Latest / The Operator's Manual with Fexingo: Daily Decisions, Tactical Wins, and Business Execution / How One Factory Cut Scrap 55 Percent With One Die Casting Change
Transcript
- Lucas: So there's a die-casting shop outside Cleveland, family-owned, been around since the 1970s. They make aluminum housings for automotive sensors — nothing glamorous, but they're good at it. Except one part, a particular housing, was running an 18 percent scrap rate. Almost one in five castings ended up in the scrap bin. Luna: Eighteen percent — that's brutal. And I'm guessing they'd been living with it for a while? Lucas: Years. They'd tried adjusting temperature, injection speed, die lube — the usual knobs. Nothing got it below 18. Then a new process engineer showed up, fresh out of school, and looked at the die itself — specifically the gating system, the channels that feed molten aluminum into the cavity. She noticed the gate was too narrow and placed at an angle that caused the metal to jet in, creating turbulence. That turbulence trapped air, which caused porosity — weak spots that showed up in x-ray inspection. Luna: So the fix was changing the geometry of the gate? Lucas: Exactly. They widened the gate by about three millimeters and repositioned it so the metal entered the cavity in a laminar flow — smooth, no jetting. The mold modification cost them roughly $2,500. In the first month after the change, scrap dropped to 8 percent. After a few more tweaks, it settled at 7.5. The plant was producing 120,000 of these housings a year at roughly $18 per part. The scrap reduction saved them about $340,000 annually. Luna: That's a 130-plus times return on a $2,500 investment. And it wasn't new machinery or software — it was just understanding the physics of metal flow. Lucas: Right. And this is a pattern we see across manufacturing: the most impactful changes are often at the design stage — what you could call 'shift-left' quality. The earlier you address a defect, the cheaper and easier it is to fix. By the time you're inspecting finished castings, you've already baked in the cost of scrap material, energy, machine time, and labor. Luna: I've heard shift-left used a lot in software — catching bugs before they reach production. But in manufacturing, it's the same concept: a change in the mold design costs hundreds, a change after casting costs thousands in rework or scrap. Lucas: Exactly. And what made this case stick with me is that the engineer didn't do anything exotic. She didn't use simulation software — she looked at the existing die, sketched a wider gate on a napkin, and convinced the tool room to try it. The toolroom foreman had been there 25 years and said, 'We've always done it this way.' The engineer had to earn his trust by running a small production trial. They made a hundred parts with the modified die and x-rayed every one. Zero porosity. Luna: That's the human side that often gets ignored. It's not just the technical fix — it's getting the team to try something different. Lucas: Precisely. And that's a big part of what this show is about — the real-world execution challenges. Speaking of which, if you find these stories useful for your own work, we keep this show ad-free thanks to listener support. You can help at buymeacoffee.com/fexingo. It's a simple way to say these conversations move your work forward. We appreciate every bit of it. Luna: Yeah, it really does help keep the lights on. And it lets us dig into stories like this one — the kind that don't make headlines but change how a shop runs. Lucas: Okay, back to the Cleveland shop. After the gate fix, they had another problem: the new gate left a slightly bigger gate vestige — the nub where the runner attaches to the part. The customer hadn't specified a tolerance on that vestige, but the customer's assembly robot was occasionally misreading the part in a vision system. So the shop had to hand-trim the vestige on a few parts, which added labor cost. Luna: So they traded scrap for a secondary operation. That sounds like a classic unintended consequence. Lucas: Right. But here's what happened next: the engineer looked at the gate again and realized she could thin the gate slightly — just half a millimeter — without reintroducing turbulence. That reduced the vestige to nearly flush, and the robot stopped rejecting parts. Total additional cost: zero. The gate was already being modified; they just changed the depth. That brought scrap down another half a percent and eliminated the hand-trimming. Luna: It's a good reminder that 'done' is rarely done. You have to watch for second-order effects. Lucas: Exactly. And this ties into a broader principle called 'process capability.' The shop didn't just fix a defect — they made the process more robust. Before the change, the process was running at a Cpk — that's a capability index — of about 0.8. After both gate changes, it hit 1.5. In practical terms, that means the process was producing so few defects that it could run with less inspection. They eventually moved from 100 percent x-ray inspection to a sampling plan, which saved another $40,000 a year in labor and machine time. Luna: So the total savings from that $2,500 investment was closer to $380,000 a year when you