480 cafe chairs passed the buyer's own thickness check, and 3 mm of every seat panel was compressible foam
Illustrative example This case is written from the patterns we see in furniture export orders, and the figures are representative rather than from a specific client file. It shows how a check works and what it finds. We will replace it with a real, anonymised case as soon as we can publish one with the buyer's permission.
The order, and the four lines in the specification
A European cafe group was refitting fourteen sites and placed a single order for 480 stacking cafe chairs with a factory in Foshan. Steel tube frame, plywood seat shell with a veneered face, contract grade for daily commercial use. The unit price was $46, the order value $22,080, and the group paid 40 per cent as a deposit and held the remaining $13,248 back.
The buyer's specification for the seat shell ran to four lines. Two were about appearance: the veneer species and the sheen of the lacquer. One was about geometry: finished seat height and seat depth. The fourth was the line that decided the case, and it read: seat shell 9 mm birch plywood, 0.6 mm natural veneer face, finished panel thickness 9.6 mm.
A load requirement sat under the same clause. The seat shall withstand 1,300 N applied at the seat centre in accordance with EN 12520, ten applications of ten seconds. That number is not administrative. It is what decides whether a chair that a customer sits on for two hours at a time is still flat in the third year of a contract.
The group's own quality team was on site during production and inspected the first forty chairs with a caliper, as its incoming procedure required. Every reading came back at 9.6 mm. Nothing in that procedure was looking at the panel's construction, because the thickness was what the order had specified.
This is an illustrative example drawn from a recurring pattern in cafe and contract seating orders, where a construction requirement is reduced to a single number on the incoming check sheet. The figures are representative of the category rather than taken from a named customer file.
Where to go next
→ our dining and chair verification scope, including seat load and panel construction
What the caliper could not see
The order went into production in February and ran for eight weeks. For the first four weeks the factory laminated seat shells from 9 mm birch ply, and the panel was exactly what the buyer's calendar sample showed: nine millimetres of plywood laminations with a thin veneer line glued to the top face.
In the fifth week the 9 mm stock ran out and was not replaced with the same thing. What came in was 6 mm birch ply faced with a 3 mm layer of low-density foam and a 0.6 mm veneer bonded over the foam. The finished panel measured 9.6 mm at the exposed edge, exactly as before, because the foam took up the difference.
This is the whole mechanism of the case, and it is worth stating plainly because it defeats the most ordinary check in the trade. Nothing was falsified and no measurement was wrong. The factory built to the number the buyer's own schedule gave them, because the schedule gave a total and a total was what got produced. A caliper at the edge cannot tell you what the 9.6 mm is made of.
The buyer's incoming check had been done once, in week three, on the first forty chairs. Those forty were genuine, because the substitution had not started. The check was never repeated on the later production, and the incoming procedure did not require it to be. By the time the container was due, the lot held two different seat constructions: 120 chairs built from 9 mm ply, and 360 built from 6 mm ply with a foam-filled face.
Reading the section on the exposed edge
The seat shell is bolted to the frame, and the shell is wider than the frame on all four sides. Its raw edge is exposed on every chair, which means the panel's construction can be read directly off the finished product, without cutting anything and without asking anyone for a factory offcut.
The read is a matter of looking at the panel edge with a steel rule and a magnifier, and photographing it with a scale in frame. On the first chairs, the edge showed nine millimetres of plywood laminations with a thin veneer line sitting directly on top of them and a visible glue line between the two. On the later chairs, the edge showed six millimetres of laminations, then a three millimetre band of pale grey foam, then the veneer. The total at the edge was 9.6 mm on both.
Where the edge is measured matters as much as what it shows. The front nose of a seat shell is eased and rounded for comfort, so a caliper there reads low and reads differently on every chair. The crowned face of a veneered panel reads high. The jaws have to sit square to the panel or the reading is low by a few tenths of a millimetre. The edge should be read at least 100 mm back from the front nose, along the side edge, and at the front third, the middle and the rear.
And the position in the run is a measurement too. Panels were read from the front of the first run, from the middle and from the rear, and the changeover point fell at chair 121. A single panel from the middle of a stack tells you what the factory is doing now. Three panels from three positions tell you what the factory did across the lot, which is a different and much more useful statement.
