Cure presses run the plant. Everything else has to match the cycle.
Cure presses are the bottleneck. A 12 to 25 minute cycle per cured unit across eight presses, each running a different tread spec against a different mould against a different compound batch. The constraint moves weekly between mixing, curing, and finishing. When promise-date math uses an average press cycle instead of the per-press cycle, customer commitments slip on press P-208 every week. When mould M-118 runs 400 cycles past its 8,000-cycle refurb interval, tread depth on units 8,101 to 8,400 measures soft and shows up as a field-failure warranty replacement six weeks later. Polymr keeps the press log, the mould refurb counter, the compound mix-step plan, and the tread-wear QC gate on the same loop.
Built for the tires and rubber shape.
Polymr reads compound-mix recipes into batch genealogy with downstream-product trace per consumed lot. Compound B-2284 ties forward to every WO that drew from it, every cure-press shot that ran the resulting tread, and every shipped pallet, in one query. A field-complaint walk-back lands the affected lot list in minutes rather than the next morning.
The cure-press capacity heatmap reads across bays and slots so the press lead sees Bay C saturated before the next batch lands at the dock. Original-equipment versus aftermarket priority routing rides on every WO so an OE-tied tread gets the next bay slot ahead of a stock-replenishment run when the schedule contends.
Six failure modes that govern a compound-and-cure-driven plant.
Each is what a production planner, a maintenance lead, or a shift supervisor will name when asked why this week's schedule held and last week's slipped. And why promise-date integrity sits at 81 percent on the customer scorecard.
- Cure-press scrap from compound driftCure press P-203 hits 14% scrap on tread spec T-44 because compound batch B-2284 ran 1 point outside the carbon-black spec
B-2284 of compound CMP-217 mixes at 13:42 Tuesday with a measured carbon-black loading of 22.1% against the v4 recipe spec of 23.0% ± 0.5%. The bench QA flags the drift on the post-mix sample at 14:30 but the batch is already staged at press P-203 for the T-44 tread cure run. The shift supervisor decides to push the batch through anyway against the customer commitment; the cure-time math against the under-spec compound runs 30 seconds short of full cure and the tread surface measures soft on units 8101-8480. Scrap rate against the T-44 spec runs 14% for the day. The compound flag did not block the press.
- Mould overrun past refurbMould M-118 due for refurb at 8,000 cycles runs to 8,400 before the flag and defective tread surfaces on units 8101-8400
M-118 carries an OEM-recommended refurb interval of 8,000 cycles before mould-cavity wear begins to affect tread-pattern depth on cured units. The maintenance system at the plant logs cycles against the mould but the flag-to-supervisor handoff is a weekly review. M-118 crosses 8,000 cycles at 09:42 Wednesday during second-shift run on press P-208; the flag does not surface until the Friday maintenance review at 11:00. The 400 units cured between 8,001 and 8,400 carry tread-depth measurements 0.4 mm shallow against the T-44 spec, surface as field failures across two customer accounts over the following six weeks, and pull a 1,280-unit warranty replacement.
- Compound supplier viscosity driftCompound supplier V-218 ships a carbon-black lot tackier than spec and downstream cure cycle drifts 30 seconds off
V-218 ships lot V218-CB-441 of carbon black against PMR-CB-204 at standing weekly volume. The supplier COA reads in-spec on the assay sheet but the actual viscosity profile at the mix step runs 6% tackier than the rolling 12-week baseline. The plant mix step accommodates by extending the mix cycle 90 seconds. The cure cycle downstream, however, calibrates against the standard compound viscosity and runs 30 seconds short of the optimal cure window for tread spec T-44. The press OEE numbers look fine for the week but the scrap rate ticks up 2 points and the operator narrative blames the press, not the upstream compound drift.
- Promise-date math against average press cyclePromise-date math uses the average press cycle and misses bottleneck press P-208 saturation
The promise-date engine at sales takes the average press cycle across the eight cure presses (mean 17.4 minutes) and computes customer commitment dates against that mean. Press P-208. the larger-mould press dedicated to the T-44 family. runs a 21.8-minute cycle and is the bottleneck for that customer commitment family every week. The mean-based math overpromises by 25% on P-208-pegged customer commitments; the saturation surfaces as missed dates rather than a planning view. Polymr re-computes promise dates against the per-press cycle and surfaces the P-208 saturation Sunday night for the operator to re-price the at-risk commitments.
- Cure log entry lagCure log entries are manual and the 11-minute lag between press end and ERP update means the planning view is always behind
Each cure cycle ends with the operator scanning the press out-tag and entering the cycle count, the batch ID, the mould ID, and the cure-window actual into the ERP terminal at the line. The walk from the press to the terminal, the entry itself, and the supervisor sign-off run an average 11 minutes per cycle. With eight presses running 24/6, the cumulative ERP lag puts the planning view 90 minutes behind reality at any given moment. Polymr ingests the press out-tag scan directly from the line, eliminates the terminal walk, and brings the lag to under two minutes.
