Skip to content
Faultpath
Chapter 1255 min read

Mechanical Power Transmission

Belts, chains, couplings, gearboxes and bearings — wear patterns and what they mean.

  • The worn part is the evidence
  • Belts, tension and glazing
  • Chain wear and matched sets
  • Gearboxes, oil and breathers
  • How bearings give warning

What it is

Everything in the previous chapters exists to turn a shaft. This one is about what happens between that shaft and the thing that actually does the work — belts, chains, couplings, gearboxes and the bearings that carry all of it.

The worn part is the evidence

That is the whole chapter in a sentence. A belt, a bearing or a sprocket that has failed is a physical record of the conditions it was working under, and those conditions are still there. Fit a new one without reading the old one and you have ordered the same failure again, usually with a delivery date.

This is where Conveyor 04's root cause lives: glazed drive pulley lagging, with black rubber dust under the frame and the take-up at the end of its adjustment. None of that is subtle once you are looking at it, and all of it was invisible from the panel.

The four families

FamilyTransmits byToleratesFails by
Belt driveFriction (V, flat) or teeth (timing)Misalignment, shock, dirtSlip, glazing, cracking, snapping
Chain drivePositive engagementDirt, heat, shock loadsStretch, wear, seizure
CouplingDirect connectionSmall misalignment onlyElement wear, backlash, breakage
GearboxMeshing teethAlmost nothing out of specTooth wear, bearing failure, oil loss

The ordering matters. Belts are the most forgiving and the most likely to be slipping quietly; gearboxes are the least forgiving and the most expensive to get wrong.

How it works

Belts

A V-belt transmits by wedging into the pulley groove: tension pulls it in, and friction on the flanks does the work. Two consequences follow.

First, a V-belt should never touch the bottom of its groove. If it does, the belt or the pulley is worn, the wedging action is gone, and the drive is running on whatever friction is left.

Second, tension is a specification, not a preference. Too loose and it slips, glazes and burns; too tight and it destroys the bearings either side of it. That second failure is the sneaky one, because it produces a bearing fault that gets diagnosed as a bearing fault.

THE WEAR IS THE EVIDENCEWhere it is, and whether it is even, names the cause.GLAZEDSlip — the belt is not grippingPolished, shiny, hard. Conveyor 04.ONE-SIDEDMisalignmentWear on one flank only.FRAYED EDGERubbing a guard or flangeDamage outside the working face.CRACKEDHeat, age, or a pulley too smallCracks across the belt back.A worn part is a written record of what was wrong with the machine.Fit a new one without reading it and you have ordered the same failure again.
Fig 12.1 — Four wear patterns, four causesGlazing means slip. Wear on one flank means misalignment. Damage outside the working face means it has been rubbing something. Cracks across the back mean heat, age, or a pulley too small for the section. The pattern is the diagnosis and it takes ten seconds to read.

Timing belts transmit by teeth rather than friction, so they do not slip — they skip, and a skipped tooth is a lost position rather than a lost speed. Machines that home themselves after a stop are frequently compensating for exactly this.

Chains

A chain wears at the pins and bushes, which lengthens the pitch — so a "stretched" chain has not stretched at all, it has worn. The consequence is that it no longer sits correctly in the sprocket teeth, and it starts riding up toward the tips, which then wears the sprocket into a hooked profile.

Two practical points:

  • Fit a new chain to worn sprockets and it will not last. The hooked teeth destroy the new chain quickly. Chains and sprockets are replaced together, and a technician who has been given only a chain should say so.
  • Chain drives need lubrication where the wear is — between pin and bush, not on the outside where it is easy to reach and where it collects dirt.

Couplings

A coupling exists to join two shafts, and every type has a limit on how much misalignment it will accept. Flexible couplings tolerate a little; rigid ones tolerate none.

The thing to understand is what a coupling does with misalignment it cannot absorb: it passes the load into the bearings on both sides, as a cyclic force at running speed. So the coupling's element wears, and the bearings fail — and if you only replace the element, both come back. Chapter 13 covers alignment properly.

Gearboxes

The three things that kill gearboxes, in order of frequency: oil (wrong level, wrong grade, contaminated, or never changed), overload (shock loads, jams, a drive with its current limit set generously), and misalignment transmitted through the input or output shafts.

Reading a gearbox is mostly reading its oil. Metal in the oil is normal at running-in and abnormal later; water makes it milky; a burned smell means it has been hot. A breather blocked with product is a common and cheap cause of seal failure, because the box cannot equalize pressure and pushes oil out past the seals.

Bearings

Most rotating failures end at a bearing, and most bearings fail for a reason that is not the bearing: misalignment, over-tensioned belts, contamination, the wrong lubricant, too much lubricant, or a bearing current from a drive (Chapter 6).

