We built a four-cavity tool for a 500 ml PET thin-wall tub with a 0.75 mm nominal wall, holding rim flatness to 0.10 mm and balancing fill across all four cavities to within 1.8%. First article passed on the second trial. The deciding factor was not cooling water; it was hot-runner nozzle sizing and gate drop diameter.
What makes PET thin-wall different from any other container
Thin-wall packaging looks simple: a tapered box, a rim to seal a lid, no cosmetic A-surface.
It is the hardest category of tool we quote, and the difficulty is arithmetic, not opinion.
A 500 ml tub with a 0.75 mm wall has a flow length of roughly 145 mm to the last corner, a flow-length-to-thickness ratio near 193:1, where a 2.0 mm housing sits near 70:1.
PET also has a narrow processing window. Amorphous PET softens at a low temperature and crystallises quickly above it, so a melt that is 8 °C too cold fails to fill and a melt that is 10 °C too hot starts to haze.
The result: the practical window is roughly 265–280 °C, and inside it the only variable left to tune is pressure and balance.
This is why packaging tools are quoted differently: a 30 mm change in runner length matters more than a 1 mm change in wall.
The challenge: five constraints that argued against each other
The customer is a European food-packaging converter who fills and seals in-house, already running the part on two suppliers’ tools. The drawing set five hard constraints:
- Wall: 0.75 mm nominal (1,260 mm² of projected sidewall plus 1,900 mm² of base). The customer would not accept more, because resin cost is the dominant cost line at this volume.
- Rim: a 4.2 mm wide sealing flange, flat within 0.10 mm around the full perimeter. The lid peel-strength test at the filler fails if the rim is out by more than 0.15 mm, because the seal head cannot compensate for a tilted flange.
- Balance: top-to-bottom fill weight difference under 2.0% across four cavities. Above that, the filler’s checkweigher rejects and the customer eats the loss.
- Volume: 4.5 million parts a year on a 24/7 operation, which on a four-cavity tool means a 13 s cycle, run continuously.
- Tooling budget and date: $46,000 cap, T1 in 9 weeks.
The old tool had failed on balance, not on shape. Cavity 3 ran 6.4% heavier than cavity 1, and cavity 2 short-shot at the far rim corner on roughly 400 ppm. The converter had already raised melt temperature to compensate, which produced hazy rims on cavity 2 and pushed rejects to 1.9%.
What we did: four decisions, and the third one was the fix
1. Choose 1.2343 ESR and a hardened rim insert, not a uniform 1.2738 block
At 4.5 million cycles a year, this tool is designed for a two-million-shot service life before any major rework, and we expect the rim edge to be the first wear point.
The cavity and core plates are 1.2343 ESR hardened to 46–48 HRC. The rim area is a separate insert in the same steel, installed and preloaded so it can be replaced in one working day without re-cutting the cavity.
A pre-hardened 1.2738 block at 30–34 HRC is cheaper by about $3,800 and is the right answer for 500,000 parts. At this volume the rim edge polishes away instead.
Surface quality matters too: acetaldehyde traces from a corroded cavity surface shorten shelf life, and the customer’s spec capped them. Hardened 1.2343 ESR, polished to SPI A2 on sealing areas and SPI A3 elsewhere, is the working compromise.
2. Size the hot-runner nozzles for pressure, then verify with Moldflow
The old tool used 2.0 mm nozzle orifices — generous on a 2.0 mm part, and the primary cause of imbalance on a 0.75 mm PET wall.
Pressure drop through a circular channel scales with the fourth power of the diameter, so a small difference in nozzle bore becomes a large difference in the pressure actually delivered to each cavity.
We went to 2.5 mm orifices and 3.6 mm runners in a naturally balanced layout: four equal runner lengths from a central sprue, no balancing inserts.
Moldflow predicted the last-filled point at 2.6 MPa below the maximum injection pressure available on the customer’s 260 tonne press, and predicted cavity-to-cavity fill difference of 1.4%.
That margin is deliberately thin: on a thin-wall PET tool the only remaining cycle lever is melt temperature, so pressure headroom has to be designed in.
3. Drop the gates into the base fillet, and use a longer gate land
This is the correction that closed the balance gap, and it came from the Trial 1 data.
On Trial 1 we had gated on the base centre with a 0.8 mm gate and a 0.5 mm land. Fill was 3.1% out of balance, and cavity 2 short-shot at the rim corner whenever melt temperature went below 268 °C.
The problem was not the gate diameter; it was gate freeze-off. A short land lets the gate freeze while the cavity is still filling, so each cavity sees a different effective packing pressure depending on tiny differences in gate temperature.
We moved the gates onto the 1.5 mm radius fillet where the base meets the sidewall, so the flow enters the wall at full thickness instead of turning a 90° corner in a 0.75 mm section.
The gate diameter went to 0.9 mm with a 1.2 mm land, and we added a 0.4 mm cold-slug well behind each gate.
Balance went from 3.1% to 1.8%. Cavity 2 no longer short-shots, and the melt temperature window opened from about 8 °C to 14 °C.
The gate vestige sits inside the base recess and is not visible in the filled pack, which mattered because the customer’s print runs on the tub side, not the base.
4. Cool the rim separately from the body, and accept the cycle cost
A 0.75 mm wall cools in a few seconds; a 4.2 mm rim is nine times thicker in section and dictates the cycle.
We used two independent circuits: baffled circuits following the body profile at 14 °C, and a separate rim circuit at 9 °C with a 12 mm baffle blade pushed to within 8 mm of the sealing edge.
The separation is what lets the rim be pulled harder than the body without creating differential shrinkage that would tilt the sealing flange.