- Tread-wear QC hold reboundTread-wear QC sample fails on 1 of 9 dimensions and the full lot is held pending re-test, the planning view replans against nothing
The end-of-cure tread-wear QC samples 9 dimensions per cured unit (tread depth, sidewall thickness, bead diameter, six others) against the customer spec. Lot L-3318 of T-44 produces 480 units; the 9-dimension QC sample fails on tread depth on 1 of 24 sampled units. Per the customer spec the full lot is held pending a 96-unit expanded re-test. Pre-Polymr, the planning view does not know the lot is held and continues to allocate it against the open customer commitment for two days. The re-test passes; the lot releases late; the customer commitment slips one day. With the hold surfaced in real time, the planner re-allocates an alternate lot inside the hour and the customer commitment holds.
The cure-press queue, four bays, eighty minutes ahead.
Bay C, running press P-208 against mould M-118 on the T-44 large cycle, is the bottleneck this week. Two consecutive 21.8-minute cycles plus a cooldown saturate the bay before the next load slot. Bay D enters a planned preventive-maintenance window for Press 4. Each cell carries the WO, the tread spec, the cycle time, and the mould; the queue is the artifact the press lead and the planner share.
| Bay | 00-20 min | 20-40 min | 40-60 min | 60-80 min |
|---|---|---|---|---|
Bay A P-201 | WO-77041 T-44 / 14.2 m m M-118 | WO-77041 T-44 / cooldown m M-118 | WO-77051 T-22 / load m M-090 | WO-77051 T-22 / 14.8 m m M-090 |
Bay B P-203 | WO-77042 T-22 / 14.2 m m M-101 | WO-77042 T-22 / 14.2 m m M-101 | WO-77052 T-22 / load m M-101 | WO-77052 T-22 / 14.2 m m M-101 |
Bay C P-208 | WO-77043 T-44 large / 21.8 m m M-118 | WO-77043 T-44 large / 21.8 m m M-118 | WO-77043 T-44 / cooldown m M-118 | WO-77053 T-44 / 17.6 m m M-204 |
Bay D P-210 | idle | MAINT Press 4 PM cycle m . | MAINT Press 4 PM cycle m . | WO-77054 T-22 / load m M-090 |
Compound lot B-2284 traces to three downstream WOs and one of them is at risk.
The card below is the operating view on B-2284 of CMP-217 v4. Bench QA flagged the post-mix viscosity at +6 percent against the 12-week baseline. The compound is in-spec on the assay sheet but the profile drift opens cure-window risk on WO-77043 because the T-44 large cycle is the most sensitive of the three downstream WOs. The card surfaces the re-allocation against alternate lot B-2286 before the WO releases.
Five work orders, one floor, one row each.
WO-77043 is on hold against the B-2284 compound flag. WO-77046 has already re-pegged against alternate lot B-2286 and is queued behind the changeover on press P-210. WO-77049 is at 88 percent against plan and lands inside the shift. The shift supervisor reads the same row the planner reads.
What this looked like at a two-plant industrial rubber maker.
- Situation
- A handful of mixing-and-curing presses ran 24/6. Compound batching, mould changeovers, and cure-time variance were the top three throughput governors. Throughput dictated promise-date integrity.
- What was breaking
- Bottleneck shifted weekly between the two curing presses and the mixing line, but the planning team rediscovered it on Tuesday standup rather than seeing it Monday morning. Compound recipe revisions were not propagated to open POs; out-of-spec ingredient lots reached the floor 8% of the time.
- Margin and bottleneck analysis
- Engineering revisions
- Planning + purchasing
Three workflow families do the heavy lifting on a cure-press-driven floor.
- Margin + press bottleneck
Live per-press cycle vs capacity. Constraint-shift surfaces Sunday night so customer commitments re-price before the W22 schedule publishes. Promise-date math uses per-press cycle, not the mean.
- Compound + recipe revisions + purchasing
CMP-217 v3 to v4 recipe rev diffs against open POs and WIP, every affected mix-step lot updated, cure-window field on the routing propagates to the press scheduler. Net requirements refresh as compound revisions and cure-window changes land.
- Compound drift + QC hold recovery
Inbound supplier viscosity drift or end-of-cure tread-wear QC hold surfaces affected customer commitments and the alternate-lot re-allocation path inside the hour rather than after two days of unresolved hold.
Three siblings carry the same cycle-and-mould shape.
- Plastics manufacturing
Mould cousin. injection moulding and tire cure presses share the same mould-cycle refurb tracking, the same per-mould cycle-time math, and the same changeover-vs-throughput trade-offs.
- Batch process
Batch cousin. compound mixing is a batch step with reactor scheduling, yield variance, and weekly plan stability identical in shape to the broader batch-process world.
- High-volume manufacturing
Cycle-driven sibling. takt-driven high-volume plants and cure-cycle-driven tire plants share the morning-queue discipline and the bottleneck-shift dynamic on a shorter cadence.