TL;DR: Flat pouch and sachet failures almost always trace back to tolerance stackup errors introduced during CAD handoff — not to film grade selection or sealing parameters.
TL;DR: A ±0.5mm dimensional tolerance on each of four laminate layers compounds to ±2.0mm total, which is enough to misregister a tear notch by a full tooth and cause consumer opening failures at scale.
Where Flat Pouch CAD Files Break Down Before Production Starts #
A brand sends us a structural dieline for a three-side-seal sachet. The file looks clean: 80mm × 120mm finished dimensions, 8mm seal borders, centered tear notch at 30mm from the top seal. The graphic file registers correctly in prepress. But when we run the first 500-piece pilot on our BOBST SX 165 lamination line, the tear notch is landing 3.5mm off-center on roughly 40% of the pieces.
The problem wasn’t the file. The problem was that the dieline was drawn as a single-layer flat geometry — it didn’t account for the cumulative dimensional shift that happens when four substrate layers (PET/adhesive/AL foil/CPP, in this case) go through three adhesive lamination passes, each introducing its own registration tolerance, and then get wound on a tension-controlled unwind at slightly different caliper thicknesses across the web.
This is the geometry-to-reality gap that CAD integration for flexible packaging has to close. Unlike rigid boxes — where a 2.5mm greyboard panel has a predictable, isotropic behavior — flexible laminate structures behave differently in the machine direction versus the cross direction, and that anisotropy must be built into your master template before the first artwork pixel is placed.
We created what we internally call the FP-DFM01 constraint matrix: a structured input form our applications team completes for every new flat pouch or sachet project before we open a dieline in Illustrator or ArtiosCAD. It forces us to lock in laminate stack, web tension spec, and seal jaw geometry before any structural dimension is committed to a file.
The Parameters That Govern Dimensional Integrity in Flat Pouch Structures #
Four engineering inputs dominate flat pouch dimensional behavior, and they interact.
Film caliper and ply count. A standard two-ply PET12/PE80 sachet laminate runs at roughly 92–95 microns total caliper. Upgrade to a four-ply PET12/AL7/PET12/CPP50 structure and you’re at 150–160 microns — and that extra thickness changes the minimum bend radius around a seal jaw, which means your seal border width minimum changes too. We specify a minimum seal land of 6mm for two-ply structures, increasing to 8mm for four-ply foil laminates, because anything narrower risks the seal jaw crimping through the outer layer on tight radii corners.
Machine direction vs. cross-direction elongation. Biaxially oriented PET (BOPET) has an MD elongation at break around 120–130% and CD elongation around 100–115% per ASTM D882. CPP seal webs run higher: MD 400–600%. These numbers feed directly into how much your finished dimension drifts from your dieline dimension under web tension. On our forming lines we compensate by scaling CAD templates 0.3–0.5% in the machine direction as a baseline correction. We don’t apply this correction for CD dimensions because transverse tension is controlled differently.
Adhesive thickness and open time. Solvent-based polyurethane adhesive applied at 3–5 g/m² adds 3–5 microns per lamination pass. On a four-ply structure that’s 9–15 microns additional caliper. The more consequential factor is adhesive cure: if a laminate is slit and pouched before full cure (typically 40–50 hours at 45°C per our internal bond development protocol, based on the adhesive supplier’s TDS), the PU layer is still slightly compliant and the bond shear strength will be 15–25% below its 72-hour rated value. We track this under our internal QC-F12 cure release log, and nothing leaves our lamination store for pouching unless the lot timestamp clears the hold window.
