TL;DR: The single biggest cause of requotes and delayed samples is an incomplete brief — missing one dimension or an unspecified substrate can add 2–3 sample iterations to your timeline.
TL;DR: A production sample for a folding carton inspection project typically takes 15–20 working days from brief approval; white samples can be ready in 5–7 working days if structural files are complete on day one.
What Your Brief Must Contain Before We Can Quote Accurately #
A quotation for an automated inspection system integration — whether that means a standalone camera inspection unit, an inline print defect detection module, or a full 100% vision inspection line — cannot be accurate without a complete technical brief. When we receive an incomplete RFQ, we either hold it pending clarification (which delays your timeline) or quote with assumptions built in (which means the price changes when the spec gets locked). Neither outcome helps you.
The minimum viable brief contains:
- Line speed and format range: web width (mm) or sheet size, and maximum run speed in metres per minute or sheets per hour. A system specified for 80 m/min will not reliably inspect at 150 m/min — the camera trigger interval, lighting pulse duration, and image buffer are all sized to the speed window.
- Substrate type and surface finish: coated vs uncoated paper, BOPP, PET film, foil laminate, or corrugated board. Surface reflectance affects illumination design; a dark matte substrate and a high-gloss metallised film require different light angles and often different camera sensor types.
- Defect classification list: which defect types are in scope (colour delta, register error, barcode grade, missing print, hickeys, streak, splice detection) and what the reject threshold is for each. This directly drives algorithm configuration and ultimately system cost.
- Integration point: inline (mounted on press or converting line), offline (rewinder or die-cut line), or standalone inspection table. The mechanical interface — whether we’re supplying a stroboscopic unit, a contact image sensor bar, or a line-scan camera array — depends entirely on where the system will sit in the production flow.
- Quantity tiers: annual volume in million impressions or linear metres, and whether the quantity is a committed forecast or an indicative budget figure. Inspection system pricing is not linear — at below 50 million impressions/year, a modular entry-level configuration is typically more cost-efficient than a full 100%-coverage broadweb system.
| Brief Element | Impact if Missing | What We Assume (and the Risk) |
|---|---|---|
| Line speed | Cannot size camera trigger rate | We assume 80 m/min — if actual is 150 m/min, hardware changes |
| Substrate type | Cannot specify illumination module | We assume coated paper — metallised film may require full redesign |
| Defect threshold per class | Cannot configure algorithm accuracy targets | We assume ISO 12647-2 delta-E ≤2.0 — tighter tolerances cost more |
| Integration point | Cannot specify mechanical mount | We assume standalone — inline may require press-side safety compliance |
| Annual volume | Cannot recommend system tier | We assume 50M impressions — underspecification leads to overselling |
The table above is our internal RFQ intake checklist — what we call the QC-AIS-01 completeness gate. If two or more fields are blank, the brief goes back before engineering looks at it.
Where Briefing Errors Translate into Sample and Cost Rework #
Incomplete or ambiguous briefs create three distinct failure modes, and each one adds real calendar time.
The first is substrate-illumination mismatch. A brand team specifies “flexible packaging” without confirming whether the film is transparent, opaque white, or metallised. We configure the illumination rig for diffuse backlighting suited to transparent film. The actual production material arrives as a white LDPE with a hot-stamp foil panel. The illumination geometry produces glare hotspots across 15–20% of the image frame, and the false reject rate climbs above 3% — which is operationally unacceptable on any production line running more than 200 ppm. The rework requires a new light-angle fixture, a requalification run, and at minimum two weeks of delay. The cost delta is measurable and not trivial.
The second failure mode is register tolerance ambiguity. Inspection systems that flag register errors must be configured with a pass/fail threshold. A ±0.3mm tolerance is our default for premium folding carton work, based on consumer-detectable misregister benchmarks we’ve validated on our own offset press lines. But if the client’s press has a mechanical register variance of ±0.5mm and the inspection system is set to ±0.3mm, the false reject rate will be high enough to shut the line down within the first hour of production. This happens when buyers submit the brand’s print specification without also supplying the press capability data. Both documents are required for correct system calibration.
The third, and most common, failure mode is resolution undersizing. A camera array with 150 dpi effective resolution will not reliably detect a 0.2mm pinhole in a pharmaceutical blister lidding film. Pharmaceutical and medical device packaging typically requires 200–400 dpi effective resolution under ASTM F2250 leak detection guidance and EU Good Manufacturing Practice Annex 1 visual inspection standards. When a buyer’s brief says “pharma packaging” without specifying the defect size tolerance, we sometimes see systems quoted at 150 dpi to hit a budget target — then rejected by the QA team on first validation. Specifying minimum detectable defect size in millimetres removes this ambiguity entirely.
Do You Need a White Sample Before Committing to a Production Build? #
For inspection system projects, the equivalent of a white sample is a functional demo run on your actual substrate with your actual artwork file — sometimes called a proof-of-concept trial. Yes, you should request one before committing to a full production build.
