Quick Answer: A canopy inspection happens in the 5–7 days before the container is packed, not the day it ships — after that, every defect becomes a freight problem instead of a factory problem. The 12 points below are what we check on every order, in this order, and each one has a specific failure it catches. Points 3, 7 and 10 are the three that fail most often on a first production run.
Most canopy problems are decided before the container closes. A millimetre of seal gap, a rail that was drilled from the wrong side, a gas strut that was tested once instead of cycled — none of them look like much on a loading dock, and all of them become the customer's problem three weeks later.
This is the checklist we run, why each line exists, and what to do when a line fails.
1. What inspection catches, and what it cannot
Pre-shipment inspection catches build errors: the wrong bracket, an unsealed seam, a door that binds. It does not catch design errors, and it cannot predict a three-year failure.
That distinction matters because buyers often ask inspection to do the second job. A 12-point check will tell you the unit in the container matches the drawing. It will not tell you the drawing is right for a truck that has not been built yet.
Two rules follow:
- Inspect against a drawing, not against taste. "Looks fine" is not a criterion; "rail height 1,712 mm ± 2 mm" is.
- Inspect before packing, not before shipping. Once the crate is closed and strapped, the cost of finding point 7 has gone up by a container.
Three things inspection is often asked to do and cannot:
Confirm the design fits the truck. That is verification against the vehicle, done from scan data or from a physical trial fit. If the rail height was specified for the wrong cab, every unit will pass inspection and none will fit.
Predict durability. A canopy that passes all twelve points today can still leak in month 30 because a seal was bonded to a surface that was not properly prepared — a condition you can see if you watch the process and cannot see on a finished unit.
Cover transit. Inspection ends when the doors close. What happens between the loading dock and the destination is a packing and strapping question, which is why points 11 and 12 sit on the same sheet rather than on a different one.
2. Timing: why 5–7 days before loading
Production runs 15–25 working days. Inspection at day 20 of 25 leaves enough time to rework a bracket or re-cut a seal; inspection on the loading day leaves time only to argue.
We schedule three checkpoints:
| Checkpoint | When | What it decides |
|---|---|---|
| First-article check | First unit off the line | Whether the batch is allowed to continue |
| In-process check | Mid-batch | Whether the same error is repeating |
| Pre-shipment check | 5–7 days before loading | Whether the batch ships |
The first-article check is the one buyers skip and the one that pays. If a bracket is drilled 8 mm off on unit 1, it is drilled 8 mm off on unit 25.
Worked backwards from a loading date, the schedule looks like this:
- Day 0 — PO confirmed, drawing revision locked
- Day 3–5 — first article off the line, checked against the drawing, batch released or corrected
- Day 12–15 — in-process check on a mid-batch unit, same 12 points on the structural four
- Day 18–20 — pre-shipment check on the sample, photographs taken, open items listed
- Day 22–25 — rework window, then packing and loading
If the drawing revision is not locked on day 0, every later checkpoint inspects the wrong thing. We have seen a batch built to a superseded rail height because the revision was confirmed verbally.
3. The 12-point checklist
The order matters: the first four are structural and cannot be fixed later, the middle five are sealing and function, the last three are paperwork and packing.
| # | Point | What is measured | Failure it catches |
|---|---|---|---|
| 1 | Frame squareness | Diagonal difference ≤ 3 mm | Doors that bind after installation |
| 2 | Rail height vs drawing | ± 2 mm | Fitment failure on the vehicle |
| 3 | Panel thickness at 5 points | 2.0 / 2.5 / 3.0 mm per spec | Panels rolled from the wrong coil |
| 4 | Weld and rivet spacing | Per drawing, no gaps > 1 mm | Water paths that open in service |
| 5 | Door alignment and gap | 4–6 mm uniform | Wind noise, seal wear |
| 6 | Seal compression | Continuous contact, no dry section | The leak that starts in month 3 |
| 7 | Hinge torque and operation | 20 full open-close cycles | Sag within the first year |
| 8 | Gas strut cycle test | 10 cycles, both struts | Failure in pairs, months later |
| 9 | Latch and lock function | Open, close, lock, unlock × 5 | Latch that releases in transit |
| 10 | Seal adhesion at corners | No lifting under 90° peel by hand | The corner that peels first |
| 11 | Hardware count vs packing list | 100% count | Missing bolts at the install site |
| 12 | Crate, wrap and load plan | Per drawing, strapped | Transit damage and moisture |
Points 3, 7 and 10 are, in our experience on first production runs, the three that fail most often. They share a cause: each depends on a process that looks correct from a distance — a coil spec, a torque setting, an adhesive cure.
