Views: 0 Author: Tonney Shao Publish Time: 2026-08-06 Origin: Abery Mold
In one sentence: In 2012 Schneider Electric asked us to design, build, and sample 26 injection molds — for a modular low-voltage electrical distribution product family molded in glass-fiber-reinforced polycarbonate, with wall thicknesses as thin as 1.0 mm, deep multi-face ribs, thousands of tiny EDM-machined shutoff features, and undercuts requiring CUMSA PX Xtra Sprung Cores — and to finish the whole program inside a single month; we finished on time and Schneider's engineering team, who flew to our factory to inspect the tools in person, approved the entire family after the T0 trial — before we even ran T1 — because our chief engineer had front-loaded a full Autodesk Moldflow analysis on every part before we cut a single millimeter of steel, and fourteen years later we are still running Schneider mold projects out of the same building.
Item | Detail |
Customer | Schneider Electric — global leader in energy management and industrial automation (Fortune Global 500) |
Year | 2012 |
Program scope | 26 injection molds in one program, spanning small precision inserts to large multi-cavity structural bases |
Product family | Internal structural components for a modular low-voltage electrical distribution product line (identity of specific SKU family withheld under NDA) |
Material | Polycarbonate + glass fiber (PC-GF) — flame-retardant / dielectric grade for electrical enclosure applications |
Design characteristics | Wall thicknesses 1.0–2.1 mm (thin-wall precision), thousands of small EDM-machined shutoff features, deep parallel ribs, undercuts on multiple faces |
Timeline | 30 days from PO to sample-approved tooling — approximately one-third of the normal industry lead time for a program of this scope |
Engineering approach | Full Autodesk Moldflow simulation on every part before steel-cutting: mesh thickness diagnostic, balanced fill, gate-freeze timing, deflection prediction, clamp force estimation |
Premium components used | CUMSA PX Xtra Sprung Cores for spring-loaded undercut release; industry-standard EDM electrodes for shutoff features |
Sign-off milestone | Schneider engineers approved at T0 (initial trial) — not the more common T1 or T2 gate |
Relationship since | 14 years and counting — Schneider has continued to place new mold programs with Abery every year since 2012 |
Above: three views from the Schneider Electric 26-mold program — the mold assemblies (top), the finished molded parts (middle), and a close-up of the parallel-rib architecture that made the Moldflow analysis non-negotiable (bottom).
I have already published five case studies on this site. Three describe how we behaved when a customer was under pressure or when we made a mistake — Russia, the US customer through bankruptcy, Spain KNX during Covid. Two describe how we handle engineering — the stud finder rescue project where a competitor's tool had failed, and the JCB ECU housing project where our own DFM was imperfect.
This one is different. This is the case study where nothing went wrong, and the reason nothing went wrong is worth documenting in detail because it tells you what Abery Mold actually looks like on our best days, not just on the days we had to recover.
I am naming Schneider Electric because the technical content of this case study — thin-wall PC-GF molding, Moldflow-driven DFM, T0 sign-off on a 26-mold program in 30 days — is meaningful only if a serious buyer can locate the project against the actual end application. Our specific contractual relationship with Schneider remains confidential; I am naming Schneider only as the recognizable global brand whose engineering team accepted the tools and whose products the molds have been producing for the last fourteen years.
The RFQ that arrived in our office in early 2012 was one of the most concentrated pieces of engineering work Abery Mold had ever been asked to quote. Twenty-six molds, delivered as a single program:
Some tools were compact — precision inserts with tight cosmetic and dielectric requirements
Some tools were large and complex — multi-cavity structural bases with deep parallel ribs on the interior face, mounting flanges around the perimeter, dense terminal apertures on the exterior face, and undercuts on side faces requiring slide actions
Every part was to be molded in polycarbonate with glass-fiber reinforcement — a class of material that gives you the electrical, mechanical, and thermal performance you need for a certified low-voltage distribution product, but which is genuinely difficult to mold well because of its high melt viscosity, sharp shear-heating behavior, and unforgiving gate-freeze window on thin walls
The timeline was 30 days. In our industry, a program of this scope on this material would normally run 90 days — sometimes longer. Schneider knew this. They asked us anyway because their downstream product-launch commitments to their own European distribution channels did not allow for a normal timeline. The alternative was to split the 26 molds across three or four different Chinese suppliers, accept the inconsistency, and hope for the best. They asked us to try holding the whole program in one house.
I said yes on the condition that we could front-load all of the engineering work into the first week and cut steel the moment simulation confirmed each part was mold-viable. That is the single decision this case study is really about.
