Hot Industrial Gearbox Coupling After Cold Alignment: What to Check

An alignment check passes. The readings are inside tolerance, the report is signed, the machine goes into production, and two weeks later the coupling is running hot enough to be uncomfortable to touch, with grease weeping from the seal and a vibration reading that has crept up.

The natural conclusion is that something is wrong with the gearbox or the coupling. One hypothesis worth testing early is that the alignment was correct for a machine that no longer exists, because the machine that is running is hot, and thermal growth has moved shafts relative to one another. It is one candidate among several: the coupling itself, its grease, a change in process duty, a loosened base, and an overloaded drive all produce the same symptom. Nothing in this article is an argument for a verdict about the gearbox before the evidence is in.

A cold alignment measures one condition: the relative position of two shafts at ambient temperature with no process heat in the machine. What the machine runs in is a different condition, and the difference is not a small correction.

What follows is why the thermal explanation deserves the first look: not because it is the only one that fits, but because it is the one a passed alignment report has no information about at all.

What "Aligned" Actually Has to Mean

The engineering literature on shaft alignment makes a distinction that gets lost in practice. A complete alignment has two elements, not one.

The first is the shaft-to-shaft measurement, which is what a dial indicator or a laser system produces: the offset and angular relationship between two shaft centerlines at the coupling.

The second is the measurement of thermal growth and dynamic displacement at each bearing housing, from ambient to operating temperature. Both elements are required to determine the cold alignment position that will put the machine correctly aligned while it carries load at temperature.

Without the second element, the first is a measurement of a machine in a state it never operates in.

The reason this is not merely theoretical is the nature of the task. Aligning a machine while it is running is difficult or impossible, so a cold alignment target has to be set in advance to compensate for thermal growth, with the objective that the shafts are correctly aligned when the machine reaches normal operating temperature. That is what a cold alignment target is for.

Alignment readings taken promptly after an approved shutdown can help estimate thermal displacement relative to the cold condition. Cooling during measurement and the loss of operating forces may affect the result, so post-shutdown readings do not necessarily capture dynamic displacement that exists only while the machine is running.

Post-shutdown readings can help estimate displacement relative to the cold condition, but cooling during measurement and the loss of operating forces can affect the result. They do not necessarily capture the full displacement present while the machine is running.

Why the Growth Is Not Symmetrical

If every component grew by the same amount in the same direction, thermal growth would be irrelevant. It is not, for three reasons.

The components run at different temperatures. A gearbox housing, its high-speed shaft, a motor frame, and a driven machine's bearing pedestal all settle at different temperatures, because they have different heat sources, different cooling, and different masses. The gearbox is heated by its gears and bearings. The motor is heated electrically. A driven roll or a fan is heated by the process.

They are made of different materials. A cast iron housing, a steel shaft, and an aluminum component expand at different rates, and growth is proportional to length as well as to coefficient, so a long shaft grows more than a short one at the same temperature.

The growth is in three dimensions. A machine sits on feet, and the feet are at different heights. Vertical growth of a gearbox relative to its base changes the shaft height, and a cold alignment carries no information about it at all. Horizontal growth acts along each component's own axis, and because the gearbox and the driven machine react against anchors that are not shared, it does not cancel. Which component of the total movement is the larger is a property of the installation: it depends on the pedestal heights, the anchor arrangement, the temperature each part reaches and the direction its growth is constrained to take. There is no general answer, which is why both have to be measured on the machine in hand rather than one assumed to dominate.

A machine aligned within tolerance at ambient temperature may operate outside the intended alignment condition after thermal growth and load-induced displacement. Whether this explains an elevated coupling temperature, and when symptoms develop, must be established from measurements on the actual machine rather than assumed from the cold alignment result.

Thermal displacement is one possible explanation for a coupling that runs hot after installation. Its contribution and the timing of the symptoms must be established from measurements on the actual machine.

What Misalignment Costs, and Where It Shows

The consequences of running misaligned are documented in the technical literature from bearing and alignment equipment manufacturers, and the list is consistent across sources: increased friction and energy consumption, premature bearing and seal failure, premature shaft and coupling failure, excessive oil leakage at seals, coupling and foundation bolt failure, and increased vibration and noise.

Published guidance from one bearing manufacturer attributes up to half of the cost associated with rotating machinery failures to misalignment. Read that figure for what it is: a manufacturer's published assessment, not an independent industry survey. Its value is in the order of magnitude, not in the decimal place. Half or a fifth, the conclusion holds, which is that misalignment is not a marginal condition to be tolerated until a convenient shutdown.

Where it shows first, on a machine that was aligned cold and runs hot:

Coupling temperature. A flexible coupling absorbs misalignment by deforming. Deformation costs energy, and the energy becomes heat in the elastomer or in the metallic flexing elements. An infrared reading of the coupling, compared against the adjacent shafts and against the same machine's own history at the same duty, is the fastest available indicator. There is no universal temperature threshold to quote across couplings, because the permitted operating temperature is a property of the coupling's materials and its own rating. What the comparison is for is the change: a coupling that has moved up relative to its own baseline at the same load, or that sits hotter than the shafts either side of it, is absorbing more than it did. The number that governs on any particular unit is the coupling manufacturer's own temperature limit.

Coupling grease or elastomer condition. Grease that has separated, turned dark, or been thrown out of the coupling indicates sustained heating. An elastomer element that has taken a permanent set or shows cracking at the flex points indicates that it has been running at high deflection continuously.

Bearing temperature at the points nearest the coupling. The bearing closest to a misaligned coupling carries the restoring forces. A temperature difference between the drive-end and non-drive-end bearings of the same shaft is a signal, as is a difference between two otherwise identical machines.

