that is extraordinarily strong on some surfaces and nearly useless on others, coated with a film that may or may not be undergoing an irreversible chemical reaction, and installed into a building framed to tolerances an order of magnitude coarser than its own. Every one of those facts has a standard written about it, and most of what separates a cabinet that is right in thirty years from one that is tired in five is decided in that engineering rather than in the door style.
The organizing insight is that these systems are coupled. Substrate choice sets the screw-holding capacity available to the hardware, which sets the load a drawer can carry, which sets the joinery the drawer box needs. Species and cut set how much a solid panel will move, which sets whether it can be captured or must float, which sets the door construction, which sets how much the finish film will be asked to flex across a joint line. Installation tolerance is the last constraint in the chain and the least forgiving, because it is the only one measured against a building nobody in the shop controls.
Where this guide gives a number, it is attributable. The dimensional data come from the USDA Forest Products Laboratory Wood Handbook, publication FPL-GTR-190, which remains the reference work for the physical and mechanical properties of wood. The construction, finishing, and installation criteria come from the Architectural Woodwork Standards and ANSI/AWI 0620. The cabinet performance loads and cycle counts come from ANSI/KCMA A161.1 and ANSI/BHMA A156.9. Where a figure could not be attributed to a source of that kind, the claim is written qualitatively instead — a real shop works in ranges and directions, and false precision is worse than none.
Each of the six sections below summarizes what matters and links to a longer article devoted to it. Read straight through for the shape of the subject, or go directly to whichever system your project is actually constrained by.
Key Takeaways
- Substrate is the first decision and the most consequential: it sets screw retention, shelf deflection, weight on the wall, and how the cabinet responds to a leak. ANSI/KCMA A161.1 loads shelves and bottoms at 15 pounds per square foot for seven days.
- Glue is weak on end grain. The Forest Products Laboratory reports that end-grain butt joints reach only about 25 percent of the tensile strength of wood parallel to the grain, which is the entire reason traditional joinery exists.
- Wood barely moves along the grain — roughly 0.1 to 0.2 percent longitudinally green to oven-dry — and moves substantially across it, with tangential shrinkage about twice radial. That asymmetry is why frames hold shape and panels must float.
- Cut matters as much as species. A 20-inch flat-sawn white oak panel changes about 0.29 inch in width over a 4-point moisture-content swing; the same panel quartersawn changes about 0.14 inch.
- Finish durability follows cure chemistry, not appearance. The Architectural Woodwork Standards rate nitrocellulose lacquer 1 for solvent resistance and conversion varnish 5, while reversing the ranking on repairability.
- Installation tolerance is measured against the building, not the shop: ANSI/AWI 0620 requires product installed plumb and level within an eighth of an inch over 96 inches, and holds Premium-grade wood-to-wood joint gaps to 0.016 inch.
The Box: Substrate Choice Drives Everything Downstream
Before a single joint is cut, the panel you choose has already decided how the cabinet holds a screw, how it survives a dishwasher leak, how much a full run weighs on the wall, and how flat a painted door will look in raking light.
A cabinet box is a structural assembly disguised as a container. It carries a static load on its shelves and bottom, a live load every time a drawer is pulled out and dropped back, a torsional load whenever it is racked during delivery or shimmed on an uneven floor, and — in the case of a wall cabinet — a cantilever load transferred through a mounting rail into whatever the wall is made of. Substrate is what determines whether all of that is absorbed or accumulated.
The standards put numbers on the loads. ANSI/KCMA A161.1, the performance standard behind the certification seal on a great deal of North American casework, loads every shelf and cabinet bottom at 15 pounds per square foot and leaves the weight in place for seven days, looking for excessive deflection or any sign of joint separation. It loads a wall cabinet toward 600 pounds total. It applies 250 pounds against the inside stiles of a base cabinet front with drawer rails present, or 200 pounds without them, to see whether the front joints open. And it drops a three-pound steel ball from six inches onto the assembly to check that nothing has been held together by finish and optimism.