PC-GF is not a forgiving material. On the specific parts Schneider was asking us to build, the risks stacked in a way that would guarantee failure without simulation:
Wall thicknesses down to 1.0 mm on some ribs — near the practical minimum for PC-GF flow, and highly sensitive to gate location and injection speed
Thousands of small shutoff features — the tiny EDM-machined details that create the openings, snap arms, and terminal apertures on a modular electrical enclosure. Each one is a potential source of flash, short-shot, or air trap if flow is not balanced through it
Deep parallel ribs on the interior face — glass-fiber orientation runs along these ribs, and mismanaged flow creates weld lines and fiber-orientation-driven warp that no amount of process tuning can rescue
Undercuts requiring slide-core release — geometric features that hard tooling alone cannot form; specialty spring-loaded slide cores were required, and their placement had to be validated against fill and packing behavior before the steel was cut around them
A 26-mold family that had to be internally consistent — inserts had to seat cleanly with mating parts across the assembly. Getting any single mold's shrinkage or warp wrong would break the assembly downstream
Every one of these risks is well-understood by any experienced polycarbonate molder. What made the Schneider program different was that we had no time to fail at T1 and correct at T2. On a normal 90-day program, you can lose a week to a bad first trial and still hit your window. On a 30-day, 26-mold program, one bad T1 propagates cost and delay across the entire family and the timeline collapses.
The only technical answer is to move the failure mode from the trial press to the simulation software — which is what we did.
For each part in the family, our engineering team ran full-suite Autodesk Moldflow analysis before signing off on the mold design for steel-cut. I am going to walk through the specific studies we ran, using screenshots from the actual project analyses that Schneider's engineers reviewed at their factory visit, because I want a buyer reading this case study to see exactly what "full Moldflow before steel" means at Abery Mold — not the marketing version.
Before you run flow analysis, you run mesh thickness diagnostic to confirm what you are actually about to simulate. This screenshot from the actual project shows measured wall thicknesses at critical locations: 1.047 mm at the thinnest features, 1.500 mm across most ribs, 2.000 mm at the base sections, up to 2.093 mm at the thickest transitions. That distribution — a 2× swing between thin and thick regions — is exactly what causes differential shrinkage and warp in PC-GF if the gate is placed wrongly. The mesh diagnostic told us where the risk regions were before we designed a single runner.
For the medium-cavity parts in the family we specified cold-sprue systems with cold runners (7×5×6 mm trapezoidal cross-section), fed by ⌀4.0 mm sprues, dropping to ⌀1.6 mm pin gates on the parts. For selected parts we used cashew-gate configurations (1×7 layout) and valve-gate control to sequence the fill front. This is not the standard "one hot drop, hope for the best" approach — it is a gate strategy tuned to the flow-length-to-thickness ratio of each specific part. Schneider's engineering team specified some of the gate dimensions themselves; where they did, we implemented to spec and simulated the result.
The most important single Moldflow result on a program like this is the fill-time contour. Evenly-spaced isochrones (the coloured contour bands) mean the flow front is moving at a uniform rate — the plastic is not stalling in narrow sections or racing through wide ones. This screenshot shows one of the larger parts filling in 2.091 seconds with clean, evenly-spaced contours from the six-point runner and no short shot anywhere on the part. That is what a mold-viable PC-GF design looks like. If those contours had been visibly bunched or fragmented, we would have moved gates and re-simulated before quoting the steel.
PC-GF melt temperature has to sit inside a fairly narrow window — typically 280–320°C nominal, but with strong sensitivity to shear heating from injection speed. The screenshot shows measured melt temperatures at multiple locations across one of the parts (267–286°C in the flow front, 210–218°C at the leading edge where fresh cold-material contact drops the local temperature). The distribution stays inside the PC-GF process window; the injection speed we specified is not over-shearing the polymer. On a thin-wall PC-GF part this check is critical — over-shear degrades the material and produces brittle parts at the customer's factory, which no amount of geometry cleverness can fix.
The frozen-layer study shows how solidification progresses through the part over time. On this specific tool, the gate freezes at approximately 12 seconds and the full part solidifies at approximately 13 seconds — a 12-second packing window during which hold pressure can still push material into the shrinking cavity to compensate for volumetric shrinkage. That number is the difference between a part that comes out to spec and a part that comes out sunk. It also drives cooling-channel design: we had to make sure the mold's cooling circuits kept the cavity surface below the freeze threshold at the required cycle time, and the frozen-layer simulation gave us that number directly.
Every finished Moldflow report at Abery ends with a clamp-force curve — because a tool that cannot be closed on the customer's target press is worse than useless. On this specific part the peak clamping force reached 262 tons at approximately 2.1 seconds into the injection stroke. The analysis note flags: "A 300-ton machine cannot satisfy the requirements." That directly drove the production-press specification for the customer's downstream molder. If we had not simulated clamp force, the tool might have been declared "finished" only for Schneider's contract molder to find they could not run it on the presses they had scheduled.
