Density is not hardness, and the confusion is expensive
Two different numbers decide a seat pad and buyers routinely treat them as one. Density is mass per cubic metre: it says how much polymer is packed into the pad, and therefore how much material has to share the load. Hardness is measured as indentation load deflection, the force in newtons needed to compress a 50 mm thick sample by 25 percent, and it says how the pad feels in the first ten seconds on a showroom floor. A 35 kg/m3 foam can be made softer than a 25 kg/m3 foam, because hardness is set by the formulation, not by the mass.
That distinction matters because the two numbers fail on different timescales. Hardness is what a buyer judges when they press a sample chair with their hand and decide it feels right. Density is what decides how the pad looks after 24 months of service, and by then the container has been signed for and the complaint lands on the operator instead of the factory.
The arithmetic is simple enough to check. A 60 mm seat pad on a 450 mm square seat is 0.0122 cubic metres of foam. At 25 kg/m3 it weighs 305 g, at 35 kg/m3 it weighs 427 g, at 45 kg/m3 it weighs 549 g. The difference between the cheapest grade and the best is about 244 g of material per chair, on a chair that will take roughly 2,000 sittings a year at 8 turns a day over 250 trading days. That is the whole argument in one line: 244 g against 6,000 load cycles.
The three grades in service
25 kg/m3 is the grade most quotations default to when the buyer does not specify. It is legitimate foam and it is used in domestic dining chairs without trouble, because a domestic chair sees perhaps 700 sittings a year with one or two people in the house. In a restaurant it loses 20 to 25 percent of its 25 percent indentation load deflection by month 12 and 35 to 40 percent by month 24. What the guest feels is a dip of 8 to 12 mm in the middle of the seat and a hard stop at the front edge where the pad has been crushed against the frame.
35 kg/m3 is the grade we quote for dining rooms with normal turnover. At the same duty it loses 8 to 12 percent by month 24 and holds a serviceable shape to year four or five, provided the pad is at least 50 mm thick and sits on a deck that spreads the load rather than on two webs 200 mm apart. This is the sensible default for commercial restaurant seating in a full-service room.
45 kg/m3, usually as a high-resilience formulation, is what belongs in seating that is also a lounging seat: banquette backs, booth backs, deep armchairs where a guest sits for 70 minutes rather than 45. High-resilience foam at this density recovers about 95 percent of its height within one second of release, against 80 to 85 percent for a standard 25 kg/m3 block. It costs roughly 35 to 50 percent more per pad and it is wasted on a chair that is turned every 40 minutes.
Why the seat pad fails before the backrest
A seated adult puts about 100 percent of body weight through the seat and 15 to 20 percent through the backrest at a 15 degree rake. The seat therefore compresses roughly five times as much per sitting, and it does so against a rigid plane: plywood deck, webbing or a steel frame, depending on the model. The bottom 10 mm of seat foam is crushed against that plane on every single sitting, which is why the collapse always starts at the bottom of the pad and works upward.
The backrest has an easier life. In most dining seating for restaurants it is 40 mm thick against the seat's 60 mm, sits on a curved shell with some give, and carries a fraction of the load. A dining room that reports sagging backs is usually reporting a frame problem, not a foam problem; a dining room that reports hard seats is reporting foam.
The third factor is cleaning. Seat pads take the spills, the mopping water at the frame line and the daily wipe with a quaternary disinfectant. Water at the deck line softens the foam surface, and repeated wetting accelerates the cell wall fatigue that density was supposed to resist. A 25 kg/m3 pad that stays dry lasts noticeably longer than the same pad in a room that is mopped twice a day.
Rebound time and the 40 percent line
Laboratory measurement uses a compression set test: hold the pad at 25 percent deflection for 22 hours at 23 degrees Celsius, release for 30 minutes, then measure the height lost. High-resilience foam comes back within 3 to 5 percent. Standard foam at 8 to 10 percent is acceptable for contract use. Anything over 12 percent on a new sample means the block was over-blown to hit a density it cannot hold.
In the field you can approximate the same thing with a thumb. Press 10 mm into the pad, release, and time the recovery. Under one second is healthy. Two to three seconds, or a visible thumbprint left after five seconds, means the cell structure has begun to fatigue and the pad is on its way out. Do the test in the middle of the seat, 100 mm back from the front edge, not at the corner where the frame is close to the surface.