Each point is recorded as a value, not a tick. A sheet of twelve ticks tells you nothing on the next order; a sheet of twelve numbers shows a trend — rail height drifting 1 mm per batch, seal compression tightening as a new adhesive supplier beds in. The value is what makes the second order cheaper than the first.
The order of the twelve is also deliberate. Points 1–4 are structural: once a frame is welded and a rail drilled, the error is in the product, not in the process. Points 5–10 are functional and can usually be corrected in place. Points 11–12 are logistics, and they are the cheapest to fix and the most expensive to discover late.
4. The three checks that fail most often
Point 3 — panel thickness. A coil that is 0.1 mm under spec still forms, still welds, and still looks identical. It only shows up when the canopy is loaded, or two years later when it dents. Measure at five points per panel, not one: thickness varies across a coil.
Point 7 — hinge cycle and torque. A hinge that opens once is not tested. Twenty full cycles is the minimum that separates "moves" from "will still align in a year". Torque is checked after the cycles, not before, because cycling settles the fastener.
Point 10 — corner seal adhesion. Corners are where the seal is stretched most and where adhesive cure is least even. A hand peel at 90° is crude and effective: if it lifts, the corner will open.
None of the three requires equipment beyond calipers, a torque wrench and a hand. That is the point: they fail because they are cheap to skip, not because they are hard.
The other nine fail less often for a structural reason: they are checked by people whose next step depends on them. A door that binds will not close, so the fitter notices. A seal gap shows at the first water test. A short hardware count stops the packing line. Points 3, 7 and 10 are the ones where nothing downstream complains immediately — which is exactly why they need a written checkpoint rather than an experienced eye.
5. Sampling when the order is 25 units
A 20HQ holds roughly 25 canopies, so most orders are small enough that sampling rules written for thousands do not apply. We use a three-tier approach:
| Batch size | Sample | What the sample must include |
|---|---|---|
| 1–10 | Every unit | All 12 points |
| 11–30 | 5 units | Points 1–4 on all 5; 5–12 on 2 |
| 31+ | 8 units + 1 per 20 | Points 1–4 on all; 5–12 rotating |
The sample must also be pulled at random from the finished line, not selected by the person being inspected. If the five units were chosen for you, you inspected a showroom.
For buyers working with an AQL-based third-party inspector, the same logic translates: AQL 2.5 major / 4.0 minor on a 25-unit lot accepts 2 major defects in a sample of 5. On a canopy, a seal gap is not minor.
A worked example. A 25-unit order, sample of 5, AQL 2.5 major. The inspector finds one unit with a corner seal lifting (point 10) and one with a missing bolt kit (point 11). The seal is a major defect because it leads to water ingress; the bolt kit is minor because it can be shipped separately. One major in a sample of 5 is inside the acceptance number, so the lot passes — and both findings still go into the report as open items.
That is the limitation buyers need to understand: AQL decides whether the lot is accepted, not whether the unit in front of you is right. If the unit matters — a display canopy, a first order for a new market — inspect it individually regardless of the sampling plan.
6. Photos: what to demand, what they hide
Photos are the cheapest inspection tool and the easiest to stage. Ask for these five, unedited, with a timestamp:
| Photo | Why | What it hides |
|---|---|---|
| Unit on the line, wide shot | Shows the batch exists | Nothing — this is the honest one |
| Rail height with caliper in frame | The measurement, not the claim | A caliper on a different unit |
| Corner seal close-up, 3 corners | The adhesion point | The fourth corner |
| Hinge with torque wrench fitted | That torque was applied | Whether it was applied after cycling |
| Crate before closing + after strapping | Packing and load pattern | What happens in transit |
A photo of a canopy on a showroom floor is not inspection evidence. A photo with a caliper in frame, showing a reading, is.
Two practical rules for photographs. First, ask for the same five angles on every order, so that a series across shipments is comparable — an inspector who improvises produces images you cannot line up. Second, insist on unedited files with the original timestamp, not a compiled PDF; a PDF can be assembled from any order, a timestamp plus a container number in frame cannot.