Finally — and closing the loop back to the mesh thickness map we started from — we predicted total part deflection after cooling and demolding. On this part the Z-component deflection was 0.46 mm at maximum and -0.18 mm at minimum, magnified 3× in the display for visibility. Those numbers were fed back to the Schneider assembly team so they could confirm the modeled deflection was compatible with the mating-part tolerances on the finished distribution product. It was. The tool went to steel-cut the same week.
A Moldflow-perfect design still fails at the trial press if the physical mold components are the wrong ones. For the Schneider program, several parts had geometric undercuts that hard tooling alone could not form — the finished parts needed features that "hooked" underneath the mold's main draw direction, which meant the mold had to release those features during opening, before the ejector stroke.
For that class of undercut, our tooling team specified CUMSA PX Xtra Sprung Cores — spring-loaded slide cores from a specialist Spanish mold-component manufacturer.
CUMSA sprung cores are not the cheapest slide component you can buy. They are the reliable ones. The PX line is patented, uses hardened alloy steel (56±2 HRC on the core body, 45±2 HRC on the bushing), and is rated for continuous use up to 150°C mold temperature — which matters on PC-GF where the mold has to run at 90–110°C to control fibre orientation. On a 30-day program with no time to correct a slide failure at T1, choosing the premium component was the correct engineering economy. It cost more up front. It saved the timeline.
There is a version of this case study that I could write where the entire story is engineering — Moldflow this, CUMSA that, clamp force the other. That version would be honest and it would sell the technical work. It would also leave out the part of the story that I think matters most.
Twenty-six molds in thirty days did not happen because our software was good. It happened because our tooling team decided, on the second week of the program, that the timeline was going to be met and that they were going to sleep in the factory to meet it if that was what it took. Some of them did. The EDM operators worked in staggered shifts to keep the electrode machining running through the nights. The chief engineer reviewed DFM files at his desk with a bowl of instant noodles for the better part of a fortnight. The tooling manager coordinated steel supply across three different vendor accounts to keep every mold's blanks arriving exactly when its slot on the machining floor opened up.
I am not sentimental about this. I am not going to claim the team's overtime was inspired by a shared vision of European electrical distribution. What they were inspired by was that Schneider Electric had said yes to a Chinese mold maker for a program that most European engineering teams would not have trusted a Chinese mold maker with in 2012, and everyone in our shop understood that the outcome of those thirty days would tell Schneider — and, indirectly, tell the whole European buying market — what a Chinese tooling operation was actually capable of when it decided to bring its best work.
They brought it. That is what happened.
Toward the end of the thirty days, Schneider engineers flew to Shenzhen to inspect the completed tools in person. This was not a formality — they had a list of specific dimensional, cosmetic, and process-parameter checks for each of the 26 molds, they ran through them systematically on our sample floor, and they stayed a week.
At the end of that week they signed off the entire family after T0 — the very first trial run of each tool.
For readers who do not spend their days on injection-mold approval gates: T0 sign-off is unusual. The normal industry gate is T1 or, more commonly on programs of this complexity, T2 — with dimensional inspection, fitment checks, and typically one or two minor tool corrections between T0 and T2. Approving after T0 means the customer's engineering team looked at the first sample off the trial press and concluded there was nothing they needed to see us adjust. On a 26-mold family in PC-GF with thin walls and slide-actioned undercuts, that outcome is what Moldflow-driven front-loaded engineering is supposed to produce. This time it did.
I want to add one small human detail, because it is my favorite part of the story and I do not think it makes anyone look bad to tell it. At the end of the sign-off week, the visiting Schneider engineers — this was their own personal gesture, not a company transaction — handed our project team a small red envelope containing RMB 1,000 as a symbolic thank-you for the overtime the team had worked. In Chinese business culture, a red envelope of that size from a visiting client team to a supplier's engineers is an unmistakable signal of respect between colleagues. My tooling team framed it and hung it on the office wall. As far as I know it is still there.
Since 2012 Schneider Electric has continued to place mold and production work with Abery Mold. I am not going to publish specific SKU counts or program revenues under our confidentiality agreement, but I can say the following, factually:
Every year since 2012 has included at least one new mold program from the Schneider relationship
The technical scope has broadened — from the original PC-GF distribution component family into other material classes and other product categories inside Schneider's LV product portfolio
The commercial terms have never had to be renegotiated as a rescue exercise. The relationship has been stable, transparent, and adult on both sides
The 30-day precedent set in 2012 has never been repeated, because it did not need to be — the point of that program was to prove capability, not to establish it as a normal delivery rhythm
I want to close by being specific about what a serious buyer should conclude from reading this article. Three things.
