Why a New Noise Isn't Always a New Problem

Why this matters

A customer calls saying the unit "just started making a noise," and the tech's first instinct is to look for what changed today. Half the time, nothing changed today. A bearing that has been drying out for eight months, a fastener that has been backing out for a season, a belt that has been glazing since spring: none of these are new. What is new is that the wear finally crossed the threshold where it makes sound a person notices, or the person finally stood in the room long enough to notice it. Chase "what happened today" on a slow-wear fault and you will find nothing, because nothing happened today except the wear finishing its arc. Understand that noise onset and fault onset are two different timelines, and you stop wasting the visit looking for a trigger event that does not exist.

Two different timelines, one confusing symptom

Every audible fault has two dates that matter, and they are usually not the same date.

• Fault onset: when the underlying condition actually began. A bearing starts losing lubricant, a mount starts loosening, a valve seat starts eroding, months or years before anyone hears it.
• Noise onset: when the sound crosses the threshold of being audible, or being audible enough that a person in the space actually registers it as "new."

The gap between these two dates can be long, and it is exactly proportional to how gradual the wear is. A bearing degrades on a curve, not a cliff: quiet, then a faint high-pitched hint only a trained ear catches, then a growl anyone would hear, then a grind. The customer's report ("it just started") is accurate about the noise. It says nothing reliable about the fault.

Why customers genuinely cannot tell the difference

This is not a customer being inattentive. A few real mechanisms explain why a slow-building noise gets reported as sudden:

• Habituation. People stop consciously hearing a sound that has been present and gradually increasing for a long time, the same way you stop noticing a clock ticking in a room you live in. The sound crosses back into awareness only when it steps up a level (a new pitch component, a new loudness) or when something draws attention to the space (a guest comments, the person changes rooms, a quiet night highlights it).
• A schedule change exposes an existing noise. A unit that has always made this sound during the day, when the house is busy and loud, gets noticed the first night someone is home and the house goes quiet. Nothing changed on the equipment; the listening conditions changed.
• A threshold effect in the fault itself. Some wear genuinely does produce a step change in sound level at a specific point, not a smooth ramp: a bearing race finally pits enough to rattle, a belt finally glazes enough to slip audibly. In these cases the customer's "sudden" report is closer to literally true, even though the underlying wear was gradual.

What this means for how you diagnose

Do not spend the first ten minutes hunting for an event that triggered the noise, unless the customer can independently confirm one (a storm, a power blip, a recent repair, a recent modification). Absent a real trigger event, treat the noise as a progressive wear fault reaching an audible threshold, and diagnose it the way you would any wear fault: identify the component class by the sound's character (see the field guide on distinguishing bearing sounds from belt sounds), then confirm by touch, by isolation, or by direct inspection, not by asking what changed.

Ask the right version of the timeline question. Instead of "when did this start," which invites an answer anchored to when the customer first noticed, ask "has this space been quieter or louder than usual lately" or "did anything change about when you're normally in this room." That question surfaces a listening-condition change if one happened, without leading the customer toward a false trigger event they will otherwise invent to be helpful.

Do not let this cut the other way

None of this means every noise report is old wear finally surfacing. A genuinely new problem does exist, and a real trigger event (a part swap, a power event, a repair visit, an install) deserves the same attention it always did. The skill here is not assuming every noise is old; it is not assuming every noise is new either. Cross-check the customer's timeline against physical evidence: a wear pattern consistent with months of degradation (glazing, uneven pitting, a worn mount) confirms the slow-onset read. Fresh tool marks, a recently disturbed connection, or a component that looks new but sounds wrong points to an actual recent event, and that changes where you look first.

The conversation that keeps trust intact

Tell the customer plainly what you found: "this bearing has been wearing for a while, it just crossed the point where you could hear it." That is more useful to them than confirming their assumption that something broke today, because it sets the right expectation (this was progressing, not sudden failure) and it explains why you are recommending a repair for something they only just noticed. Framing it as "caught early" rather than "ignored a long time" keeps the conversation forward-looking instead of defensive on either side.

References

• Trade-standard practice for progressive wear-fault identification by sound signature
• Manufacturer documentation on component wear-life curves and audible-threshold behavior
• See related: The Difference Between a Bearing Fault Sound and a Belt Fault Sound
• See related: Two Different Noises Appear at Once (decision tree)