BEARINGS GIVE WARNING — MOST OF IT EARLYThe symptoms arrive in a fixed order, and they are not evenly spaced.newfailedULTRASOUNDmeasurable changeVIBRATIONspectrum shiftsAUDIBLEyou can hear itHEATyou can feel itFAILUREBy the time you can hear it, most of the warning has already been spent.A bearing you can smell or feel is a bearing to plan around today, not next month.
Fig 12.2 — The order the warnings arrive inUltrasound and vibration change first, then it becomes audible, then hot, then it fails. The intervals are not equal — the gap between audible and failed is a small fraction of the gap between detectable and audible, which is the entire argument for condition monitoring.

That ordering is worth internalizing because it changes what a noisy bearing means. A bearing you can hear is not at the start of its decline; it is near the end. One you can smell or feel through the housing is a machine to plan around today rather than next month.

Tension, alignment and lubrication are the three levers

Nearly every mechanical drive fault comes back to one of them being wrong. Tension has a specification; alignment has a tolerance; lubrication has a type, a quantity and an interval. All three are routinely set by feel, and all three produce failures that look like component quality problems.

What normally fails

Symptom
A drive squeals on start-up, and it is worse under heavy load
Likely cause
Belt slip — insufficient tension, glazed belt or pulley, or a load that has increased
How common
Very common

The squeal is the belt slipping against the pulley, and slip polishes both surfaces, which reduces grip further. Conveyor 04 is this failure allowed to run for a week. Check the tension against specification, look for glazing, and ask whether the load has changed before simply retensioning.

Symptom
Bearings failing repeatedly on the same shaft, often within months
Likely cause
Something is loading them — misalignment, over-tension, or a coupling passing what it cannot absorb
How common
Very common

The bearing is the consequence, not the cause. Two bearings in a year on the same position is a machine telling you clearly that the fault is not the bearing, and fitting a third without addressing tension or alignment is buying the same failure a third time.

Symptom
A chain runs rough, climbs the sprocket, or jumps under load
Likely cause
Pitch worn beyond limit, and sprocket teeth hooked to match
How common
Common

Wear at the pins and bushes lengthens the effective pitch until the chain no longer seats correctly. Because the sprocket wears to match, replacing only one of the two fails quickly. Measure the chain against its wear limit rather than judging by how it looks.

Symptom
Oil weeping from a gearbox seal, or oil level dropping steadily
Likely cause
A blocked breather pressurizing the case, or a seal that has hardened with heat
How common
Common

The breather is the cheap cause and the one that gets overlooked. If the box cannot equalize pressure it pushes oil out past the seals, and replacing the seal without clearing the breather repeats the fault. Check what the oil looks like while you are there.

Symptom
A coupling element wears out repeatedly, sometimes with a rhythmic knock
Likely cause
Misalignment the coupling is absorbing on every revolution
How common
Common

A flexible element is designed to accommodate a small amount and it wears in proportion to how much it is being asked to take. Repeated element failure is an alignment measurement waiting to happen — and it is also loading the bearings each side. Chapter 13.

Symptom
A timing belt drive loses position, and the machine needs homing more often
Likely cause
Belt teeth skipping under load — worn belt, worn pulley, or insufficient tension
How common
Occasional

Timing belts do not slip, they skip, and the symptom is lost position rather than lost speed. Machines that have quietly acquired a more frequent homing routine are often compensating for this.

Symptom
A drive runs hot with no obvious mechanical fault
Likely cause
Over-tension, over-lubrication, or misalignment — all of which turn energy into heat
How common
Occasional

Grease is not free: a bearing packed full has no space for the rolling elements to move lubricant around, and churns it instead, which produces heat. More is not better, and this is one of the few maintenance tasks routinely made worse by enthusiasm.

How to troubleshoot it

  1. Isolate, prove dead, and account for stored rotation

    Safety

    Chapter 2, with the mechanical specifics: a large rotating mass coasts for a long time after power is removed, and springs, tensioners and raised loads all hold energy. Confirm the machine has actually stopped rather than assuming it has.

  2. Turn it by hand before measuring anything

    The single most informative action in the chapter and the one most often skipped because it is not electrical. Stiff, notchy, rough or seized tells you more in ten seconds than any instrument, and it distinguishes a mechanical overload from an electrical fault immediately — which is Chapter 5's whole diagnostic branch.

  3. Read the wear before you remove it

    Photograph it in place. Which flank, which side, how even, how far around. Once a belt is off the machine, the information about which way it was fitted and where it was rubbing is gone, and that information is usually the diagnosis.

  4. Check tension against the specification, not against feel

    A gauge, or the deflection method from the manufacturer's figures. "About right" is how belts get fitted too tight, and an over-tensioned belt destroys bearings on both sides of the drive — a failure that then presents as a bearing problem.