One — front-loaded Moldflow is not a marketing checkbox. On a program like this one, Moldflow is the only tool that lets you compress a normal 90-day tooling timeline into 30 days without moving the failure risk from steel-cut to the trial press. If a Chinese mold supplier says they "run Moldflow when the customer requests it," that supplier is treating simulation as a sales feature. If they say they run mesh-thickness / gate-design / balanced-fill / melt-temperature / frozen-layer / clamp-force / deflection studies on every part before steel-cut, they are treating simulation as the actual engineering method. The difference shows up on your first trial.
Two — premium mold components are cheap on a compressed timeline. CUMSA sprung cores cost more than the generic slide mechanisms you can buy from any Chinese tooling catalog. On a 90-day program, you can afford to argue about the difference. On a 30-day program, you cannot afford a slide failure at T1. This is a purchasing decision as much as an engineering one — and it is the kind of decision a supplier who has actually done this work will make automatically, without waiting for you to ask.
Three — a T0 sign-off on a 26-mold program is not a talent story. It is a process story. The reason Schneider approved after T0 was not that our chief engineer was uniquely gifted. It was that we had built a design-review process where nothing went to steel until simulation confirmed it was mold-viable, and the front-loaded engineering absorbed the failure modes that would otherwise have shown up at T1. That process is now the standard way we run any tight-timeline program, and it is documented internally as our default for anything on glass-filled polycarbonate, glass-filled nylon, or PBT-GF with thin walls under 1.5 mm.
If you are running a program on any of those materials and any of those constraints and you want to have a conversation about whether Abery Mold is the right tooling partner for it, my email is at the bottom of this page. I answer serious European and North American RFQs personally.
Our specific contractual relationship with Schneider Electric is confidential. Fourteen years of continuous cooperation is a matter we describe factually because it is true and Schneider's engineering team is aware we publish it, but we do not represent ourselves as a Schneider-approved-supplier in any regulated commercial sense on this website. Buyers evaluating us for LV distribution or modular switchgear tooling should assess our capabilities on their own merit and reference this case study for technical context only.
No, and it would be dishonest to claim otherwise. We run full Moldflow analysis on every project that involves thin-wall glass-filled thermoplastics, tight timelines, complex flow paths, or geometries where our engineering team's structured checklist identifies a warpage or fill risk. On simpler single-cavity molds in stable materials with standard geometry, structured DFM review without simulation is often sufficient, and we do not add cost the project does not need. The Schneider program was in the first category; almost every serious European industrial-electronics program is.
Mesh-thickness diagnostic, gate and runner design overlay, balanced-fill isochrones, melt-temperature and shear-rate distribution, frozen-layer and gate-freeze timing, clamp-force curve, deflection prediction, and (for multi-cavity families) fill-balance across cavities. The full set is illustrated with real project screenshots in Section 4 above.
T0 is the very first sample off the newly-completed tool — parts as they come out, before any mold-side corrections. T1 is typically the first inspection-gate trial, often after minor initial cleanup of the tool. T2 is the second engineering gate, usually after one round of correction. On complex programs, some customers require T3 or higher. Approving after T0 means the engineering team saw no correction they needed to see us make. It is uncommon on a 26-mold program.
The LV electrical distribution and modular switchgear category — techno-plastic enclosures, DIN-rail modular components, distribution board internals, MCB enclosures — has been one of our continuous mold-building categories since 2012. See also our Spain KNX case study for a European KNX-family perspective on a similar product class.
Tonney Shao is the founder and CEO of Shenzhen Abery Mold & Plastics Co., Ltd., an ISO 9001-certified injection mold maker and plastic parts manufacturer serving automotive, consumer electronics, medical device, industrial equipment, and consumer goods customers worldwide. Abery has built more than 3,000 molds since 2009 and operates a tooling and molding facility in Shenzhen, China. Tonney personally reviews every case study published on this site.
Contact: tonney@a-mold.com · +86-13925214356
Case study published 7 August 2026. The specific product SKU family this program produced, and the specific Schneider Electric business unit that engaged the program, are withheld under a mutual non-disclosure agreement. Schneider Electric is named in this case study as the recognizable global end-application manufacturer whose engineering team engaged our tooling work in 2012; Abery Mold does not claim regulated-approved-supplier status with Schneider Electric on this website and buyers should not infer any such status from this article. Autodesk Moldflow is a registered trademark of Autodesk, Inc. CUMSA and the PX Xtra Sprung Core designation are trademarks of Comercial de Utiles y Moldes, S.A. (CUMSA). All Moldflow screenshots on this page are from the actual 2012 Schneider project analyses; they are published with the specific dimensional annotations and material identifiers redacted where confidentiality applies. Molded-part photographs on this page are of components from the ongoing Schneider mold-project relationship and are used with the customer's awareness.