Seal jaw temperature, pressure, and dwell time. These three form an interdependent triangle. For CPP inner webs, our standard starting point is 160–175°C jaw temperature, 0.3–0.4 MPa jaw pressure, 0.4–0.6 seconds dwell, tested against a target seal strength of ≥30 N/15mm per ASTM F88 under our standard T-peel protocol. If a brand partner asks us to reduce jaw dwell to increase line speed, we rerun the seal integrity validation before confirming — the relationship between dwell and peel strength is not linear and drops off sharply below a threshold that varies by web thickness.
| Parameter | 2-Ply PET/PE Structure | 4-Ply PET/AL/PET/CPP Structure |
|---|---|---|
| Total laminate caliper | 92–95 µm | 150–160 µm |
| Minimum seal land width | 6 mm | 8 mm |
| Recommended jaw temp (CPP) | 140–160°C | 160–175°C |
| Target seal strength (ASTM F88) | ≥25 N/15mm | ≥30 N/15mm |
| MD template scale correction | +0.3% | +0.5% |
The most commonly overlooked parameter in this list is the MD scale correction. Graphic designers working from a brand-provided dieline never apply it — they shouldn’t have to. But if our prepress team receives a dieline and merges it directly into the artwork file without applying the substrate-specific correction factor, every positional element that matters for function (tear notch, zip track cutout, registration target for clear window laminate) will drift.
Decision Framework — Matching Design Constraints to Actual Run Conditions #
If your sachet is a single-serve food application filled on a vertical form-fill-seal (VFFS) line with a target fill weight ≤ 10g, the dominant constraint is VFFS forming tube clearance, not laminate caliper. At that fill weight, the tube internal diameter typically runs 28–35mm and the film has to conform around it without wrinkling at the back seal. We specify a minimum film stiffness of 150 mN (Gurley) for sachets running on VFFS forming tubes, because films below that threshold crease into the back seal jaw and produce intermittent weak seals. A PET12/PE60 at 72 microns barely passes that threshold; swapping in PE80 or adding an LDPE flood coat gets you to 190–220 mN comfortably.
If your application involves a tear notch that must function reliably without a notch-press tool (i.e., laser-scored or die-cut on the pouch press), the design constraint shifts to film orientation and notch geometry. Laser scoring on BOPET works consistently at 0.3–0.5J/cm² energy density, creating a groove depth of 20–30% of film thickness without penetrating through. For AL foil laminates, we do not recommend laser-only tear initiation — the foil layer reflects energy inconsistently, and in our experience the notch-to-tear propagation direction across a foil structure is unreliable unless backed by a physical micro-perforation. This is a point where practices vary: some converters accept laser-only on foil sachets; our position, based on observed tear-test variance on 12 production lots, is that mechanical notch-press remains the standard for any foil-containing laminate.
If the project requires print registration between front and back panels to align at the gusset fold (an effect used by several skincare brands we work with), the CAD template needs to carry explicit panel-fold compensation. The standard flat geometry misaligns by 1.5–2.5× the laminate caliper for every 180° fold — for a 155-micron four-ply structure, that’s 0.23–0.39mm of deliberate offset you need to build in to make the front-to-back print appear registered at the edge. We embed this offset calculation in our FP-DFM01 constraint matrix and verify it during color proof approval, not during press makeready.
For any pouch targeting recycling compatibility under current EU PPWR guidelines or aligned with APR Design Guidelines for Recyclability, the structure selection decision narrows significantly. All-PE structures (BOPE/MDOPE/LLDPE) are currently the most straightforward path to mono-material recyclability, but all-PE laminates have lower stiffness, higher haze, and typically require recalibrating jaw temperature downward by 15–20°C compared to PET-based structures. If a brand brief specifies both high barrier (OTR ≤ 1.0 cc/m²/day) and recyclability, an all-PE structure won’t reach that barrier threshold without a ceramic oxide (AlOx or SiOx) vacuum-deposited coating, which adds roughly 15–20 working days to the laminate sourcing lead time on our current AVL.
Specification Notes for Brand Partners #
When you brief us on a new flat pouch or sachet project, the information that makes the biggest difference to first-sample accuracy is: your filling equipment make and model (or filling method if it’s manual), the intended fill weight or fill volume range, and the consumer opening method you’re designing for (tear, zipper, spout, none). Without the filling equipment spec, we’re guessing at forming tube geometry and web tension settings — and that guess is often wrong the first time.