Vibration at twice running speed. Misalignment can produce vibration at 2× shaft speed, often with axial components that are less typical of other fault types. It is one of several things that raise a 2× component, and it cannot confirm anything on its own: a bent or bowed shaft, looseness, a coupling with worn elements, resonance, and a mis-set machine all show up in the same region of the spectrum. A rising 2× with a strong axial reading is a reason to look at alignment and a reason to keep the other causes open until each has been tested. A 2× component is a clue, not a diagnosis. Assess it alongside coupling condition, lubrication, load, thermal movement, phase, and measurements from appropriate locations; no single spectrum feature uniquely identifies alignment.

Seal leakage at the gearbox input. Misalignment loads the input shaft bearing and the seal, and the seal's ability to maintain its film depends on the shaft running concentric. A seal that starts weeping after an alignment or a motor change is reporting the alignment.

Rebuilding the Alignment Requirement

The practical route from a hot coupling to a fix runs through five measurements, and doing them in this order prevents a wasted shutdown.

Post-shutdown readings can help refine an alignment target, but their limitations must be recorded. The estimated displacement depends on measurement timing, cooling, mounting conditions, and any load-induced movement that disappears after shutdown.

The tolerance to which the hot alignment should be set is a separate question, and it is not a universal figure. Permissible offset and angular deviation are set by the coupling manufacturer, who knows what that coupling can absorb, and by the machine manufacturer, who knows what the bearings can accept. Quoting a general tolerance across different machines and different couplings is not conservative, it is arbitrary.

That is also where the industrial gearboxes side of the question sits: the bearing arrangement at the input shaft and the shaft end geometry determine what the unit can accept from a coupling that is working at the edge of its range, and those are configuration facts rather than catalog ones.

The Second Element: Dynamic Displacement

Thermal growth is not the only thing that moves a shaft between the stopped condition and the running condition.

A machine under load deflects. Belt tension pulls a shaft toward the driven pulley. Gear tooth forces push shafts apart along the line of centers. A driven fan or a pump develops hydraulic or aerodynamic thrust. Flexible mounts compress under load and under torque reaction.

All of these are present at operating condition and absent when the machine is stopped. So an alignment that accounts for thermal growth but ignores load-induced displacement is still incomplete on a machine where those effects are large.

The practical way to judge whether they matter: if the machine has flexible mounts, a belt drive with significant tension, or a driven machine with high thrust, the static-to-running displacement is worth measuring. The route to that measurement is through permanently installed monitoring points, a closed inspection window, or a reading taken with the machine stopped and isolated, and the last of these is often enough: a dial indicator on a shaft showing the deflection between the unloaded and the loaded state gives the magnitude directly. A guard is not something to remove from a running machine to get a reading, whatever the alignment question it would answer.

Method Matters Less Than Repeatability

Alignment instruments differ in accuracy and convenience, and the published comparison is worth knowing because it sets expectations: a straightedge is fast and easy and inaccurate, a dial indicator is accurate and slow and awkward, and a laser system is accurate and fast and easy to use.

What matters more than which instrument is used is that the same instrument, in the same setup, produces the readings compared against each other. Cold readings and hot readings taken with different instruments, from different reference points, or by different people are not comparable, and the thermal growth figure derived from them will contain the difference.

The other practical requirement is recording. An alignment that is not documented cannot be repeated, and the second alignment of the same machine is where the value of the first one's record becomes obvious.

Before You Move the Machine

A hot coupling is evidence of energy being dissipated somewhere it should not be, and misalignment is only one mechanism that produces it. Working through the alternatives costs an hour and can save a shutdown.

Three of them are mechanisms in the coupling itself. Over-greasing is the cheapest and the most often overlooked: a coupling packed beyond its intended fill churns its own grease, and churning generates heat without any misalignment at all. The signature is a coupling that runs hot from the first hour after a maintenance grease-up and cools down when some of the excess is purged. The wrong grease behaves the same way for a different reason, because a coupling specified for a high-speed application and greased with a heavy, high-viscosity product will run hot at speed regardless of alignment. The coupling manufacturer's specification is the reference, not whatever is in the plant's grease gun. Wear in the coupling itself raises temperature in two ways: a worn element deflects further than a new one under the same misalignment, and lost material changes the balance.

The fourth is not a mechanism at all. It is the case where the heat is real but the coupling is a casualty rather than a cause, because the driven equipment has been pushed to a higher duty and the whole machine runs hotter. A coupling that has been getting hotter alongside its gearbox and its driven machine is reporting the process. The check there is a comparison against the machine's own history at the same production rate, not against a rule of thumb.

What helps separate misalignment from those is not a single reading but a pattern: misalignment heat tends to track load and speed, to recur at the same measured position after a shutdown and restart, and to correlate with a change in the machine's condition rather than with a change in its maintenance. Each of those is consistent with other causes as well, so the pattern is evidence to weigh rather than a test that returns an answer.

Thermal displacement is one possible explanation for a coupling that runs hot after installation. Its contribution and the timing of symptoms must be established from measurements on the actual machine.

If the evidence points at alignment, use measured thermal growth and an alignment target calculated for the machine's actual operating condition. The required adjustment depends on the installation and duty, and the symptom may recur depending on those conditions; no fixed recurrence period can be inferred from a cold alignment alone.

Recorded measurements and validated estimates provide the basis for assessing the alignment target; they do not automatically establish the full displacement present during operation. Share the relevant readings to discuss the drive interface against the measured evidence.