The Architectural Woodwork Standards go at the same question from the deflection side. Their shelf load test uniformly loads a shelf toward 200 pounds and allows a maximum deflection of one one-hundred-eightieth of the span, and never more than a quarter of an inch. That single criterion is the most useful shelf rule a homeowner can carry around, because it explains why a 36-inch shelf that would be fine in three-quarter-inch plywood sags visibly in half-inch particleboard: the allowance is a function of span, and the stiffness available to meet it is a function of material and thickness.
Moisture is the second axis, and it separates the three substrates more sharply than stiffness does. The Architectural Woodwork Standards include a submersion test for casework specified for laboratory use — the cabinet stands in two inches of water for four hours, and the base material is allowed to gain no more than 4 percent in thickness. Four percent is a small number, and it is the reason a toe kick detail, an edge-sealing routine, and a shut-off valve you can actually reach are construction decisions rather than afterthoughts. The Forest Products Laboratory notes separately that hot-pressed panel products such as particleboard and hardboard do not reach the same equilibrium moisture content as solid lumber and recommends conditioning them at 30 to 40 percent relative humidity for interior service.
| Substrate | Dimensional stability | Screw retention | Weight | Moisture behaviour | Relative cost |
|---|---|---|---|---|---|
| Veneer-core hardwood plywood | High. Cross-banded plies restrain movement in both directions, so a panel stays close to the dimension it was cut at. | Best of the three. A screw driven into a plywood edge engages multiple glue lines and alternating grain directions. | Lightest. Matters most on tall or wide wall cabinets, where dead load is carried by the mounting rail and the wall behind it. | Swells and can delaminate at the edges under sustained wetting, but generally recovers its shape after a short exposure. | Highest material cost, and the cost climbs steeply with veneer grade and core quality. |
| Industrial particleboard | Very stable in the plane of the panel. It has no grain direction, so it has no strong or weak axis to plan around. | Weakest, especially in the panel edge, where a screw is pulling apart a matrix of chips and resin rather than continuous fibre. | Heavy. A run of tall pantry cabinets built from it is meaningfully harder to carry, hang, and shim. | The most vulnerable. Once the resin bond around the chips is broken by standing water, the swelling is permanent. | Lowest. It is the reason a great deal of production casework is built from it. |
| Medium-density fibreboard (MDF) | Very stable in the plane of the panel, and flatter than plywood out of the press — which is why it is the default core for painted doors. | Better than particleboard, still short of plywood. Face screws hold well; edge screws want a pilot hole and a coarse thread. | Heaviest of the three. Weight is the practical limit on how large an MDF door can be before the hinges start to complain. | Swells rapidly at cut edges if water reaches raw fibre, and does not recover. Sealing every cut edge is not optional. | Between the other two, and cheap relative to what it does for a painted finish. |
Go Deeper
Cabinet Box Construction →Inset vs Frameless Cabinetry: Which Is Right for Your Luxury Kitchen? [2026]
4 min read
Read the Article →Joinery: The Load Paths in a Cabinet, and Which Joints Carry Them
Joinery is not decoration and it is not nostalgia. It is the answer to a specific structural question: how do you get a load across the place where two pieces of wood meet, given that wood glue is very strong on some surfaces and nearly useless on others?
The Forest Products Laboratory is blunt about the limiting case. End-grain surfaces are many times more porous than radial or tangential surfaces, and adhesive runs so freely into the open cell lumina along the grain that over-penetration is normal rather than exceptional. That over-penetration, the Wood Handbook says, is the primary reason it is so difficult to form strong load-bearing bonds in butt joints. Even with special techniques, an end-grain butt joint reaches only about 25 percent of the tensile strength of the wood parallel to the grain.
