TL;DR: A batch release decision should never rest on a single viscosity reading — our protocol requires at least four independent test parameters to pass before ink is cleared for press.
TL;DR: We reject incoming ink lots where colorimetric drift exceeds ΔE 1.5 against the approved standard, which filters out roughly one in eight lots from new suppliers in the first six months.
Acceptance Criteria by Test Parameter: How We Define Pass/Fail Before Ink Touches the Press #
Before any ink lot is approved for production use, it goes through what we call our INK-QC-04 incoming validation sequence. This is a structured seven-parameter check, not a spot test. The sequence was developed after a run of colour inconsistency complaints on a folding carton programme in 2022 — fourteen SKUs, 180,000 units, and the root cause traced back to a single batch of process cyan that had shifted in grind fineness without any accompanying test certificate change.
The table below shows our standard acceptance thresholds for lithographic and flexographic ink systems. Gravure has a separate protocol (INK-QC-05) due to the much lower viscosity range involved.
| Test Parameter | Method | Acceptance Threshold |
|---|---|---|
| Colorimetric match (ΔE) | ISO 13655 / spectrophotometer | ΔE ≤ 1.5 vs. approved standard |
| Viscosity (flexo, water-based) | DIN Cup #4 / 25°C | 18–25 seconds |
| Viscosity (offset litho) | Laray falling-rod / 25°C | 40–65 Pa·s |
| Fineness of grind | Hegman gauge per ASTM D1210 | ≤ 10 µm (process) / ≤ 25 µm (opaque white) |
| pH (water-based systems) | Digital pH meter, calibrated | 8.2–9.2 |
| Drawdown adhesion | Tape pull per GB/T 13217.4 | No ink transfer at 90° peel |
| Tack (offset, 30°C) | Inkometer per ISO 12634 | ±15% of approved reference |
A lot that passes six of seven still fails. There are no partial approvals in INK-QC-04. If a parameter is borderline, the lot goes into quarantine and we request a re-test certificate from the supplier before any disposition decision.
The colorimetric and viscosity thresholds carry the most weight operationally. Colour drift above ΔE 1.5 is perceptible to a trained observer under D50 illuminant and would show up as a brand colour complaint within the first production run. For viscosity, a ±5-second deviation from the approved centre point for water-based flexo systems changes dot gain by roughly 3–5% on mid-tone values — enough to shift a skin tone or a gradient background noticeably on press.
What Causes Batch-to-Batch Drift — and Where Validation Actually Breaks Down #
Pigment dispersion is the variable that causes the most recurring failures in our incoming data. When a supplier changes their grinding media without notifying customers (which happens more often with smaller regional suppliers), the fineness of grind shifts even though the formulation specification on paper stays the same. A lot may arrive with a valid CoA showing 8 µm grind, but our Hegman gauge reads 14–16 µm. The CoA was accurate for the reference batch, not for the current production run. The consequence on press is ink film uniformity loss and mottling in solid areas — and it only shows up after 2,000–3,000 sheets, by which point the issue is already embedded in the production schedule.
Resin ratio drift is a subtler failure mode. Offset inks contain a balance of hard and soft resins that determines tack and setting speed. When a resin component is substituted due to supply chain pressure and the supplier doesn’t trigger a change notification, the tack value can shift by 20–25% while the colour spec remains within tolerance. We catch this via our inkometer check in INK-QC-04, but if a factory skips tack testing and relies only on colour drawdown, this goes undetected until press sheets show picking or linting on coated substrates.
The third failure mode is pH drift in water-based systems, and this one is almost entirely a storage and handling problem rather than a manufacturing defect. Water-based flexo inks are formulated at pH 8.2–9.2, but if stored above 35°C or exposed to CO₂ in poorly sealed containers, pH can drop to 7.4–7.8 within four to six weeks. Below pH 8.0, the resin emulsion destabilises, viscosity climbs unpredictably, and dot sharpness on fine screen work degrades. We’ve seen this specifically with lots shipped by sea freight during summer months in sealed containers without climate control. Our current practice is to test pH on every water-based lot regardless of CoA date — we don’t give pH a “freshness pass.”
Calibration frequency is where validation protocols lose integrity over time. Our pH meters are calibrated against two-point buffer solution (pH 4.01 and pH 7.00 per ISO 3696) every 48 hours of use. The Hegman gauge is verified against a certified block standard at the start of every incoming inspection week. Inkometer calibration follows the manufacturer’s quarterly schedule, cross-checked against a reference fluid with a known viscosity of 1,000 mPa·s. If any calibration check fails, all test results since the last successful calibration are invalidated and the affected lots are re-tested. This has happened three times in the past two years — each time caught at the calibration check, not at a customer complaint.