The single most common brief gap we see is artwork files supplied without embedded laminate structure information. Graphic designers often supply a flat-format dieline (PDF or AI) that matches finished bag dimensions, but carries no indication of which edge is the machine direction on the web. For directional features — tear notches, textured finishes, registered die-cuts — this gap causes at least one extra sample iteration. Supply us with a substrate specification sheet alongside your artwork file and we can eliminate that round.
Our standard sampling timeline for a flat sachet or three-side-seal pouch with a new laminate structure is 18–22 working days from confirmed spec to physical sample. For projects reusing a laminate structure already on our AVL, that timeline compresses to 10–14 working days. Rush sampling below 10 working days is possible for two-ply commodity structures but not for four-ply or specialty coated webs.
Is ±0.3mm print register achievable on a flat pouch production run?
On our gravure printing lines, our standard register tolerance is ±0.2mm in the machine direction, verified by inline camera at 100% inspection. Cross-direction tolerance is ±0.3mm. These hold consistently on BOPET-based webs; on all-PE webs, CD register opens slightly to ±0.4mm because of higher transverse elongation under printing tension. If your design has graphic elements that must register across a fold line, the relevant number is not the press register tolerance — it’s the fold compensation offset described above.
What film structure do you recommend for sachets with OTR requirements below 0.5 cc/m²/day?
It depends on whether the brief includes recyclability requirements. For non-recyclable structures, a PET12/AL7/CPP50 laminate will reach OTR of 0.01–0.05 cc/m²/day at 23°C/50%RH, well inside that threshold. For recyclable structures, you’d need an AlOx-coated BOPE or SiOx-coated BOPE layer, which brings OTR to 0.3–0.8 cc/m²/day — borderline for that spec, and it depends on how aggressive the product headspace environment is. We’d want to review product water activity and storage conditions before committing to a structure.
What’s the minimum order quantity for a custom-structure flat sachet?
For a three-side-seal sachet in a laminate structure already on our approved vendor list, our standard MOQ is 50,000 units. For a custom laminate structure (new film grade, new barrier coating, new ply combination), the minimum laminate run typically requires 500kg of printed web — how many pouches that yields depends on format size, but for a 60mm × 90mm sachet it’s approximately 120,000–150,000 units. Going below that doesn’t make economic sense because setup and lamination cost per unit rises steeply.
Can you integrate our existing dieline file from another supplier?
Yes, with a caveat: we treat every externally supplied dieline as a starting point, not a production file. We run every incoming dieline through our FP-DFM01 constraint check before it touches our prepress workflow. This often surfaces substrate-specific corrections (MD scale, seal land width, fold compensation) that the original file doesn’t carry. We’ll flag those to you before making changes — we don’t modify structural dimensions without written approval.
Do tear notch depth and position specifications need to be re-validated if we switch from foil to metallized film?
Yes, and this is a point where we’d push back on treating it as a simple material swap. AL foil and vacuum-deposited metallized film (typically 500–700 Å aluminum on BOPET) behave differently under notch initiation: foil has a defined yield point and tears cleanly in the transverse direction; metallized film has a thinner, more brittle deposition layer that can flake rather than propagate cleanly if the notch geometry isn’t adjusted. We re-run tear propagation testing per ISO 6383-2 on any laminate structure change that involves a barrier layer swap, regardless of whether finished dimensions are unchanged.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.
The 3.5mm notch drift on 40% of pilot pieces tracks with what we’ve seen, though in our case the bigger culprit was unwind tension variation on the AL foil layer specifically — we run a 4-ply PET/AL/PET/CPP on our vertical sachet line and had to set foil unwind tension 15–18% lower than the PET webs before the caliper-driven registration stabilized. The machine direction vs. cross direction anisotropy point is accurate, but it’s worth noting that foil-containing structures also introduce a yield-point variable that purely polymer plies don’t have, so the constraint matrix probably needs a separate row for metallic interlayers.
