Vibration analysis: when it's worth it

The vibration survey report arrives as a PDF. Forty machines, colour coded. Thirty-two green, six amber, two red. One of the reds is the main extract fan. The note says "bearing defect frequencies present, plan replacement".

It goes in the shared drive. The fan runs for another seven weeks. Then it goes on a Saturday night, and takes the shaft with it.

You paid for the warning. You just didn't use it.

That's the real question with vibration analysis. It works. Whether it pays depends on which machines you put on it, and what you do when it tells you something.

What it actually finds

Every rotating machine vibrates. A healthy one has a steady pattern. When something starts to go wrong, the pattern changes, often long before you'd hear or feel it.

A trained analyst can read that pattern and tell you a lot:

• imbalance in a fan or a rotor
• misalignment between a motor and a pump or gearbox
• looseness in a base, a bearing housing or a foot
• early bearing damage, sometimes months before it fails
• gear tooth wear or damage
• belt problems and soft feet

The early bearing warning is where the money usually is. A bearing change you plan for a Tuesday morning is a few hours. The same bearing at 3am, with a wrecked shaft and no spare, can be a day or more of lost output.

When it's worth paying for

We'd put a machine on vibration monitoring when most of these are true:

1. It's critical. If it stops, the line or the site stops with it.
2. There's no standby. No duty and assist pump, no spare fan to switch over to.
3. The parts have a long lead time, or the repair needs a specialist.
4. It runs most of the time, at a steady speed, so the readings compare like with like.
5. It's big or fast enough that a failure does real damage. Main fans, big pumps, compressors, gearboxes on critical drives.

HSE's code of practice for PUWER points the same way. Maintenance "should be targeted at the parts of work equipment where failure or deterioration could lead to health and safety risks" (L22, paragraph 73). Where safety-critical parts could fail, it calls for "a formal system of planned preventative or condition-based maintenance" (L22, paragraph 75). Vibration monitoring is one of the main ways sites do the condition-based part.

If you haven't ranked your machines, criticality ranking is where to start. Your top ten is usually your vibration list.

When it isn't

Plenty of kit doesn't need it.

Small motors you hold spares for are one. If a 1.5kW conveyor motor takes your fitter 40 minutes to swap and there's one on the shelf, a monthly survey costs more than the failure. Run it to failure and keep the spare. Our post on critical spares covers that choice.

Slow-turning kit can be hard to read with a standard survey. Machines that start and stop all day, or change speed every few seconds, give readings that are hard to compare. Talk to the analyst before you add them.

And if nobody will act on the report, don't buy it. A survey with no follow-up is just an expensive record of what's about to break.

Three ways to do it

In-house, with a handheld meter. A basic overall vibration meter isn't expensive. Take readings on the same marked points each week and trend them. It won't tell you what's wrong, but it will tell you something has changed. That's often enough to call in someone who can find out. Our guide to condition monitoring on a small budget covers how to set up the round.

A route survey from a specialist. An analyst visits monthly or quarterly with a data collector, reads every point on your list and sends a report. You get the diagnosis without training your own team. The gap between visits is the weakness. A fast-failing bearing can go from fine to failed between rounds.

Fixed sensors. Wireless or wired sensors on your most critical machines, sending data all the time. Good for kit that's hard to reach or too important to miss. Costs more up front, and someone still has to look at the alerts.

Most sites end up with a mix. Fixed sensors on the two or three machines that matter most. A route survey for the next tier. A handheld round for the rest.

Before you sign a contract

Ask the provider:

• Who reads the data, and what training do they have?
• What does the report tell me to do, and by when?
• Will you come back and check after we've done the repair?
• Can I see the trend for each machine as well as this month's colour?
• How quickly will you tell me about a red?

The last one matters. A red sat in an inbox for a week is wasted.

The report is the easy part

Here's our view. Reading the data is the easy bit. The hard bit is having someone free to do the work when the report lands.

An amber on a gearbox means plan a job. That needs a fitter, a stop, a part, and often an alignment check after. If your team is already flat out on breakdowns, the planned job gets pushed back a week, then another. Meanwhile the amber turns red. Then it turns into the breakdown you paid to avoid.

So before you buy a survey, decide who'll do the follow-up work. If it's your own team, protect their time for it. If they can't, line up outside help for the repairs, so the warning turns into a planned job. After any repair, check the fix. A new bearing on a misaligned shaft fails again. Our guide to shaft alignment covers the basics, and the PUWER code of practice is a reminder of what's at stake. It notes that "failure to lubricate bearings or replace clogged filters might lead to danger because of seized parts or overheating" (L22, paragraph 70).

What to do on Monday

1. List your ten most critical rotating machines and mark which have no standby and long-lead spares.
2. Pull your last vibration report, if you have one, and check every amber and red has a job and a date.
3. Decide who will do the repair work when the next red comes in.
4. Take a first set of handheld readings on your top three machines, so you have a baseline.

* Paragraph numbers refer to L22, fourth edition (2014, amended 2018), correct as of 2 October 2026. Check HSE's L22 page for the latest.