Everything traditional joinery does follows from that one fact. If a glued end-grain surface gives you a quarter of the available strength, then the joint has to be reshaped until the glue is working on long grain instead. A mortise and tenon does exactly that: the tenon cheeks present long grain to long grain inside the mortise, and the shoulders resist racking mechanically. The Wood Handbook states the principle directly for end-to-edge joints — it is difficult to design a plain end-to-edge-grain joint capable of carrying appreciable loading, so these joints must be designed with interlocking surfaces such that the edge grain of one piece bonds to the edge grain of the other. That sentence is a mortise-and-tenon joint described in structural language.
The same logic explains why a dovetail is the drawer joint that survives. A drawer box fails in one direction above all others: the front is pulled away from the sides, thousands of times, often with the drawer overloaded. A rabbet resists that with glue and fasteners alone. A dovetail resists it with geometry — the pins and tails physically cannot pass one another in the direction of pull, so the adhesive is asked to hold the joint together rather than to carry the load. Where end-grain joining is genuinely unavoidable, the Wood Handbook shows the alternative: a well-made scarf, finger, or lap joint in end grain can reach as much as 90 percent of the tensile strength of clear wood, because the cut increases the bonded surface area. The reason a scarf slope of one in twelve is specified for the strongest joints is that wood is roughly ten times stronger in tension than in shear, so the glue area has to exceed the cross-section by a comparable factor.
One subtlety gets missed even by careful shops. The Wood Handbook warns that adhesive bonds are damaged when adjacent pieces have different swelling or shrinkage coefficients — different species, different grain type, radial grain bonded to tangential, end grain bonded to cross grain — and that the stress is worse still when only one part of an assembly changes moisture content. Its remedy is to bond pieces with compatible grain directions and low shrinkage coefficients, and to bond at the moisture content the assembly will actually see in service. A joint can be cut perfectly and still be engineered to fail if those two conditions are ignored.
Go Deeper
Joinery →Wood Movement: Why Solid Panels Float and Frames Do Not
A frame-and-panel door is the oldest running answer to a problem that never goes away: wood changes width with the seasons and barely changes length at all, so the parts that move must be separated from the parts that hold the shape.
The asymmetry is the whole story. According to the Wood Handbook, longitudinal shrinkage — along the grain, green to oven-dry — runs between roughly 0.1 and 0.2 percent for most species. Across the grain it is one to two orders of magnitude larger, and tangential shrinkage is about twice radial shrinkage. For white oak the Forest Products Laboratory puts radial shrinkage at 5.6 percent and tangential at 10.5 percent from green to oven-dry; sugar maple runs 4.8 and 9.9 percent; black walnut 5.5 and 7.8; black cherry 3.7 and 7.1. A door stile, cut with its length along the grain, is dimensionally almost inert. A wide solid panel is not.
Nobody builds cabinets from green wood, so the more useful figures are the dimensional change coefficients the Forest Products Laboratory publishes for the 6 to 14 percent moisture content range that furniture actually lives in. They express change per 1 percent change in moisture content, referenced to the dimension at 10 percent. For commercial white oak the tangential coefficient is 0.00365 and the radial coefficient is 0.00180. Work an example: a 20-inch-wide flat-sawn white oak panel moving from 6 to 10 percent moisture content changes roughly 20 by 0.00365 by 4, which is about 0.29 inch — call it five-sixteenths of an inch of seasonal width change in a single panel. Cut the same panel quartersawn and the radial coefficient applies instead: about 0.14 inch. That is the entire practical argument for quartersawn stock, in one line of arithmetic.
Those moisture swings are ordinary, not extreme. At 70 degrees Fahrenheit the Wood Handbook puts the equilibrium moisture content of wood at about 6.2 percent in 30 percent relative humidity and about 9.2 percent in 50 percent — a three-point swing inside a comfort range no one would think twice about. The Forest Products Laboratory recommends installing interior woodwork, flooring, and furniture at an average of 8 percent moisture content across most of the United States, 6 percent in the dry southwestern area, and 11 percent in the damp warm coastal area, with individual pieces held to 6 to 10, 4 to 9, and 8 to 13 percent respectively. Californian conditions cover a wide band of that range, and the outdoor data show the amplitude involved: the Wood Handbook lists outside conditions in Fresno at an equilibrium moisture content of 16.4 percent in January and 7.8 percent in July.