The 3.5mm notch drift on the BOBST pilot run is exactly what bit us on a 75mm × 110mm lavender sachet we ran in late 2022 — we’d been using a single-layer dieline template for years without issue, then switched to a PET/AL/PET/CPP structure for a retail chain’s private label line and suddenly 30-odd percent of pieces were failing tear initiation. Took us two production runs to isolate that the CPP layer was unwinding at a slightly higher tension than our previous 2-ply, which was compounding the cross-direction shift right at the notch tooth. We’ve since locked our minimum notch-to-seal-edge distance at 9mm on any 4-ply structure, but honestly that’s a workaround, not a fix.
The 40% notch misregister rate on that pilot run tracks — we saw something similar on a 4-ply PET/AL/PET/CPP sachet last year, except ours was landing 2.8mm off after lamination on a Nordmeccanica line, and the root cause traced back to the CPP layer alone accounting for 0.4mm of that drift in the cross-web direction at our standard 4 N/m² wind tension.
Switching from the PET/AL/CPP stack to a mono-material PE structure (we trialled Mondi’s BarrierPack Recyclable last year) made the tolerance stackup problem significantly worse — the caliper consistency across a 4-ply mono-PE build is nowhere near what you get with AL foil acting as a dimensional anchor, so our tear notch misregistration went from acceptable to a full requalification event. We didn’t anticipate that the recyclability gain would essentially reset our FDM constraints from scratch.
We caught something similar building our master template in ArtiosCAD — if you don’t lock the tear notch position relative to the finished bag edge rather than the unwind centerline, any cross-web caliper drift on the CPP layer will walk that notch every single run, and you won’t see it until pilot.
The 6mm minimum seal land on the 2-ply PET/PE structure holds in most cases, but we’ve found that anything running through a rotary jaw sealer (we use a Totani BH-60 in our Osaka facility) needs at least 7.5mm once you factor in the jaw parallelism drift that accumulates after roughly 800,000 cycles. At 6mm you start seeing incomplete fusion at the seal edge intermittently — not a catastrophic failure rate, but enough to show up in post-market QC on units that passed inline.
One thing the caliper discussion doesn’t touch on is cross-web caliper variation within a single roll — we measured a 4.2 µm range across the 600mm usable web width on a PET/AL/CPP rollstock from our primary converter last quarter, and that alone was enough to push our cross-direction seal land below the 8mm minimum on the operator-side edge without triggering any mid-run QC flag.
The anisotropy point is the one that keeps catching brand-side teams off guard — we had a structural engineer from our Paris fragrance division insist the CPP orientation didn’t matter on a 80mm × 65mm single-dose pouch until we showed him the MD vs CD elongation delta at seal temp.
Catching the tolerance stackup issue at the DFM stage rather than after a pilot run saves more than people realize — we were absorbing roughly $1,200–1,400 per pilot run in wasted rollstock and operator time before we built a proper constraint checklist into our quoting workflow, and that’s on a mid-volume sachet around 50k units/month.
The geometry-to-reality gap showed up for us in a different way — our Shenzhen converter was building the dieline from the finished bag spec we sent them, but nobody had communicated that the CPP layer on our 4-ply structure ran at a slightly tighter web tension on their line than the PET outer, so the whole substrate was arriving at the sealing station with a consistent 1.8mm cross-web bow that their QC team had just… normalized over time. Took us three pilot runs and a plant visit in Q1 2023 to figure out it wasn’t a notch placement issue at all but a tension profile mismatch that had been quietly baked into every dieline they’d built for us for two years.
The geometry-to-reality gap hitting at pilot stage is the painful version — we’ve started requiring converters to return a completed substrate spec sheet before we even open the dieline, which added about 4 working days to our pre-production intake but cut our second-sample cycle from an average of 3 rounds down to 1.5 across 11 sachet SKUs we launched out of our Guangzhou co-packer in Q3 last year.