Cost: the rim circuit adds about 1.6 s of cycle, taking the tool from 11.2 s to 12.8 s at the customer’s press.
We wrote that trade-off into the quotation rather than hiding it, because at 4.5 million parts a year 1.6 s is roughly 2,000 machine hours. The customer accepted it because the alternative was a rim that fails peel-seal testing at 1.9% reject.
The result: two trials and the release numbers
Trial 1 ran as first cut and confirmed the fill pattern but failed balance. Trial 2 implemented the gate relocation and the rim circuit split, and the tool was released.
The table below compares the old tool (as documented by the customer), Trial 1 and the released tool. Dimensions are from CMM reports on 30 parts per cavity at 24 hours; weights are from the customer’s checkweigher over three production shifts.
| Parameter | Old tool | Trial 1 | Released | Spec |
|---|---|---|---|---|
| Nominal wall, sidewall | 0.75 mm | 0.76 mm | 0.75 mm | 0.75 ±0.03 |
| Cavity-to-cavity fill weight spread | 6.4% | 3.1% | 1.8% | ≤2.0% |
| Rim flatness, full perimeter | 0.21 mm | 0.14 mm | 0.08 mm | ≤0.10 mm |
| Shot weight, 4 cavities | 27.6 g | 26.4 g | 26.2 g | — |
| Short shots, far rim corner | 400 ppm | 90 ppm | 0 ppm in 120,000 | — |
| Peel-seal rejects at filler | 1.9% | 0.6% | 0.1% | ≤0.3% |
| Cycle time at customer press | 12.1 s | 12.8 s | 12.8 s | ≤13 s |
| Melt temperature window | 7 °C | 8 °C | 14 °C | — |
| Peak injection pressure | 148 MPa | 131 MPa | 124 MPa | — |
| Cpk on rim flatness | 0.68 | 1.09 | 1.47 | ≥1.33 |
| Cavity plate hardness | 32 HRC | 47 HRC | 47 HRC | — |
| Shot life to first rework (projected) | 600,000 | — | 2,000,000 | — |
Tooling landed at $44,600 against the $46,000 cap, and T1 shipped on day 61 against the 9-week (63 day) commitment. Customer approval came on day 68.
One thing did not go to plan and is worth reporting honestly: the first polishing pass on the rim insert took 11 hours against an estimated 6, because the gate land change left a heavier witness mark in the base recess than we expected. That is a one-time cost inside a $44,600 tool, not a defect, but it is why the release was day 61 and not day 55.
Projected life to first rework is 2,000,000 shots based on the hardened 1.2343 ESR and the replaceable rim insert, assuming the customer’s standard maintenance interval. Life depends on steel grade and maintenance; a converter running abrasive-filled regrind at 100% should expect less.
What carries to your packaging project
If you are quoting a thin-wall container, these five points are the short version of what this project cost us to learn.
- Gate land, not gate diameter, sets balance on a thin-wall part. A short land lets the gate freeze mid-fill, and then every cavity sees a different packing pressure. Start at 1.2–1.5× gate diameter of land on PET below 1.0 mm wall.
- Gate into a fillet, never into a corner in the thin section. Entering the wall at full thickness removes the pressure loss that produces the short shot at the far rim corner.
- Size nozzles for the pressure you are willing to spend, not for the bore you have in stock. Going from 2.0 mm to 2.5 mm orifices cost about $1,900 on a four-drop manifold and bought 24 MPa of pressure headroom and a 2.7 percentage-point balance improvement.
- Budget the rim as its own cooling problem. On any container with a sealing flange thicker than twice the wall, a dedicated rim circuit with a baffle within 10 mm of the sealing edge is not optional, and it will cost 1–2 s of cycle. Quote it as a line item.
- Specify hardness by shot count, not by habit. Below 500,000 shots a pre-hardened block is correct. Above two million, a 46–48 HRC ESR cavity with a replaceable wear insert is cheaper than a rebuild.
Our answer on PET thin-wall tooling
Our answer: on a PET container below 1.0 mm wall, design the tool around gate freeze-off and rim cooling, not around cycle time. A gate that lands in a fillet with a long land and a nozzle sized for pressure headroom will hold 2% balance without temperature tricks, and it will keep holding it after a million shots.
RCH Plastic reference — this tool was designed and built at our Dongguan shop with mold making in-house, molding through vetted partners, which is why the sequencing above was decided at the drawing stage rather than on the press.
The tool ran on a four-cavity 1.2343 ESR base, 2.5 mm hot-runner orifices, 0.9 mm gates with 1.2 mm land, and a separate 9 °C rim circuit. T1 delivery included trial samples, photos, video, the full process parameter sheet and a CMM dimensional report.
If you have a thin-wall container that is running out of balance, send the part drawing and the current process sheet. Our engineers will return a written DFM report in three working days at no cost, and it will name the gate geometry we would change before any steel is cut.
Related guides in this series
- Mold design and DFM guide — the hub for the design decisions behind this tool
- Injection molding cooling system design — circuit layout and baffle placement
- Multi-cavity mold design — balance, runner layout and cavitation economics
- Injection mold gate design FAQ — gate types, sizing and lands
- Defects and quality guide — short shots, warpage and dimensional drift
- Mold steel comparison — choosing hardness by shot count
- Injection mold tooling cost — what drives the quote on a packaging tool
- Case study: LED lighting housing — the opposite problem: a textured part with an optical tolerance
- Case study: heat sink housing — balancing a filled resin around an insert
Send your 3D file, wall section and fill weight specification to our engineering team and we will tell you where your gate geometry is going to lose balance.
