What a Rejected House Cast Taught Me About Sweco Zelzate Quality

In early November, I was standing in the Sweco Zelzate service center with a printout in one hand and a caliper in the other. The service bay smelled like cutting oil and freshly painted steel. Our lead assembler looked up from a partially disassembled vibratory screen and said, "You need to see this."

I've been working in quality at Sweco for four years now. My job is to review incoming components before they go into the vibratory separators, finishing machines, and wellhead equipment we ship. It's not glamorous. I check dimensions, material certs, surface roughness, and sometimes a casting sits on my bench for an hour while I verify a single feature. I review maybe 200 part numbers a year, so I have a feel for what looks right. That morning, something didn't.

The Part That Looked Fine

The component was a motor housing for one of our screen drives. On the drawing it's called a "housing casting." On the shop floor, we call it a house cast. Not because we cast it in-house—we don't run a foundry—but because someone once typed "house cast" on an internal form and it stuck. You won't find that term on any Sweco datasheet. It's just our word.

The house cast looked fine. The flange thickness was within 0.5 mm. The outside diameter was within 0.3 mm. Surface finish was acceptable. All the dimensions on our incoming inspection form passed. But the form didn't include the internal counterbore depth. It was on the original drawing, but nobody had transferred it to the inspection checklist. That counterbore locates the vibratory motor bracket. And it was off by 2.5 mm.

To be fair, 2.5 mm doesn't sound like much. In a general sand casting, it can still be acceptable. But for this part, the drawing tolerance was 0.8 mm. More importantly, a 2.5 mm error means the motor bracket sits at a slight angle. Under continuous vibration, that angle gets worse, and the machine eventually destroys its own isolation springs. This wasn't a cosmetic problem.

The Call with Chauvin

The foundry's quality manager, Chauvin, called me after we sent the rejection report. His first words were, "It's within industry standard." I get why he said that. In a foundry context, a 2.5 mm deviation on an internal feature is often no big deal. But the standard that matters is the one on the drawing, not the general industry guideline. According to ISO 9001:2015, Clause 8.4.1, we are responsible for verifying that externally provided products fit our specified requirements. The requirement here was 35.6 mm, and the part measured 38.1 mm. That's a reject.

We rejected 38 out of 45 house casts in that batch. Chauvin's team redid them at their cost, which I appreciate. But the rework still cost us about three weeks and roughly $28,000 in schedule delays—overtime, rushed freight for replacement parts, and the usual chaos that follows a surprise rejection. This was back in November 2024, and I remember thinking that we had been lucky up to that point. We'd never had to reject an entire batch of critical castings.

What This Means for Sweco vs Eagle

That same week, a customer asked for a "Sweco vs Eagle" comparison for a new separation line. They already had Eagle screens at another plant, and they wanted to know which brand to put in the new facility. I don't think they expected my answer.

I can't give you a universal Sweco vs Eagle verdict. My sample is Sweco equipment, and not all of it. I've never taken apart an Eagle machine, so I can't tell you theirs is worse. What I can tell you is what we do to keep a defective house cast from reaching a customer. To me, that's the more useful question: how much do you trust the quality process behind the name?

If you're here looking for a Sweco aktie analys (stock analysis), I'm not your guy. The publicly traded Sweco is the engineering consultancy in Stockholm, not the company I work for. This Sweco makes vibratory screens, separators, finishing machines, and oilfield wellhead equipment. I don't follow the stock. I follow tolerance reports.

After that bad batch, we built a simple fixture with a dial indicator to check the counterbore depth in under 30 seconds. It cost about $1,800 to build. We also started doing a first-article inspection whenever the foundry changes tooling, not just once a year. In Q1 2025 (as of March, at least), our incoming rejection rate dropped to 1.5%—from the 6% we saw in 2024. That's not a perfect number, but it's a real improvement.

The customer asking about Sweco vs Eagle ended up ordering two of our machines. I don't know if they chose us because of our quality story or because the specs were better for their application. Probably both. But I do know that we had a better story to tell because we actually found the problem instead of shipping it.

The Lesson

If I have a lesson, it's this: an inspection form is only as good as the features on it. We had a process. We checked parts. But we checked the visible features and missed the one that mattered. That's a common mistake. Most buyers focus on brand reputation and price, and completely miss the small internal dimensions that decide whether a machine runs reliably. Hidden costs add up fast (like overtime, freight, and certification delays).

This approach worked for us because our suppliers are stable and our batch sizes are fairly predictable. If you're working with a new foundry or with suppliers in different time zones, you probably need even more checks than we do. Your mileage may vary. But the principle is the same: ask for the drawing tolerance, put it on the inspection form, and measure it before production starts. If a vendor tells you something is "within industry standard," ask which standard. In writing.