Vibration Acceptance Testing: What Zone A and Newly Commissioned Actually Require

There is a phrase buried in the vibration standards that ends up in a surprising number of contracts: newly commissioned. It appears in the description of evaluation zone A, and it is the reason a machine that would be perfectly acceptable after five years of service can be rejected on the day it is handed over.

People searching for what zone A means are usually not doing academic reading. They are either about to accept a machine, or arguing about one that has already been installed and is reading higher than somebody expected. This article is written for both situations: what the zone actually says, how to set acceptance criteria before the machine arrives so the argument does not happen, and what to do when a new machine lands outside zone A.

What the Zones Actually Say

The vibration standards for evaluating machines in situ divide measured vibration into four evaluation zones, and the wording of each one matters more than the numbers attached to it.

Zone A is described as the range into which the vibration of newly commissioned machines normally falls. Note what that is and is not. It is a statement about where new machines typically sit, not a statement that anything above it is faulty.

Zone B is the range considered acceptable for unrestricted long-term operation. A machine in zone B can run indefinitely. It is not damaged, it is not in distress, and nothing needs to be done about it.

Zone C is judged unsatisfactory for long-term continuous operation. The machine can generally be run for a limited period while remedial action is arranged, but it is on notice.

Zone D is severe enough to be capable of causing damage.

The numeric boundaries between these zones are not universal. They vary by machine group, essentially by size and power, and by whether the machine sits on a support that is rigid or flexible relative to its running speed. A large machine on a flexible foundation is allowed considerably more vibration than a medium-sized machine on a rigid one, and applying the wrong group or the wrong support classification is the most common way acceptance arguments start from a false premise. The tables and their conditions live in the standard itself, which is available through the ISO catalogue, and our guide to the vibration limits covers how the groups are assigned.

One further trap: many people are working from ISO 10816 while the current numbering is ISO 20816. The zone structure carried across, but the part numbering and some of the detail did not, and quoting the superseded part in a contract creates ambiguity nobody needs. We covered that transition in ISO 10816 vs ISO 20816.

The four evaluation zones and what each one actually says. The wording matters more than the numbers attached to it. Source: MIMAH engineering practice, after the ISO in-situ evaluation framework.

Why New Machines Are Held to a Tighter Limit

The logic behind zone A is not that new machines are fragile. It is that a new machine is the only time you will ever see the equipment free of accumulated defects, and that makes its vibration reading the most informative measurement in its life.

A newly installed machine has fresh bearings, a balanced rotor, a correct alignment, clean lubricant and an uncontaminated flow path. If it is not in zone A, one of those things is already wrong, and the cheapest moment to fix any of them is before the machine has been accepted and the contractor has left site. A reading in zone B on day one is not dangerous, but it is unexplained, and unexplained vibration on a new machine tends to be an installation defect that will get worse.

That is the practical case for holding acceptance to zone A: not because zone B would damage the machine, but because accepting zone B means accepting an unidentified defect and paying to correct it later.

Setting Acceptance Criteria Before the Machine Arrives

Most acceptance disputes are contract failures rather than engineering failures. They happen because the criteria were never agreed in enough detail, and both parties later read the same standard differently. A specification that prevents this names six things.

The standard and edition. Not "ISO vibration limits" but the specific standard, part and edition. Parts differ by machine type, and editions change.

The machine group and support classification, agreed in advance with the reasoning recorded. This single line resolves most later disputes, because it fixes which column of the table applies.

The measurement quantity and units. Velocity in millimetres per second RMS is the usual basis for these zones, but displacement and acceleration are used in other contexts and a specification that just says vibration invites confusion. Our guide on reading vibration readings covers the conversions and where they go wrong.

Measurement locations and directions. Bearing housings, which bearings, horizontal, vertical and axial, and how the points are marked so that the acceptance measurement and every future measurement are taken in the same place.

Operating conditions for the test. Load, speed, temperature, and the process conditions the machine must be at. This is where most of the argument actually lives, and it gets its own section below.

What happens if the machine fails. Who investigates, who pays, what the retest looks like, and what the tolerance is for accepting a machine outside zone A with a documented reason.

Test Conditions Decide the Result

A vibration acceptance test is not a single number; it is a number plus the conditions under which it was taken. Change the conditions and the number changes, sometimes by more than the width of a zone.