So the panel floats. It sits in a groove, dry, with clearance at its edges and — if there is glue anywhere — a spot at the centre of the top and bottom rails only, so that expansion is shared evenly to both sides rather than concentrated at one joint. The Forest Products Laboratory gives the instruction almost verbatim for interior finish work: design and install large solid pieces, such as wood paneling, so that the panels are free to move across the grain, and prefer narrow widths. A frame-and-panel door does both. It is not a style choice that happens to be structural; it is a structural solution that happens to have become a style.
| Species | Radial shrinkage (green to oven-dry) | Tangential shrinkage (green to oven-dry) | Dimensional change coefficient per 1% MC | What it means at the bench |
|---|---|---|---|---|
| White oak | 5.6% | 10.5% | 0.00365 tangential / 0.00180 radial | Nearly a 2:1 tangential-to-radial ratio. Quartersawing it roughly halves the seasonal width change, which is most of why quartersawn white oak has the reputation it does. |
| Sugar (hard) maple | 4.8% | 9.9% | 0.00353 tangential / 0.00165 radial | Moves almost as much as white oak. A wide flat-sawn maple panel needs the same allowance, even though maple reads as the calmer material. |
| Black walnut | 5.5% | 7.8% | 0.00274 tangential / 0.00190 radial | The flattest tangential-to-radial ratio of the four. Walnut distorts less in section, which is part of why it behaves so well in wide solid components. |
| Black cherry | 3.7% | 7.1% | 0.00248 tangential / 0.00126 radial | The least movement of the four in absolute terms. Cherry is forgiving dimensionally; its real design constraint is that it darkens with light exposure. |
Go Deeper
Wood Movement →Finish Systems: Cure Chemistry, Not Appearance, Determines Durability
Two finishes can be indistinguishable on a sample door and behave completely differently five years into a working kitchen. What separates them is not sheen or colour but whether the film hardened by losing solvent or by crosslinking into a new material.
The Architectural Woodwork Standards enumerate thirteen finishing systems and require that a specification name both the number and the description — System 5, Varnish, Conversion, and so on — precisely because the trade names are unreliable and the chemistry is not. The standards then rate every system across general durability, moisture resistance, solvent resistance, stain resistance, repairability, and finish flexibility on a five-point scale, with 5 excellent and 1 poor, evaluated against ASTM D1308 for chemical resistance, ASTM D4060 for abrasion, ASTM D1211 for cold-check resistance, and ASTM D3359 for cross-hatch adhesion.
The pattern in those ratings is unambiguous. Nitrocellulose lacquer, System 1, rates 2 for general durability and 1 for solvent resistance — it never stops being soluble in its own thinner, which is exactly why it is also rated 5 for repairability. Conversion varnish, System 5, rates 4 for general durability and 5 for solvent, stain, and heat resistance. Catalyzed polyurethane, System 11, rates 5 for general durability and 5 for solvent resistance. Synthetic penetrating oil, System 6, rates 1 for durability, moisture resistance, and solvent resistance, and 5 for repairability and finish flexibility. Nothing in that table is a compromise-free winner; the specification question is which failure mode you would rather live with.
Once a catalyzed finish is mixed, it is on a clock, and that clock is the practical constraint the finishing schedule is built around. M.L. Campbell states, for one of its conversion varnishes, that crosslinking begins with the addition of 12.8 ounces of catalyst per gallon — a 10 to 1 ratio — and that pot life after catalyzation is 12 hours. The same data sheet gives 15 to 25 minutes dry to touch, 30 to 45 minutes to sanding, 5 to 8 hours to stacking, and a target final dry film of 4 to 5 mils built from two coats at 4 to 5 mils wet. Those are not suggestions. Exceed the film build and a conversion varnish is more prone to cracking; rush the stacking window and the damage is invisible until the doors are hung.