factor in reduced inspection. Lucas: Correct. And here's what I find compelling: this shop had been running that die for four years, accepting 18 percent scrap as normal. The engineer's fix was not a breakthrough in materials science. It was a basic application of fluid dynamics that any first-year engineering student could explain. The barrier was organizational — nobody had questioned the gate design because the process had been 'good enough.' Luna: And that's where the real lesson is, isn't it? How many processes in our own work have become 'good enough' — and we've stopped asking if there's a better way? Lucas: Exactly. The engineer later told me that her supervisor initially pushed back because modifying the die would take the press down for two shifts. But the cost of two shifts of downtime — about $6,000 — was dwarfed by the monthly scrap savings. The supervisor was thinking about short-term utilization. The engineer was thinking about total cost. And that's the operational mindset that this show is about. Luna: I want to zoom out a little. Die casting is a specific process, but the pattern — design change, small investment, large savings — shows up everywhere. Are there other examples you've come across? Lucas: Sure. There's a well-known story from a Toyota supplier that makes brake calipers. They had a defect where machining marks on a mating surface caused a leak. The traditional fix would be better machining or tighter tolerances — both expensive. Instead, an operator suggested adding a thin gasket — a $0.02 part. That eliminated the defect. The savings were in the millions because they avoided a recall for the entire vehicle line. Luna: Two cents. And that came from an operator on the floor, not an engineer. Lucas: Right. Which is another crucial point: the people closest to the work often see the simplest fix. But they need an environment where they can speak up. In the Cleveland shop, the engineer was new and not yet embedded in the culture. She had to fight to be heard. The challenge for leaders is to create a culture where the default answer to 'why don't we try something different?' is not 'that's how we've always done it.' Luna: So what happened after the gate change? Did the shop adopt a more systematic approach? Lucas: They did. The engineer was promoted to process improvement lead, and they started a monthly 'scrap review' where operators, engineers, and management review the top three defect categories and brainstorm changes. In the first year, they cut overall plant scrap from about 12 percent to 6.5 percent. The gate change was the catalyst. After that, people believed that small changes could have big effects. Luna: That's a great outcome. I wonder, though — for a listener who doesn't work in manufacturing, what's the takeaway? The specific fix is about metal flow, but the principle applies anywhere. Lucas: I think the principle is this: when you see a recurring problem — a defect, a delay, a customer complaint — don't just treat the symptoms. Look at the fundamental design of the process. Ask: is there a constraint, a step, a tool that could be modified? And don't assume that the current way is the only way. The Cleveland shop assumed that 18 percent scrap was inherent to that part. It wasn't. The geometry of the gate was arbitrary — it had been designed that way because that's how the toolmaker had always done it. Luna: So question your constraints. Which sounds simple but is incredibly hard when you're inside the process. Lucas: It is. And that's why we tell these stories — to give you a concrete example you can point to when you're trying to make a case for a small experiment. The next time someone says 'we can't change that,' you can say 'there's a die-cast shop in Ohio that spent $2,500 and saved $380,000 a year by widening a gate by three millimeters.' Luna: I'm going to use that line. Alright, but I have to ask: is there a downside? Are there cases where this kind of tweak backfires? Lucas: Absolutely. The risk is that you optimize one part of the system at the expense of another. The gate change initially created the vestige problem. If the engineer hadn't caught that, they might have saved scrap but created a labor cost that ate up the savings. The key is to measure the whole system — not just the scrap rate, but downstream effects. That's why the best process improvements are iterative: change, measure, adjust, measure again. Luna: So it's not a 'set it and forget it' approach. It's continuous. Lucas: Exactly. And that's the core of operational excellence. It's not about one big improvement — it's about the discipline to keep looking for the next small one. The Cleveland shop is now looking at their cooling time cycle. They think they can shave 4 seconds off the cycle per part. That would give them about 2 percent more capacity without any capital investment. Luna: Four seconds. That's the size of these improvements — three millimeters, four seconds, two cents. It's almost comically small, but the compounding effect is enormous. Lucas: It is. And I think that's a hopeful message for anyone running an operation. You don't need a massive transformation. You just need to find one gate, one gasket, one tweak, and see where it takes you.