Application note
I've Documented $11,800 in Coriolis Flow Meter Reinstallation Costs. Here's the 20-Minute Check That Would Have Prevented Every One.
-
The Mistake That Took Four Months to Surface
-
Why Coriolis Meters Are Less Forgiving Than You Think
-
The Check I Now Refuse to Skip (And It's Not What You'd Expect)
-
"But We Don't Have Time for This" — I Get It. I Used to Say the Same Thing.
-
The 20-Minute Pre-Check That Actually Works
-
"You Can't Check Everything" — You're Right. But You Can Check This.
If you're commissioning a Coriolis flow meter and you haven't blocked out at least 20 minutes for a pre-startup verification, you haven't actually saved time — you've just scheduled a future disaster.
I'm a process instrumentation engineer. I've been specifying and commissioning flow meters — mostly Endress+Hauser, but plenty of others — in chemical and food processing plants for about 8 years. In that time, I've personally documented $11,800 in wasted budget across maybe 14 installation-related incidents. Every single one was preventable. And every single one started with someone (usually me) saying "it looks fine."
The Mistake That Took Four Months to Surface
Our worst one was a Promass F installation back in September 2021 — actually, late September, I remember because it was right after the shutdown. We were adding the meter to a glycerin line. The piping was tight, we were under a deadline, and I approved the install after a visual check. Looked good. Flanges aligned, bolts torqued, cable runs clean.
Four months later, our production data started drifting. QA flagged a density discrepancy in the finished product. Turns out the meter had been reading 0.4% low since day one — enough to throw off our mass balance calculations without triggering any alarm.
The cause: pipe stress. The upstream flange was slightly misaligned. Not enough to leak. Not enough to see. But enough to pre-load the vibrating tube and shift the zero point.
Cost of the fix: $3,800 in re-verification, plus two days of production hold while we re-ran QA on four months of batch records. Cost of the check I skipped: about 20 minutes with a straightedge and a torque wrench.
Why Coriolis Meters Are Less Forgiving Than You Think
Here's what I didn't understand early on: Coriolis meters don't fail loudly. They don't throw an error code when the mounting is off. They just drift. Quietly.
The working principle is actually straightforward — the meter vibrates a tube and measures the phase shift as fluid flows through it. Mass flow rate is proportional to that phase shift. But if the tube is mechanically constrained — by pipe stress, by vibration from a nearby pump, by a poorly supported manifold — the phase measurement gets contaminated. And the meter has no way of knowing.
ISO 10790:2015 — the standard for Coriolis flow measurement — Section 6.3 basically says "isolate the meter from external stress." That's the polite way of saying "if you don't support the piping properly, your precision instrument becomes a very expensive guess."
Endress+Hauser's own installation documentation is clear about this. Their Promass manual has a whole section on support and alignment. But here's the thing — when you're in the middle of a shutdown and the schedule is already two days behind, "recommended" starts sounding optional.
The Check I Now Refuse to Skip (And It's Not What You'd Expect)
I used to think the critical check was the pipe alignment. It's not. Pipe alignment you can see. The thing that got me twice is electrical.
Specifically: shield grounding.
On my third Coriolis installation — a smaller Promass Q on a food-grade line — I terminated the cable shield at both ends. Felt thorough. Felt professional. What I actually did was create a ground loop. The meter picked up a 2-mV noise signal from a nearby VFD. Nothing dramatic. Just a 0.2% error that we didn't catch for six weeks.
Now I do a continuity check between the shield and the local ground before I even power up. It takes two minutes. It's boring. It has saved me from three potential re-installations — and honestly, I consider that the best ROI of anything I do on a commissioning.
"But We Don't Have Time for This" — I Get It. I Used to Say the Same Thing.
To be fair, I understand the pressure. Shutdown windows are brutal. Procurement is riding you. The contractor wants to go home. And here I am saying "stop — let's spend 20 minutes checking what already looks fine."
I get it because I was that guy. In my first two years, I skipped the pre-check maybe half the time. Not because I was lazy — because I genuinely believed the meter would tell me if something was wrong. That's the core misjudgment. Coriolis meters are precise, but they're not diagnostic. They measure the result. They don't tell you whether the measurement is trustworthy.
The pattern isn't unique to flow meters. I've seen the same thing with a confocal microscope that looked perfectly aligned until someone checked the calibration slide. Same with a T865 thermal camera — the emissivity setting was off by 0.05, and nobody caught it for weeks because the images looked right. Even with an Eppendorf pipette — which seems like it should be straightforward to use — there's a calibration trim adjustment that almost nobody checks. That's how you end up with 0.1 µL errors across an entire plate.
Whether it's a Coriolis meter, a microscope, or a pipette — the principle is identical. Verification before operation. Not because you expect failure. Because the cost of checking is always lower than the cost of discovering.
The 20-Minute Pre-Check That Actually Works
Here's what I do now. Most of it traces back to E+H's official recommendations, but the order and specifics come from my own incident log:
Pipe alignment by measurement, not eyeball. Straightedge and feeler gauge. If there's visible daylight under the straightedge, stop.
Support brackets within 12 inches of the flange on both sides. Not optional. Vibration from unsupported piping is the most common cause of zero drift I've seen.
Continuity check between cable shield and ground. Both ends should read open. If either reads closed, you've got a ground loop.
Static zero check before flowing fluid. Power the meter with no flow and log the zero reading. Compare to factory calibration. If deviation is more than ±0.1%, something mechanical is wrong. I've caught three stressed installations this way.
That's it. Twenty minutes, maybe. I've been tempted to skip it. Especially when the contractor is standing there with his clipboard, waiting.
"You Can't Check Everything" — You're Right. But You Can Check This.
I can't tell you that this 20-minute check catches every possible installation error. It won't catch a miswired output signal or a software configuration mismatch. Those are different problems with different checks.
Also — my experience is based on roughly 60 installations, mostly in mid-size plants with relatively stable ambient conditions. If you're installing on an FPSO or in a facility with extreme vibration, the calculus might be different. You'd probably need more support, not less.
But here's what I can tell you: $11,800 in rework costs across roughly 14 incidents — all of them traceable to something that a 20-minute pre-check would have caught. Every single one.
So no, I'm not going to tell you that Coriolis meters are fragile or that Endress+Hauser's documentation is incomplete. That's not the point. The point is that precision instruments only deliver precision when they're installed precisely. And "precisely" doesn't happen by accident. It happens when someone slows down and checks.
That someone used to not be me. Now it is.