Should You Require Full Test Certificates for Every Ink Delivery? #
Yes, with a qualifier on scope. A standard CoA from the ink supplier covers colour, viscosity and fineness — which is adequate for stable, long-running jobs where the formulation has been qualified for more than 12 months without incident. For new ink suppliers, new formulations, or any lot arriving for a food-contact compliant job requiring low-migration documentation under EU 10/2011 or FDA 21 CFR §175.300, a supplier CoA is not sufficient by itself. We require independent third-party migration testing before lot release — typically a full compositional declaration and a specific migration test at 10 days / 40°C per the EU framework.
For standard commercial jobs, the supplier CoA covers the baseline. Our INK-QC-04 drawdown and viscosity check adds the incoming verification layer. Both together are the minimum for a reliable batch release decision.
Specification Notes for Brand Partners #
When you brief us on a new packaging programme requiring colour-matched inks, we need at least three things to set the validation threshold correctly: your approved Pantone or spectral reference file (Lab* values under D50, 2° observer per ISO 13655), the substrate you’re printing on (coated, uncoated, laminated film), and whether the packaging has any food-contact or child-product regulatory exposure.
The brief gap that causes the most unnecessary sample iterations is an undefined substrate. A ΔE 1.5 acceptance threshold on a 350gsm SBS board does not translate to the same visual result on a matte-laminated PP film — the ink laydown, dot gain profile and surface energy are all different, and the ink formulation needs to be requalified for each substrate class. If you send us a pantone target without confirming the substrate, we’ll build the colour standard on our default stock, and the first physical sample on your actual substrate may look different even if it technically passes the same ΔE threshold.
Our standard sampling timeline for a new ink qualification is 10–14 working days from receipt of approved substrate and colour reference. This extends to 18–22 working days if third-party migration testing is required. Substrate availability on your side is usually what controls the timeline, not press availability on ours.
Frequently Asked Questions #
What AQL level do you apply to ink lot sampling?
For incoming ink lots, we apply AQL 2.5 per ISO 2859-1 at general inspection level II — this means a sample size of 32 units from a standard delivery lot of 800–1,200 kg, with a maximum of two non-conformances allowed before the lot is rejected.
Can you match a colour from a physical sample rather than a Pantone reference?
Yes, we do this regularly using spectrophotometric reverse-engineering — we measure the physical sample under D50 illuminant, extract Lab* values, and build a target. The accuracy depends on the substrate the sample is printed on. If the sample is on a different substrate than your production packaging, expect 0.5–1.0 ΔE of residual offset that is substrate-related rather than ink-related. For brand-critical colours, we recommend confirming the target on your actual production substrate before locking the spec.
How often do you recalibrate the equipment used in ink testing?
pH meters every 48 hours of active use, Hegman gauge weekly, inkometer quarterly — all against certified reference standards. Results since the last successful calibration are invalidated if a calibration check fails, which has happened three times in the past two years.
Does a passing CoA from the ink supplier mean you’ll skip your incoming tests?
No. The supplier CoA covers their outgoing QC at the point of manufacture. Our INK-QC-04 incoming check covers conditions after transit and storage. Viscosity, pH and grind fineness can all shift during shipping, particularly for water-based systems in summer sea freight. The CoA tells us what the ink was; our incoming check tells us what it is when it arrives.
What happens if an ink lot fails one parameter but passes all others?
The lot goes into quarantine. We do not run partial approvals. If the failing parameter is borderline (within 10% of the threshold), we can request a re-test certificate from the supplier and recheck our own sample before final disposition. If it fails cleanly, the lot is returned or destroyed — it does not go to press.
Is ΔE 1.5 tight enough for brand-critical Pantone colours?
It depends on the colour. ΔE 1.5 is our standard incoming threshold, but for high-chroma brand colours — particularly saturated reds, oranges and specific blues — perceptibility can be higher because the human eye is more sensitive to hue shift in those regions. For brand-critical colours with formal brand standards documentation, we tighten the threshold to ΔE 1.0 and note this in the job specification.
Do you test every single ink lot, or just spot-check?