Ambient conditions belong in the same conversation. The Architectural Woodwork Standards record their finish performance baseline as 45 to 55 percent relative humidity at 70 to 80 degrees Fahrenheit, and note that water-borne coatings must be cured in a dehumidified atmosphere assisted by air movement. This is the honest reason that spray-finishing cabinetry on site produces a different result from finishing it in a conditioned booth, even with identical material in the gun. Manufacturers including M.L. Campbell and Milesi publish detailed technical data sheets for exactly this reason, and the useful habit for a homeowner or specifier is simply to ask which AWS system number is being quoted and read what the standard says about it.
| AWS system | How the film hardens | General durability | Solvent resistance | Repairability |
|---|---|---|---|---|
| Nitrocellulose lacquer (System 1) | Solvent evaporation only — the film redissolves in its own thinner | 2 | 1 | 5 |
| Pre-catalyzed lacquer (System 2) | Partial crosslink, catalyst added at the factory | 2 | 2 | 4 |
| Conversion varnish (System 5) | Acid-catalyzed crosslink, catalyst added in the shop | 4 | 5 | 3 |
| Catalyzed polyurethane (System 11) | Two-component crosslink with an isocyanate or equivalent hardener | 5 | 5 | 2 |
| Synthetic penetrating oil (System 6) | Penetrates rather than films; little or no surface build | 1 | 1 | 5 |
Go Deeper
Finish Systems →Tolerances: Why a Shop-Perfect Box Still Needs Scribing on Site
Cabinetry is built to thousandths and installed into buildings that are framed to eighths. Scribing is the discipline that reconciles the two, and it is where a great deal of the perceived quality of a finished kitchen is actually won or lost.
ANSI/AWI 0620, the finish carpentry and installation standard, defines three aesthetic grades — Premium, Custom, and Economy — and specifies that when a project references the standard without naming a grade, Custom grade applies by default. The numbers separating them are small and entirely visible. At wood-to-wood field joints, gap width may not exceed 0.016 inch at Premium grade, 0.031 inch at Custom, or 0.047 inch at Economy. Flushness variation between wood surfaces at flat field joints carries the same three limits. The standard states them in millimetres alongside the inch values — 0.4, 0.8, and 1.2 millimetres respectively — which is a fair indication of how tight the Premium figure is.
The door and drawer front tolerances are the ones people notice from across a room without knowing what they are noticing. Edge alignment of doors, drawer fronts, and false fronts, in both the vertical and horizontal plane, may not exceed 0.031 inch at Premium, 0.047 at Custom, or 0.063 at Economy. Fronts must sit on the same plane as one another within 0.031, 0.063, or 0.125 inch by grade. In reveal overlay frameless construction the nominal reveal is 3.2 millimetres — an eighth of an inch — with maximum uniform variance of plus or minus 0.031 inch at Premium and plus or minus 0.063 at Custom. Inset face frame work is held to the same eighth-inch nominal and the same variances, which is a large part of why inset costs what it costs. Doors themselves must stay within 0.031 inch of flat per lineal 12 inches at Premium grade and never exceed a quarter inch of warp in any single door regardless of grade.
Then the standard turns to the building. Product must be installed plumb and level within 3.2 millimetres — an eighth of an inch — over 2,438 millimetres, or 96 inches. A wall that is out of plumb by half an inch over eight feet, which is entirely ordinary in a house of any age, is off by four times the amount the installation standard allows the cabinetry itself. There is no version of that job where the boxes simply go against the wall. ANSI/AWI 0620 requires that casework be scribed to finished walls or ceilings using integral scribes, scribe fillers, or scribe moulding wherever voids occur, and it is equally clear about the limit of the installer’s obligation: it is not the installer’s responsibility to scribe or fit to walls, ceilings, floors, or openings whose plumb, level, flat, straight, or square values exceed the tolerances specified in the contract documents.
That last provision is worth reading twice, because it is the clause that decides who pays when a wall is genuinely out of true. It is also the reason a careful cabinetmaker measures the room before building rather than after, and specifies filler and scribe allowances at the drawing stage. Scribing is not a repair for a mistake. It is a planned dimension, allocated in the shop, consumed on site.
