If someone Googles “hosokawa,” they might find a tennis player named Taishi Hosokawa, or wonder what the difference is between skiing and downhill skiing. (Turns out, downhill is a subset of skiing—ask me how I know.) But for me, Hosokawa means something else entirely: the Alpine series air classifiers and ACM mills that I’ve worked with for the past seven years.

My name is Eddie, and I’m the guy who documents his own mistakes so others don’t have to make them. I’ve personally cost my company about $18,000 in wasted production over the years. The worst single incident happened in September 2022, and it started with a drift I thought was harmless.

The Surface Problem: Inconsistent Fines

We were processing calcium carbonate for a client that required a D90 of 10 microns. Our Hosokawa Alpine 100 AFG fluidised bed opposed jet mill with an integrated classifier had been running that spec for months without issues. Then the operator reported that the product looked coarser on the particle analyser. Not dramatically—maybe D90 creeping to 12 microns. But enough to fail the client’s spec.

My first instinct was to tweak the mill’s grinding pressure and classifier speed. I spent half a day adjusting parameters, running test samples, scratching my head. Nothing brought it back consistently. Classic symptom of a surface-level fix.

The Deeper Reason: I Knew Better, But Skipped It

Here’s the thing: our pre-run checklist has a step labelled “Classifier wheel gap check.” It takes about five minutes. You remove a cover, measure the gap between the rotating wheel and the stationary housing, and verify it’s within 0.3–0.5 mm. If the gap drifts wider, coarser particles slip through.

I knew I should do it. But the operator was rushing to get the order out, and we’d been running the same product for weeks. “What are the odds?” I thought. Well, the odds caught up with me.

So glad I paid for rush delivery. Almost went standard to save $50, which would have meant missing the conference entirely.

The Price Tag: Not Just Money

I submitted a 120-kg batch for QC. The report came back: D90 of 14.2 microns. Not even close. That batch went straight to re-milling—$1,600 in energy, labour, and wear. Plus a one-week delay for the client, and a 3-day production backlog for the next order. Total tangible loss: roughly $3,200.

But the intangible cost was worse. The client’s purchasing manager lost confidence. We had to provide three certified batches at no extra charge to win back trust. Lesson learned the hard way.

The Solution: Back to Basics

After that disaster, I created a 12-point pre-run checklist that includes:

• Classifier gap measurement (even if you just did it last week)
• Grinding nozzle pressure verification
• Filter bag condition check
• Pneumatic sealing system test

We’ve now caught 23 potential issues using that checklist in the past 18 months. A five-minute check that would have prevented the $3,200 mess. The conventional wisdom is to optimise parameters first. My experience with over 200 mill orders suggests that prevention beats cure every time.

Bottom line: don’t be Eddie. Take the five minutes. Your Hosokawa mill will thank you—and so will your P&L.