  5. Check alignment while everything is still assembled

    A straightedge across the pulley faces catches gross misalignment in seconds. Anything better than that needs Chapter 13, but gross misalignment is common enough that the cheap check is worth doing first.

  6. Look at the lubricant, and at the breather

    Level, grade, color, smell, and whether there is metal in it. Then check the breather is clear, because a pressurized case pushes oil past seals and produces a leak that no seal will fix.

  7. Ask what changed to load it more

    More load means more slip, more heat and more wear. Product changed, speed increased, a downstream jam becoming routine, a bearing elsewhere dragging. The component that failed is often reporting on something upstream of itself.

  8. Replace matched parts together, and record what you saw

    Chain with sprockets, belts as a matched set, coupling elements with an alignment check. Then write down the wear pattern, because the next person's diagnosis is your observation plus theirs — and on a repeat failure that comparison is the whole answer.

Common technician mistakes

  • Fitting the new part without reading the old one

    Why

    The failed component is in your hand, the replacement is on the bench, and the machine is down — so it goes in the bin and the new one goes on. But it was a written record of the conditions that killed it, those conditions have not changed, and you have just started the same clock again.

  • Tensioning a belt by feel

    Why

    Everybody develops a feel for it and the feel is almost always too tight, because a tight belt does not squeal and a squealing belt is an obvious complaint. The bearings pay for it quietly on both sides of the drive, months later, and the failure gets recorded as a bearing problem.

  • Replacing a chain and leaving the sprockets

    Why

    The chain is the part that is obviously worn and it is the cheaper of the two, so it is the one that gets ordered. Hooked sprocket teeth then destroy the new chain in a fraction of its life, and the machine acquires a reputation for eating chains.

  • Adding grease because grease is good

    Why

    More lubricant feels like more care, and greasing is a task with a visible completion. A bearing packed solid has no room for the elements to move lubricant around, so it churns it and runs hot — and over-greasing also blows seals, which lets contamination in. Quantity and interval are specifications.

  • Treating a bearing you can hear as a job for next month

    Why

    It has been making that noise for a while and the machine is still running, so it goes on the list. But audible is late in the sequence, not early: most of the warning was spent before anyone could hear it, and the interval from audible to failed is short and shortening.

  • Running it with the guard off to watch the fault

    Why

    The fault is rotational, so watching it rotate is genuinely the fastest diagnosis available — which is exactly why this one is so tempting and so common. It puts you next to unguarded moving machinery for the sake of convenience, and restoration and test running are already where a large share of injuries happen.

Hands-on challenge

Scenario

Blower B-2 — third bearing in fourteen months

Blower B-2 is having its third drive-end bearing fitted in fourteen months. The non-drive-end bearing has never been changed.

What you establish:

  • The blower is belt-driven from a 20 hp motor, three V-belts in a matched set.
  • The belts are replaced at each bearing failure and are tensioned by feel; there is no tension gauge on site.
  • The fitter who usually does it says he makes sure they are tight enough that they definitely do not slip.
  • The removed belts show even wear on both flanks and no glazing.
  • The bearing housings and the motor feet show no signs of having been disturbed since the machine was installed.
  • Grease is applied on a monthly schedule with a grease gun, no quantity specified.

Write down: what three failures in the same position tells you before any measurement, which of the three levers you suspect and why, what the belt wear pattern supports and what it argues against, why it is the drive end failing, and what you would do differently this time.

Show how to approach it

A repeat failure is a different question from a first failure. The machine has already told you the bearing is not the fault.

  1. Three bearings in fourteen months is the diagnosis, not the symptom. Bearings do not have a quality problem three times running in the same position. Something is loading them.
  2. The belt tension is the strongest candidate. Set by feel, no gauge on site, and "we make sure it doesn't slip" describes a drive that is tensioned until the squeal stops — which is reliably too tight.
  3. Over-tension loads the bearings either side of the drive with a constant radial force they were never sized for, and the failure appears months later as a bearing problem.
  4. The wear pattern on the removed belts should be read. Even wear on both flanks with no glazing supports over-tension; wear on one flank would move alignment to the top of the list instead.
  5. The drive-end failing first is consistent with a belt-tension load rather than with a general lubrication or contamination problem, which would not favor one end.
  6. The job is a tension gauge and the manufacturer's figure, an alignment check while it is apart, and a note in the record of what the tension actually measured before adjustment — so the next failure has something to be compared against.

Knowledge check

Five questions. Each is answerable by reasoning from what a worn part looks like, rather than by recalling a maintenance interval.

Question 1 of 5

The same bearing position on a machine has failed three times in fourteen months. Each time a new bearing of the correct part number was fitted. What does the repetition tell you?