Replace at 35 to 40 percent loss of original indentation load deflection. Past that point the guest is no longer sitting on foam, they are sitting on the deck through a thin layer of foam, and no amount of reupholstery will fix it. A simple field version: sit for two minutes, stand up, and look at the seat. A dent that is still visible after 30 seconds is a pad that needs replacing, not rotating.
Thickness, seat height and the 20 mm nobody plans for
A 60 mm pad at 35 kg/m3 compresses 8 to 10 mm under a 70 kg guest, so a chair drawn at 450 mm seat height sits at about 440 mm in use. That is fine, and it is what the 750 mm table height was designed against. Change the pad to 50 mm at 25 kg/m3 and the same chair compresses 15 to 18 mm, so the effective seat height drops to 432 mm and the table suddenly reads 18 mm high to every guest who sits down.
The same drift happens over the life of the pad. A 25 kg/m3 pad that loses 12 mm of thickness turns a 450 mm chair into a 438 mm chair by year two, and a room where half the chairs have been replaced and half have not will show a visible 10 to 15 mm step between tables. Operators notice this as guests choosing certain tables, and it is almost always foam, not the frame.
The rule we work to: foam under 45 mm in a commercial seat is a maintenance item rather than a cushion. Dining chairs take 50 to 70 mm of pad, banquettes and booth seats take 80 to 100 mm, and the drawing should state both the thickness and the compressed height so the upholsterer and the frame maker are working from the same number.
What holds the foam matters as much as the foam
A 35 kg/m3 pad on elastic webbing at 50 mm feels softer and lasts longer than a 45 kg/m3 pad glued to a flat plywood deck, because the webbing lets the whole pad deflect instead of concentrating load at the frame edge. A plywood deck with a 10 mm rebate is the most durable arrangement and the least forgiving; a sinuous spring deck sits between the two and is the usual choice for dining chairs in the 60 to 90 USD band.
Wrapping is the other half. A 20 mm polyester fibre wrap over the foam adds 3 to 5 mm of surface softness without changing support, keeps the fabric from abrading against the foam at the front edge, and hides the sharp line where the pad meets the frame. On a dining chair with leather seat the wrap is more important than on fabric, because leather has no loft of its own and every irregularity in the foam reads through it within a month.
Fire performance is a specification item, not a detail. Contract seating sold into most markets is expected to meet CAL TB 117-2013 or BS 5852 crib 5, and the foam grade that meets it is not always the grade that was quoted. Ask for the foam supplier's test report by grade, not a generic certificate for the chair, because the certificate usually covers the cover fabric and says nothing about what is inside it.
The four lines to put on the drawing
Write the pad, not the chair. Seat foam, 35 kg/m3 plus or minus 2, indentation load deflection 130 to 160 N at 25 percent, thickness 60 mm, high-resilience grade, wrapped in 20 mm polyester fibre, fire performance to CAL TB 117-2013. Back foam, 30 kg/m3, 40 mm. That is four lines, it fits under the upholstery note, and it removes the single largest variable between two quotations that look identical on paper.
Then tie it to a model you can actually inspect. The seat pad on a JP-C022 metal dining chair and the pad on a timber frame are built the same way but loaded differently, because the steel frame usually has a tighter deck and less give. Order a cut sample of the foam itself, 100 by 100 by 50 mm, and weigh it: 35 kg/m3 should read 175 g plus or minus 5 percent. A sample that weighs 130 g is 26 kg/m3 whatever the label says.
We run that weigh-check on incoming foam as part of restaurant furniture quality control, together with a 22 hour compression set on the first pad of each batch, because density is the one property a supplier can quietly change between the sample and the production run. On a 200 chair order the difference is around 50 kg of foam, which is exactly the kind of saving that gets taken without telling anyone.
The cost is small against the consequence. Moving from 25 to 35 kg/m3 adds roughly 2.5 to 4 USD per chair at our pricing, which across three years and 6,000 sittings works out at well under a tenth of a cent per cover. Replacing 200 seat pads at year two costs the pads, the labour and the downtime, and it happens in the busiest quarter because that is when the failure becomes visible. If the restaurant furniture specification does not name a density, the factory will pick the one that quotes best, and the room will pay for it later.