Go Deeper
Installation Tolerances →Hardware: Load Rating and Cycle Count Are the Specs That Matter
Hinges and runners are the only parts of a cabinet that are asked to perform a mechanical duty cycle for decades. They are also the parts most often chosen by feel in a showroom rather than by the two numbers that predict how they will age.
ANSI/BHMA A156.9 is the North American standard for cabinet hardware, covering hinges, pulls, catches, shelf rests, drawer slides, and the rest, with operational, cyclical, strength, and finish tests behind each. Its part numbering encodes the grade in the last digit — in the designation B01011, the B is the product class, the first 0 the predominant base material, the 1 identifies a hinge, the 01 the semi-concealed function, and the final 1 the Grade 1 performance level. That final digit is the specification, and it is the one number that is almost never quoted in a kitchen showroom.
Cycle counts are where the standards diverge in a way worth understanding. ANSI/KCMA A161.1 swings a door through a full 90 degrees for 25,000 cycles and cycles a drawer loaded at 15 pounds per square foot the same 25,000 times, with both required to remain operable afterward. The Architectural Woodwork Standards call for a door hinge test performed in conformance with the ANSI A156.9 Grade 1 procedure, running 100,000 cycles at no more than 15 cycles per minute. Both are legitimate; they are simply aimed at different service expectations, and a manufacturer quoting one should not be read as having met the other.
Static load ratings are the second number, and reputable manufacturers publish them against a named standard rather than as marketing figures. Blum, for example, lists its TANDEM runner at a 100-pound static load rating per ANSI/BHMA A156.9 testing standards, and its MOVENTO runner at 125 pounds in nominal lengths of 12 to 21 inches and 170 pounds in nominal lengths of 18 to 30 inches. Grass, Salice, and Häfele likewise publish technical documentation for their hinges and runner systems. The specification worth reading is always the pair: what load, measured against which standard.
The casework tests round out the picture, because hardware fails in company with the wood it is screwed to. The Architectural Woodwork Standards hang 200 pounds on an open door at a point 12 inches out from the hinge centreline and swing it through the full arc of the hinge, looking for distortion that would bind the door. They hang 150 pounds from a drawer head opened to 13 inches of travel for five minutes. They apply 50 pounds perpendicular to each pull. Almost every one of those tests is really a test of what the fastener is anchored into — which is the point at which a discussion about hinges turns back into a discussion about substrate, and the guide comes full circle.
Go Deeper
Hardware Engineering →Frequently Asked Questions
- What actually makes one cabinet better built than another?
- Four things decided before any of it is visible: a substrate stiff enough for the span and the load it will carry, joinery that puts adhesive on long grain rather than end grain, a design that allows solid wood to move seasonally instead of restraining it, and a finish that hardens by chemical crosslinking rather than solvent evaporation. Hardware selection and installation tolerances then determine whether that construction stays true over time.
- Is plywood always better than particleboard or MDF for cabinet boxes?
- Not always, and the honest answer depends on which property matters most in a given part of the cabinet. Veneer-core plywood holds screws best, weighs least, and recovers better from a short wetting. MDF is flatter out of the press and is the standard core for painted doors for exactly that reason. Industrial particleboard is very stable in plane and much less expensive, but it is the weakest of the three in edge screw retention and the least forgiving of standing water. A well-engineered cabinet often uses more than one of them, matched to the job each part does.
- How much weight is a cabinet shelf supposed to hold?
- ANSI/KCMA A161.1 loads shelves and cabinet bottoms at 15 pounds per square foot and maintains that load for seven days, looking for excessive deflection or joint separation. The Architectural Woodwork Standards approach it from the deflection side: uniformly load a shelf toward 200 pounds and allow a maximum deflection of one one-hundred-eightieth of the span, not exceeding a quarter of an inch. Because the allowance is a function of span, a longer shelf needs a thicker or stiffer material to meet the same standard.