Every lot, without exception, for any job with a colour-critical or food-contact requirement. For repeat jobs using a supplier we’ve qualified for more than 12 months with a clean incoming record, we run a reduced check (colorimetric and viscosity only) but maintain the full INK-QC-04 frequency for any new formulation or new delivery source. Reduced checking is an earned status, not a default.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.
The grind fineness threshold is where we’ve had the most supplier pushback — we had a litho ink lot come in at 14 µm on Hegman for a process colour, supplier argued it was “within industry norms,” but under INK-QC-04 logic that’s a hard fail. Took three resubmissions before they hit our ≤10 µm spec consistently.
The grind fineness threshold that caught us was actually on opaque white — we’d set our internal limit at 20 µm and kept seeing mottle on a PE-laminated sachet substrate until we tightened to ≤ 15 µm for that specific application, which the Hegman gauge per ASTM D1210 flagged immediately once we adjusted the spec.
The Laray falling-rod method gives you a true viscoelastic picture that a DIN Cup #4 simply can’t — but that gap matters a lot less for water-based flexo than people assume, since at 18–25 seconds you’re working in a range where the DIN result correlates well enough with press behavior. Where we’ve seen the real divergence is in cold-set offset during winter months, where a lot sitting at 64 Pa·s on the Laray can still show transfer problems that only show up on the Inkometer tack reading.
The grind fineness issue is exactly what we ran into with a Shenzhen-based ink supplier in early 2023 — they were shipping lots where the Hegman gauge was reading 14–16 µm on process colours, technically within their own spec but way outside our ≤10 µm threshold, and the drawdown adhesions were masking it because the substrate was forgiving. Took us three failed batches before we even knew to add fineness to our incoming checks.
The ΔE ≤ 1.5 threshold works fine for CMYK process work, but we had to tighten ours to ≤ 1.0 for a Pantone-heavy gifting client running spot colours on uncoated kraft — at 1.5 you’re already visually off on that substrate because the ink absorption is so variable lot to lot. We kept the 1.5 limit in INK-QC-04 equivalent for process builds but it’s essentially a two-tier system now depending on the job spec.
The pH window of 8.2–9.2 looks standard but we’ve had to narrow that to 8.5–9.0 for a foil-laminate watch box liner we run on a Nilpeter FA-4 — anything below 8.5 and we were seeing adhesion failure at the laminate interface within 72 hours of converting, confirmed across three consecutive lots from our Taipei supplier in Q3 2024.
The “no partial approvals” rule took us a while to enforce internally — we had a press supervisor who’d been signing off lots that failed tack by 18–20% because everything else looked clean, and it cost us three stripped blankets on a litho carton job before QC finally owned that decision fully.
The drawdown adhesion test is the one that stings the most in retrospect — we had a gold metallic ink pass every other parameter on a Christmas gifting box run (52,000 units, foil-blocked rigid boxes for a UK supermarket own-brand) but the tape pull was done on the wrong substrate. Lab used coated board offcuts, press was running uncoated duplex. Ink transferred clean in production, we shipped, and by the time the boxes hit the retailer’s back room the metallic layer was flaking off at the spine fold. Full recall, full reprint, four weeks before the seasonal window closed.
The Inkometer tack spec at ±15% against reference is reasonable for most offset work, but we’ve found it doesn’t tell the same story for UV-curable versus conventional litho inks on the same substrate — UV lots can sit comfortably within that window at 30°C and still cause picking on uncoated board because the tack curve shape is fundamentally different, not just the peak value. Running the Inkometer sweep at multiple speeds (we added 400 and 800 rpm alongside the standard 600) caught a UV lot last spring that would have cleared the single-point check clean.
The seven-parameter gate in INK-QC-04 gets complicated fast when you’re switching to bio-based ink systems — we moved a folding carton line to vegetable oil-based offset inks in late 2023 to hit our Sedex requirements, and three of those thresholds had to be revalidated from scratch because the reference standards were set on petroleum-based formulations. The tack range especially; 40–65 Pa·s just doesn’t map cleanly onto linseed-alkyd vehicles at the same temperature.
Running a tighter incoming QC gate like INK-QC-04 actually paid back in reduced press downtime for us — we’d been absorbing roughly 2–3 makeready hours per rejected run before we formalized the protocol, and at our converter’s press rate that was eating around $320–$410 per incident. The reject-at-intake model stings on lab labour cost (we added about 0.4 FTE to cover it) but net savings across a 12-month folding carton programme came out around $18k once we stopped paying to find problems at press.