Quick answer first
Emerald, a major jewelry manufacturer in India serving domestic retail and South Asia export markets, deployed the Cylanco CXM-C20 and reported a 50% improvement in fully automated casting efficiency across three consecutive production quarters. Unlike the LAMAR deployment, Emerald reached this outcome with a single-machine architecture — made possible by a narrower product focus and a tighter upstream workflow discipline. The operational change that carried most of the gain was not on the machine. It was the decision to standardize the preparation-to-pour sequence and enforce same-shift defect closure. This case documents what Emerald changed, why a single CXM-C20 was enough, and which operational habits other subcontinent workshops can port directly.
Case context
Emerald serves the Indian domestic retail market with additional export flow into Bangladesh, Nepal, and UAE. Their product mix is weighted toward 18K and 22K solid and semi-hollow pieces, with filigree representing a smaller share. Monthly throughput is volume-oriented rather than fine-detail-oriented, which shaped both the machine choice and the workflow strategy.
Before the CXM-C20 deployment, Emerald ran an older casting line where cycle time variance between shifts was the primary complaint. Quality was stable on any given lot, but predictability across lots was not, and the mismatch between production plan and actual output created downstream friction with the finishing and QA departments. In a volume-focused factory, inconsistent lot cadence is arguably more expensive than a slightly higher defect rate — finishing teams overstaff to absorb variance, polishing schedules slip, and export order commitments become harder to guarantee.
Why a single CXM-C20 was enough
Unlike LAMAR's two-machine filigree/solid split, Emerald's product mix sat inside a narrower envelope that the CXM-C20 could handle as a single-profile line. The decision to skip a second machine was deliberate: adding a second casting line would have introduced inter-machine variability that Emerald did not want to absorb while the workflow discipline was still being built. Their prioritized sequence was "process discipline first, then scale" — which meant a single well-understood machine was more valuable than two machines with two drift profiles.
This is a useful counter-example for workshops that assume more equipment automatically means more output. For volume production inside a narrow product envelope, the efficiency ceiling is typically set by operator discipline and upstream timing, not by machine count. Emerald's 50% efficiency gain on a single machine is evidence that adding hardware before exhausting process headroom is often premature.
Before optimization — the starting baseline
- Shift-to-shift cadence varied even with identical product families, alloy lots, and investment lots. The variation wasn't large per shift, but compounded across a week it broke the production plan.
- Defect diagnosis ran on individual expertise. A handful of senior operators carried the institutional knowledge; when they rotated off or took leave, defect close times stretched by a factor of two.
- Handoff between investment, burnout, and pour was informal. Operators "knew" when a flask was ready, but that judgment lived in heads, not in writing. New operators took months to build the same instinct.
- Cycle time was measured at the lot level, not at the sub-step level. The workshop knew how long a lot took but could not see which specific sub-step was stretching or compressing across shifts.
- Defect coding was weekly, not daily. By the time a pattern was visible, two or three days of production had already reinforced the bad habit.
Execution strategy — what Emerald changed
1. One approved process sequence per product family
Emerald committed to a single validated sequence per product family — solid 18K, solid 22K, semi-hollow — and required any deviation to be logged as a change request rather than a silent adjustment. This is the same first-principle LAMAR applied, and it is the change that carries the largest single productivity gain across most workshop case data.
2. Frozen change scope during validation
During any process validation or recovery window, only one variable was allowed to move per controlled lot. Multi-variable tuning produces an unreadable signal — you cannot know which knob helped if three changed at once. Emerald's operators were explicitly trained to resist the instinct to change several machine parameters simultaneously when a defect appeared.
3. Upstream-to-casting timing measured and enforced
The release-to-pour window — time between burnout flask exit and actual casting start — became a measured KPI rather than a soft judgment. Operators logged the actual minutes, not the scheduled minutes. Within two months, the correlation between release-to-pour drift and certain defect patterns became obvious from the logs alone. Shrinking that window tightened the cadence independent of any machine change.
4. Daily defect coding and same-shift closure
Emerald replaced the weekly defect review with a 10-minute end-of-shift review. Every defect was tagged with a standing code, close actions were committed before shift end, and the coded log became a living lookup for future shifts. This kept institutional knowledge inside the team rather than inside individual heads — the same structural change LAMAR made, arrived at independently.
5. Alignment between casting, finishing, and QA schedules
Because the original complaint was cadence unpredictability rather than defect rate, Emerald explicitly linked the casting cadence signal to finishing and QA staffing. Once the casting output stabilized, finishing could de-staff the overtime buffer that had been absorbing variance. This was the downstream dividend of the upstream discipline change.
Result pattern
The reported 50% improvement in fully automated casting efficiency covers a composite of reduced shift-to-shift variance, faster defect close time, and recovered cycle time that was previously lost to upstream slack. Emerald is careful to frame the result as a process-discipline outcome rather than a machine outcome — the CXM-C20 made the discipline enforceable by producing consistent results when the discipline was in place, but the discipline itself was the lever.
Secondary outcomes Emerald reported informally: new-operator onboarding time dropped because the training target was a documented sequence rather than shadowing a senior operator for months. Finishing overtime dropped because cadence became predictable enough to staff to plan. Export order reliability improved because weekly throughput was now a forecast rather than a hope.
Why a single-machine case matters
The LAMAR case and the Emerald case solve superficially similar problems — efficiency — but they arrive at opposite architectures: LAMAR added a second machine to eliminate regime switching, Emerald kept a single machine and tightened discipline. The lesson is not "one is better than the other." The lesson is diagnose the constraint before adding hardware. If the bottleneck is product-transition overhead, a second machine helps. If the bottleneck is upstream timing and operator discipline, a second machine just doubles the number of places where the discipline can fail.
Replication checklist — what other workshops can port
- Commit to one validated sequence per product family; forbid silent variations.
- Move only one variable per controlled lot during validation.
- Treat release-to-pour as a measured KPI, not an operator judgment.
- Run end-of-shift defect coding with same-shift close actions, not weekly reviews.
- Link casting cadence to downstream staffing so predictability creates measurable savings.
- Before adding a second machine, prove you have exhausted single-machine process headroom.
Related reading: Humid-Weekend 18K Master Troubleshooting Workflow, CXM-C20 Vacuum Pressure Vibration System, LAMAR (Dubai) CXM-C18 + CXM-C20 Case Study, Jewelry Casting FAQ.