- Why is an end-grain butt joint considered weak?
- Because the adhesive cannot stay where it is needed. The Forest Products Laboratory notes that end-grain surfaces are many times more porous than radial or tangential surfaces, so glue over-penetrates into the open cell lumina instead of forming a bond line. Even with special techniques, an end-grain butt joint reaches only about 25 percent of the tensile strength of the wood parallel to the grain. Traditional joinery exists to reshape that connection until the glue is working on long grain and the geometry is carrying the load.
- Why are drawers dovetailed instead of simply glued and screwed?
- A drawer box fails predominantly in one direction: the front pulled away from the sides, repeatedly, often under load. A dovetail resists that mechanically, because the pins and tails cannot pass one another in the direction of pull, so the adhesive holds the joint closed rather than carrying the working load. It is the same principle the Forest Products Laboratory describes for end-to-edge-grain joints, which it says must be designed with interlocking surfaces so that edge grain bonds to edge grain.
- How much does a solid wood cabinet door panel actually move?
- Enough to matter. Using the Forest Products Laboratory dimensional change coefficients for the 6 to 14 percent moisture content range, a 20-inch-wide flat-sawn white oak panel moving from 6 to 10 percent moisture content changes about 0.29 inch in width — roughly five-sixteenths of an inch. The same panel cut quartersawn changes about 0.14 inch, because the radial coefficient applies instead of the tangential one. That difference is the entire practical case for quartersawn stock in wide components.
- What indoor humidity does wood cabinetry prefer?
- A moderate and, above all, stable range. At 70 degrees Fahrenheit the Wood Handbook puts wood at about 6.2 percent moisture content in 30 percent relative humidity and about 9.2 percent in 50 percent. The Forest Products Laboratory recommends installing interior woodwork and furniture at an average of 8 percent moisture content across most of the United States, 6 percent in the dry southwest, and 11 percent in damp warm coastal areas. The absolute number matters less than avoiding rapid swings between the extremes.
- What is the difference between conversion varnish and lacquer?
- How the film hardens. A nitrocellulose lacquer dries by solvent evaporation and remains soluble in its own thinner, which is why the Architectural Woodwork Standards rate it 1 for solvent resistance and 5 for repairability. A conversion varnish is acid-catalyzed and crosslinks into a new material, which is why the same standards rate it 4 for general durability and 5 for solvent, stain, and heat resistance. Catalyzed polyurethane rates 5 for general durability and 5 for solvent resistance. Greater durability generally costs repairability — that trade is the actual specification decision.
- Why do cabinets need to be scribed to the wall?
- Because buildings are not built to cabinet tolerances. ANSI/AWI 0620 requires product to be installed plumb and level within an eighth of an inch over 96 inches, and holds Premium-grade wood-to-wood field joints to a gap of no more than 0.016 inch. A wall out of plumb by half an inch over eight feet — entirely common — is four times the installation allowance on its own. Scribing to the finished wall with integral scribes, scribe fillers, or scribe moulding is how the standard resolves that, and it is planned into the drawings rather than improvised on site.
- Which hardware specifications should I actually ask about?
- Two: the rated load and the standard the rating was measured against, and the cycle count. Blum publishes its TANDEM runner at a 100-pound static load rating per ANSI/BHMA A156.9, and MOVENTO at 125 pounds in 12 to 21 inch nominal lengths and 170 pounds in 18 to 30 inch lengths. On cycles, ANSI/KCMA A161.1 tests doors and drawers to 25,000 cycles while the Architectural Woodwork Standards call for 100,000 cycles under the ANSI A156.9 Grade 1 procedure. A number quoted without its standard tells you very little.
Explore
Cabinet Construction, in Depth
Six technical articles behind this guide, the journal pieces that support them, and the standards bodies whose published documents the figures on this page are drawn from.
The Six Systems
Related Reading
Continue Learning
Planning a kitchen and want to talk it through with the person who will build it?
Request a Design Consultation