7. Packing and loading are part of the checklist
Inspection does not stop at the product. Points 11 and 12 exist because the most common transit damage is not a manufacturing defect at all — it is a crate that was built for a shorter unit, or a strap that ran through a door seal.
Three things to confirm before the container closes:
- Crate internal dimension vs the canopy plus foam, not the canopy alone
- Strap path clear of seals, glass and gas struts
- Moisture control: desiccant sized to the container volume, not a fixed number of bags
If the load plan is wrong, the inspection report is still accurate and the delivery is still damaged.
Moisture deserves its own paragraph because it is the defect that arrives late. A canopy is steel, aluminium, EPDM and a sealed cavity; a container crossing a cold route condenses water on the coldest surface, which is usually the inside of the roof panel. Desiccant sized to container volume — not a fixed number of bags carried over from a smaller order — plus a vapour barrier on the floor covers most of it. A canopy that arrives with water marks inside the roof was not a manufacturing defect; it was a loading decision.
Stacking is the other late defect. Crates are designed for a stack height, and a container loaded one tier higher than the design will show it as a deformed top rail, not as a broken crate. The load plan on the packing list should state the tier count, and the inspector should confirm it before the doors close.
8. What the report should contain
An inspection report that says "passed" is not a report. It needs four things:
- The drawings and revisions the units were checked against — a report against an outdated drawing certifies the wrong thing
- The measurement values, not just pass/fail — so a trend across orders is visible
- The sample selection method — who pulled the units
- Open items and disposition — what was reworked, what was accepted with deviation, what was rejected
The fourth is the one that protects both sides later, when a warranty claim arrives.
A one-page skeleton that has worked for us:
- Order number, drawing revision, inspection date, inspector name
- Batch size, sample size, selection method ("5 units pulled by warehouse, not by production")
- Twelve values, one line each, with pass/fail against the drawing tolerance
- Photographs numbered to match the twelve lines
- Open items: what was reworked, by whom, and when it was re-checked
- Signature from production and from quality — two names, not one
Two names matter more than they look. A single signature means the person who built it also cleared it.
9. If a point fails: four dispositions
| Disposition | When | What it costs |
|---|---|---|
| Rework in place | Paint, seal, hardware, latch | Days |
| Replace component | Panel, door, strut | Days plus material |
| Reject unit | Structural, out-of-drawing | One unit, not the batch |
| Accept with deviation | Cosmetic, documented, priced | A credit and a written record |
Two rules we apply: a structural failure (points 1–4) is never accepted with deviation, and a deviation is always written into the shipping documents rather than agreed on a call.
A real sequence worth recognising: a batch comes off the line with a 0.6 mm gap at one corner seal (point 10). It is cosmetic-adjacent, the schedule is tight, and the customer is offered a credit. Six weeks later the canopy leaks at that corner and the credit is forgotten; the conversation is now about a replacement and a freight claim that exceeds the credit by an order of magnitude.
The economically correct call in that moment is rarely the one that protects the schedule. A corner that fails a hand peel is a rework item, not a deviation item, because the failure is in a joint that the customer cannot inspect and cannot repair.
Sources & References
- Trigallop production and inspection record, 2026 (internal, A-level): three production lines, 58 staff, 15–25 working day lead time, 28-month warranty from receipt.
- Trigallop fact ledger, 2026-11-05 edition: canopy EXW from US$538; 20HQ ≈ 25 units; HS 8708295900 (internal, A-level).
- ISO 2859-1, Sampling procedures for inspection by attributes — the AQL basis referenced in section 5 (standard, L2).
- Trigallop pre-shipment inspection sheet, rev. 2026-09 — the 12 points in section 3 are the current sheet's items (internal, L1).
We are a factory, not a third-party inspection body; the checklist above is our own process, shared so buyers can ask for it by name. Your inspector's report governs your order.
Related reading: what a supplier should send before shipping · where canopies actually leak · why canopies rust and when it matters · the questions to ask a canopy factory · panel thickness: 2.0 vs 2.5 vs 3.0 mm · how a canopy is load tested · warranty claims, step by step · shipping a canopy from China · the 2027 order cutoff before Chinese New Year · what a canopy weighs and why it matters for freight · seal maintenance between orders · measuring your own bed before ordering · becoming a distributor · our canopy line.