The proof-of-concept trial tells you two things a paper specification cannot: whether the illumination geometry handles your specific substrate surface without glare artefacts, and whether the algorithm’s sensitivity settings produce an acceptable false reject rate on your actual print variation. We run these trials on a 500–1,000 metre substrate roll or 2,000–5,000 sheet sample, which is sufficient to expose most systematic issues. Lead time for a configured demo unit is typically 10–15 working days from receipt of substrate samples and confirmed artwork files. If the trial requires FSC-certified or food-contact-compliant substrate (per FDA 21 CFR Part 176 for paper in contact with food), confirm that at the outset so we source correct material for the test run.
Specification Notes for Brand Partners #
When you brief us on an automated inspection system project, we need five things before engineering can open a quote: confirmed line speed (m/min or sph), substrate specification with surface finish, defect class list with pass/fail thresholds per class, the integration point in your production flow, and annual volume in impressions or linear metres.
The brief gap that causes the most sample iterations is missing defect threshold data. Brands often send us their artwork quality standard (usually referencing ISO 12647-2 or a proprietary brand colour specification) without telling us what their converting process is actually capable of. If your press runs at ±0.4mm mechanical register variance and your brief specifies a ±0.2mm inspection threshold, the system will fail on first production run. Before briefing us, pull your last 3–6 months of press QA data and confirm the real-world process variation — not the target specification, the actual measured performance.
Sampling timeline: proof-of-concept trial, 10–15 working days from substrate receipt. Configured demo unit with report, 15–20 working days. Full production build and FAT (Factory Acceptance Test), 35–50 working days depending on integration complexity. What extends the timeline most is late confirmation of the integration point — mechanical interface changes after engineering sign-off are the single most common cause of project delays in our experience.
Frequently Asked Questions #
What file formats do you need for artwork before you can configure an inspection system?
We need a print-ready PDF/X-4 or layered TIFF at a minimum of 300 dpi at final print size, plus a separate colour reference file showing target LAB values for each spot colour or a calibrated CMYK profile. If your artwork contains barcodes, include the barcode specification (symbology, nominal size, and minimum bar width) separately — our barcode verification module is calibrated independently from the colour inspection channel.
What’s the minimum order quantity for a configured inspection system?
There is no unit MOQ in the traditional sense — inspection systems are engineered to order, not stocked. What affects pricing is annual consumable volume (replacement lamps, calibration targets, cleaning kits) and whether you’re ordering a single standalone unit or a multi-line rollout. A single-line entry configuration starts from a different cost base than a five-press inline rollout, and volume commitment across a 3-year service agreement typically changes the unit cost by 15–25%.
Can you match our existing colour standard if we’re already approved to G7 Grayscale Verification?
Yes, but we need your G7 characterisation data and press fingerprint, not just the G7 certificate. The G7 standard (defined by Idealliance) specifies neutral grey balance targets — our inspection system calibration uses those same LAB targets as the pass/fail reference. Without the actual press ICC profile and characterisation data, we’re calibrating to the standard’s generic targets, which may not match what your press actually produces in production conditions.
How do we evaluate the trial run report to decide whether to proceed to a full build?
Look at three numbers: false reject rate (target below 1.5% for most packaging applications, below 0.5% for pharmaceutical), true defect detection rate against the seeded defect set we run in every trial, and the delta-E distribution across the colour patches. If the false reject rate is above 3% on the first trial run, the issue is almost always illumination geometry or threshold calibration — both correctable before the full build. If true defect detection on 0.3mm register errors is below 95%, that points to resolution or trigger-rate undersizing.
How do you compare quotes from different suppliers fairly when the specs look different?
It depends on whether the quotes are specifying the same camera resolution, the same defect class coverage, and the same integration scope. A quote that omits offline calibration fixtures, spare lamp sets, or FAT support will look cheaper by 20–30% but will cost more over 18 months of operation. The safest comparison method is a side-by-side bill of materials review — ask each supplier to itemise hardware, software licence, integration labour, calibration tooling, training days, and first-year service separately. We provide this breakdown as standard in our quotation format.
What regulatory standards apply to the inspection system validation process?
For pharmaceutical and medical device packaging, the relevant framework is EU GMP Annex 1 (sterile products visual inspection) and FDA 21 CFR Part 211.68 for automated equipment qualification. For general food packaging, GFSI-benchmarked schemes (BRC, FSSC 22000) require documented AQL inspection procedures — typically AQL 1.0 for critical defects, AQL 2.5 for major defects under ANSI/ASQ Z1.4 sampling plans. For print quality specifically, ISO 12647-2 for offset and ISO 12647-6 for flexo define the colour tolerances our systems are calibrated against. Which standard governs your validation protocol should be confirmed in your brief — it affects system documentation, IQ/OQ/PQ deliverables, and lead time.
What’s the most common mistake brands make when comparing inspection system quotes?
Specifying line speed as the maximum rated press speed rather than the actual operating speed. A press rated to 18,000 sph rarely runs at that speed in production — actual throughput for a decorated folding carton line with inline die-cutting is typically 10,000–13,000 sph. If you specify 18,000 sph in your brief, you’ll be quoted a higher-tier camera and buffer configuration than your operation actually needs. Spec to your 90th-percentile actual run speed, not the nameplate maximum.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.