One further check was available and it cost nothing. The foam underlay takes a knife blade and a thumbnail; plywood does not. Pressing the band on the panel edge with a fingernail leaves a lasting dent in the foam and nothing at all in the ply.
| What the edge showed | Chairs 1 to 120 | Chairs 121 to 480 |
|---|---|---|
| Plywood laminations | 9.0 mm | 6.0 mm |
| Layer between ply and veneer | None, glue line visible | 3.0 mm low-density foam fill |
| Veneer face | 0.6 mm, bonded to the ply | 0.6 mm, bonded to the foam |
| Total panel thickness at the edge | 9.6 mm | 9.6 mm |
| The buyer's own caliper check sheet | 9.6 mm, pass | 9.6 mm, pass |
The last row is the point of the whole case. The factory delivered a panel that passed the buyer's own measurement, and the panel was not the panel that was ordered.
Where to go next
The load test, and what the section is worth
A section tells you what was built. It does not tell you what that costs, and for a seat the difference matters more than the paperwork does. That is where a chair frame strength test earns its place: the same load, applied the same way, to the same chairs, with the approved sample standing alongside as the reference.
The 1,300 N seat load from the buyer's schedule was applied at the seat centre through a 300 mm square loading plate, held ten seconds and repeated ten times, in accordance with the method the order named. Deflection was read with a dial gauge against the frame, and the panels were re-read thirty minutes after unloading. The early chairs and the later chairs were tested side by side on the same day, on the same rig.
The chairs built from the 9 mm ply deflected 1.7 mm under load and came back to within 0.1 mm of where they started. The chairs built from 6 mm ply with a foam-faced panel deflected between 4.3 and 5.2 mm, and between 1.1 and 1.4 mm of that stayed. The veneer face on those chairs was visibly dished at the centre when the load came off, and it stayed dished.
The mechanism is not mysterious once the section is known. A 1,300 N load across a 300 mm square plate bears on the veneer, and the veneer on the later chairs is bonded to a 3 mm foam layer rather than to the ply. The foam carries almost nothing and simply crushes; the load then goes into six millimetres of plywood instead of nine, and the panel has two thirds of the section it was designed to have. Nothing on the outside of the chair says any of this.
The fixings were measured rather than assumed. The shell is screwed to the frame through the underside of the plywood, and with only 6 mm of ply in the panel the screws had 6 mm of solid material to bite into instead of 9 mm. On the later chairs, two of the four shell fixings showed visible movement at the head after the load cycle and needed a spanner to tighten.
| Measurement at 1,300 N seat centre | 9 mm ply, chairs 1 to 120 | 6 mm ply plus foam, chairs 121 to 480 |
|---|---|---|
| Deflection under load | 1.7 mm | 4.3 to 5.2 mm |
| Permanent set 30 minutes after unloading | 0.1 mm | 1.1 to 1.4 mm |
| Seat face after unloading | Flat | Dished at the centre, stayed dished |
| Shell fixings at the frame | No movement | Two of four moved at the head |
| Reference | Buyer's approved sample | Measured the same day on the same rig |
The order gave a load and a construction but no limit on deflection, because the buyer had never needed one before. The report did not invent a tolerance the order did not contain. It recorded the numbers and set the production against the approved sample, which is the only reference the order did supply.
Where to go next
→ our dining and chair verification scope, including seat load and panel construction
What the report recorded, and what it did not
The report established that the lot contained two seat constructions, and said where the line between them fell: chairs 1 to 120 built from 9 mm birch ply with a veneer bonded directly to the ply, and chairs 121 to 480 built from 6 mm birch ply with a 3 mm foam layer between the ply and the veneer. It recorded the edge reads with photographs and a scale, the eight panels section-read, the five chairs load-tested, and the changeover point at chair 121.
It also established that the thickness never varied. Both constructions measured 9.6 mm at the exposed edge, and the report said so, because a reader who found only the non-conformity would assume the thickness was wrong as well. It was right, and that is exactly why it was useless as a check.
What it did not establish: that all 360 chairs carry the substituted build, on the evidence of eight panels and five load tests; whether the chairs would pass a fatigue requirement, which neither the buyer nor the inspection can settle without an accredited laboratory; and that the chairs will fail in service. The report states a construction difference and a measured response to a specified load, and leaves the decision where the contract puts it.
