If you've ever watched a screen deck slowly lose recovery without a clear cause, you know the frustration. Feed is the same. Tonnage is there. But the product is getting coarser, and nobody can pin down the reason.
The first time I saw it up close was in March 2020. I was doing a field audit at a sand and gravel operation in Lincoln County, Nevada—a few hours' drive from the FLSmidth Salt Lake City Utah service center. The site had been dealing with a creeping 6-7% screen efficiency loss for months. They replaced the screen media twice. No improvement.
What they were living with is called drift—and it's likely costing your operation money right now, even if you've never heard the name.
Ask ten people in mining what drift means, and most will point to an underground tunnel. A drift is a horizontal excavation that follows the ore body. That's the geological answer, and it's correct.
But on the equipment side—the side that hits your maintenance budget—drift means something different. Drift is the slow, progressive movement of a machine away from its design operating point. It's not a sudden failure. It's a deviation. And because the change happens in small steps, plants tend to normalize it.
I've reviewed equipment specifications and service agreements for FLSmidth for the past five years—roughly 200 deliverables per year, from crusher manuals to global service contract scopes. I reject about 8% of first submissions for missing tolerance limits or vague acceptance criteria. Drift is exactly the kind of problem that thrives in vagueness.
The 2020 Lincoln County audit changed how I think about all of this. Eddie, the site's maintenance lead, pointed at the screen's outlet flange and showed me the gap between the frame and the chute. "Look," he said. I want to say the gap was 15 mm, but don't quote me on the exact number—I didn't measure it with a feeler gauge that day. The point is, it was visible, and it was affecting how material left the machine.
"That's when I knew it wasn't the media," he told me. He was right. The misaligned outlet was the symptom. The real cause was the drive sheave, which had drifted along its shaft over two years of hard running. The screen was still moving material; it was just doing it less efficiently. His team had accepted the loss as normal wear. It wasn't. It was drift.
Here's the strange thing about drift: it costs more than a breakdown, but it's invisible on your dashboards.
A breakdown gets an incident report. A bearing seizes at 2 a.m., the plant loses eight hours, and everyone knows the cost because there's a line item. Drift is the opposite. There's no single failure event, so nobody assigns it a number.
In FLSmidth site audits, we've measured plants running 3-5% below design throughput from drift-related causes alone. On a 2,000 t/h circuit, that's 60-100 tons of lost production per hour. At $15 per ton of concentrate value, the math gets uncomfortable fast. I can't point you to a public study for that number—it's what our service teams have measured across sites in North America and North Africa over the past few years, and it varies with ore type and circuit design. But directionally, it's consistent: drift is taking money out of your plant every single shift.
One example from a few years back: at a copper concentrator in Arizona, we traced a 4% recovery loss to a froth camera whose mounting bracket had sagged about 30 mm from heat and vibration. The camera was still sending a picture; the control room was making reagent decisions off an image that had drifted. Nobody suspected the camera, because cameras don't wear out—but their mounts do.
What most people don't realize is that drift compounds. One millimeter of bearing misalignment accelerates bearing wear. Vibration goes up. The higher vibration loosens the next bolt over, which adds more vibration. It's a self-reinforcing loop, and the cost grows non-linearly. The vendor who says "a little deviation is fine" is technically right—for about a quarter. By the second quarter, you're paying for it in power, wear parts, and lost uptime.
Here's something vendors won't tell you: most service agreements are written around responding to failures, not detecting drift. The break-fix model starts billing when something actually breaks. So the whole maintenance ecosystem—yours and the vendor's—points at failure response instead of drift detection.
It's not malicious. It's how the industry structured itself. But it creates an absurd situation where a $2,000 laser alignment check (done while the machine is running fine) gets questioned, while a $22,000 redo caused by a motor skidding off its base gets approved without a second thought. I've seen that dynamic from the supplier side, in invoices our own customers approved.
That's why I now insist that drift diagnostics appear as an explicit line item in every service agreement I review. Not "condition monitoring" as a vague phrase—a concrete schedule of measurements, with tolerances, and a final report that lists deviations as numbers, not adjectives.
The good news is that drift is detectable and fixable, if you look for it with intention.
At the FLSmidth Salt Lake City Utah service hub, our teams run full-structure audits: laser alignment, vibration analysis per ISO 10816 severity zones, thermography, and instrument verification against reference standards. The same protocol runs through FLSmidth Maroc, our office supporting cement and mining customers across North and West Africa. In 2024, I coordinated the audit checklist for a cement client in the Atlas Mountains who'd seen support-roller drift in their kiln. The diagnostic format we use in Utah translated directly to the Morocco job. Drift doesn't respect borders.
A caveat: not every plant needs an international audit team. A regional visit from Salt Lake City or Casablanca, combined with a capable local crew, can catch 80% of drift issues. The trick is making measurement part of the routine, not part of the emergency.
And here's the part I care about most: say upfront what the diagnostic costs. When we quote an audit, I push our teams to list every measurement, every tolerance, and every price in writing. If you've ever asked for a quote and then watched the final invoice grow with "extras," you know how that kills trust. The vendor who lists everything upfront—even if the total looks higher—usually costs less in the end. I've learned to ask "what's not included" before asking "what's the price."
Catching drift early is boring work. It doesn't make a dramatic incident report. But it's the difference between a plant that runs at 97% of design and one that runs at 91%.
So next time someone asks what drift means, the geological answer is a tunnel. The practical answer is: it's the slow leak in your plant's efficiency that nobody alarms on.
Check the sheaves. Check the inlet and outlet alignments. Check your calibration records. You don't need permission to look for drift—you just need to know it's there.
Discuss This Topic
If this article connects to an active wear issue at your plant, use the inquiry form to continue the conversation with our advisory team.
Tell Us What Is Wearing Fastest
Share plant stage, maintenance window, and the result you want to improve so our team can respond with a practical next step.