Thermal state. A machine measured cold and the same machine measured after hours at temperature can read very differently, because alignment changes as the machine grows. Acceptance readings should be taken at stable operating temperature, and stable needs defining rather than judging by eye.

Load. Vibration on many machines varies with load, and a test at whatever load the process happened to be running is not a test. The specification should name the load, or name several and require the reading at each.

Speed. On variable speed machines, a single speed is not an acceptance test. Resonances hide between the speeds you did not measure, and a machine that passes at 80 percent and fails at 60 percent has passed nothing.

Mounting of the transducer. How the sensor is attached changes the frequency range it can faithfully report. A hand-held probe and a stud-mounted accelerometer do not produce interchangeable results at higher frequencies, and mixing methods between acceptance and later monitoring destroys the comparison.

Auxiliary conditions. Piping strain, coupling condition, whether the machine is running solo or coupled. A pump that passes uncoupled and fails coupled has told you something specific about the alignment, and the test order should be deliberate rather than accidental.

The conditions that decide an acceptance reading. Change any of them and the number changes, sometimes by more than the width of a zone. Source: MIMAH engineering practice.

When a New Machine Reads Outside Zone A

The first move is not to reject the machine. It is to establish whether the reading is real.

Check the measurement before you check the machine. Wrong group or support classification, a transducer mounted badly, a reading taken at the wrong location, a machine not yet at thermal equilibrium, or a comparison against the wrong part of the standard will all produce a failing number on a healthy machine, and all of them are quicker to eliminate than a teardown.

If the reading survives that scrutiny, diagnose before negotiating. A single overall velocity number says the machine is vibrating but not why. The spectrum and phase say why, and the cause determines who is responsible. Unbalance, misalignment, looseness, a bent shaft, resonance and hydraulic or aerodynamic excitation all present differently, and they belong to different parties: unbalance is usually the manufacturer, alignment and soft foot are usually the installer, resonance is often the foundation designer, and process excitation may be nobody's fault at all.

Only then is it a commercial conversation. The useful outcomes are correction and retest, or documented acceptance in zone B with an agreed reason, a recorded baseline and a monitoring commitment. What should not happen is quiet acceptance with no record, because that converts a contractor's defect into an owner's problem at the moment of signature.

The Baseline Is the Real Prize

The acceptance test produces a pass or a fail, and everybody focuses on that. The more valuable output is the baseline: a full record of what this machine looked like when it was new and healthy.

Every later diagnosis is a comparison. A machine reading 3.1 millimetres per second means very little on its own, and means a great deal if the same point on the same machine read 1.2 at commissioning. Trend beats threshold, and the trend has to start somewhere.

A baseline worth having records the overall levels at every point and direction, the spectra behind them, phase readings, the operating conditions, the instrument and mounting method used, and the date. Filed where the maintenance team will find it in three years, not in the commissioning contractor's project folder. That record is what turns an acceptance test into the first data point of a condition monitoring programme rather than a piece of paperwork.

What a baseline must record. The acceptance test gives a pass or a fail; the baseline is what every later diagnosis is measured against. Source: MIMAH engineering practice.

The Disputes That Recur

Four arguments come up again and again, and all four are preventable in the specification.

The wrong group or support classification, which moves the acceptance threshold and is usually settled by going back to the machine's power rating, shaft height and the foundation design rather than by opinion.

Different test conditions between the manufacturer's works test and the site test. Works test readings are taken on a test bed, uncoupled, on a stiff foundation, at a convenient load. They frequently do not reproduce on site, and a contract that treats the works figure as the acceptance figure is going to fail. ASME and equivalent performance test codes are the usual reference for how a site test differs.

Comparing against the wrong reference, typically an operating alarm limit taken from a monitoring system rather than the acceptance zone in the standard. They are different numbers, set for different purposes.

No baseline recorded, so nobody can say later whether the machine has changed. This is the most expensive of the four, and it does not show up until years afterwards.

Getting It Right

Vibration acceptance is cheap to do properly and expensive to do late. The specification work costs an hour before the order goes out, the test costs a morning, and the baseline costs the time to write it down. Skipping any of it converts a fixable installation defect into a machine you own with a problem you cannot attribute.

MIMAH's maintenance and condition monitoring team writes acceptance specifications, witnesses commissioning tests, and sets up the baselines that later diagnosis depends on. If you have a machine due for handover, or one already installed and reading higher than you expected, get in touch.