The report was issued as an order verification statement against the buyer's documents, not as a test report, and it says so on its face. Its standing comes from what was recorded and how, not from a claim that EN 12520 conformity was certified.
Where to go next
Why the factory did it, and why nobody stopped it
The factory was not hiding anything, and the buyer's own specification had set the trap. The order said 9 mm plywood faced with 0.6 mm veneer, and a total of 9.6 mm. The work order that went onto the shop floor carried the total. A shop making 480 panels to a total will reach that total, and will reach it with whatever material is in the store that week.
There was a cost reason behind it, and it was not marginal. Birch ply is priced by thickness, and a drop from 9 mm to 6 mm is a third of the panel's material cost gone before the foam and the extra face are counted. The order price had been fixed at $46 when the quotation was accepted. When the 9 mm price moved, the panel was the only line on the order with any give in it.
The first 120 chairs were built before the change, and the buyer's calendar sample was taken from them. That is the entire reason a sample check passed. A sample proves what the factory can do, not what it did on the other 360 chairs, and here the difference was a material change that nobody was required to report.
The buyer's incoming check sheet had one line for the seat shell, and it was a caliper reading at the panel edge. It was a good check. It was the wrong check for the only question the order actually raised.
Where to go next
What the specification should have said
Seven lines. None of them costs anything to write, and together they replace a single thickness with a construction requirement and an acceptance limit that a report can actually defend.
- Name the ply, not the total. Seat shell: 9.0 mm birch plywood with a 0.6 mm natural veneer face bonded directly to the plywood, finished panel thickness 9.6 plus or minus 0.3 mm. The panel shall contain no foam, fibre or other fill layer between the plywood and the veneer face. Fixings shall engage not less than 8 mm of solid plywood.
- State the section check and where to look. Panel construction shall be verified on the exposed raw edge of the seat shell, at least 100 mm back from the front nose, at the front third, the middle and the rear of the shell, by reading the laminations and photographing the edge with a scale in frame. The eased front nose shall not be used for this check.
- Give the load an acceptance number. The seat shall not deflect more than 3.0 mm at the seat centre under 1,300 N applied through a 300 mm square plate for 10 seconds, and shall retain no more than 0.2 mm of permanent set 30 minutes after unloading, measured against a reference chair built from the approved sample.
- Make the change rule explicit. Any change to the panel material, ply thickness, veneer or face construction requires written approval before production continues, including a change made after the first article has been approved.
- Make the sample and the run the same material. The first article and the production run shall be cut from the same plywood lot. One unit from the final ten per cent of the run shall be section-read at the raw edge and recorded, so that a changeover cannot pass on the strength of a sample taken at the start.
- Fix the checklist as well as the order. Where the buyer's own incoming procedure measures a panel with a caliper, the check sheet shall require a section read as well. A thickness reading cannot establish construction, and the two checks answer different questions.
- Write down what happens next. Where a finding turns on the construction of a panel rather than on a measurement, the order shall state the decision: release, rework or replace, and who carries the cost. A report can tell the buyer that 360 chairs are not what was ordered. It cannot decide what to do about them without a line in the order to point at.
The point of all seven is not to catch the factory out. The factory built 480 panels to a 9.6 mm total, exactly as the buyer's own schedule asked, and the buyer's own caliper confirmed it. The construction was never the thing either side was writing down, and the order that would have caught it costs seven lines to write.
Where to go next
→ our dining and chair verification scope, including seat load and panel construction
→ what a report records when a lot passes and a lot fails
→ how sampling works under ISO 2859-1
What the buyer did with the finding
The 120 chairs built from 9 mm ply were released against the balance. The 360 were held. The factory replaced the seat shells on those 360 at their own cost, cutting new 9 mm birch panels and gluing new veneer faces to match the approved sample, and the work took eleven working days.
A re-check on 32 chairs from the reworked lot section-read the edge of the first and the last chair in the run and load-tested five of them. Every read showed 9.0 mm of laminations under the veneer and every deflection came in under 2 mm. The balance was released four days after the re-check report.
The group's own specification was amended the same week, on the seven lines above, and the amendment went to the two other factories it worked with. The second supplier read it and asked a question the original order never would have drawn: which part of the 9.6 mm is plywood?
Where to go next
Have an order you need verified? Send us your PI and factory address. Three levels, published prices, $149 per man-day from.