An Oil Pump Running Does Not Prove That Every Bearing Is Lubricated

Startup on a forced-lubrication gearbox usually follows the same script. Turn on the pump. Wait for the pressure gauge to move. Release the drive. The gauge moves, the interlock clears, production starts.

The gap in that script is the distance between the pump discharge and the last bearing in the housing. Pressure at the discharge tells you the pump is pushing. It does not tell you that oil arrived, in the right quantity, at every point that needs it.

A Healthy Gauge and a Starved Bearing Can Coexist

The fault that makes this distinction worth learning is easy to describe and hard to see, because nothing about it is visible at the instrument the startup check is built around.

A distribution manifold feeds several bearings from one supply. Each outlet is drilled, each feed line is fitted with the restrictor the drawing calls for, and the whole arrangement is designed so that a defined share of the flow reaches each point. Change the manifold without changing the drawing, and the balance changes with it. An extra drain hole drilled during a repair, an outlet left open, a line running to a point that no longer exists: each of these creates a path of lower resistance than the bearing it was supposed to feed, and oil takes the easy path.

An unintended flow path may reduce oil delivery to one or more lubrication points. The pressure response depends on the pump characteristics, bypass and relief arrangements, distribution network, and measurement location. A normal pressure reading at one point does not establish that every bearing receives the required oil flow.

The reason this class of fault is worth a section of its own is not that it is common. It is that the startup check most plants use cannot distinguish it from a healthy system. A gauge at the pump discharge is a good instrument for the question it answers and the wrong instrument for the question that matters here. Where a population of identical units shows the same symptom, with pumps that all test good, the oil path is the place to look, because that is the one element the units share and the one element a pressure reading does not describe.

Pressure and Flow Are Not the Same Measurement

A pressure reading describes conditions at the gauge location; it does not measure flow at every bearing. A downstream restriction may reduce delivery while raising upstream pressure, while a bypass or additional leakage path may reduce pressure or redistribute flow. The response depends on the pump, relief and bypass arrangement, piping, leakage paths, and where pressure is measured. A gauge reading alone cannot establish bearing oil flow.

The engineering literature draws the distinction directly. Lubrication studies separate flooded from starved inlet conditions and define starvation by the thickness of the oil layer entering the contact, not by supply pressure. Once the inlet layer thins, the consequences are measurable: film thickness falls, average contact temperature rises, and the film pressure distribution shifts toward the Hertzian profile characteristic of dry contact. At that point the contact is no longer operating in a fully separated regime.

The magnitudes involved explain why the margin is thin. Elastohydrodynamic films in rolling bearings typically run on the order of a fraction of a micrometer to a few micrometers, and gear contacts are in the same range or thinner. A gear contact film of around a micrometer is a normal, healthy value. It is also roughly the thickness of a soap bubble's wall. Reduced lubricant supply can lower film thickness and increase the risk of mixed lubrication. Whether the contact actually transitions from full-film to mixed lubrication depends on the original operating margin, load, speed, lubricant properties and temperature.

Lubrication-film conditions depend on lubricant properties, surface roughness, contact geometry, speed, load, and temperature. Insufficient lubricant supply can increase the risk of mixed lubrication and surface damage, but no single film parameter provides a universal failure threshold for every bearing or gear contact.

There is a second mechanism that runs the other way and is worth knowing about. Debris blocking an inlet nozzle restricts supply, contact temperature rises, and at sufficient temperature the surfaces can scuff. Scuffing is not gradual wear; it is a rapid adhesive failure that can destroy a gear set in a short time. The window between "supply slightly reduced" and "catastrophic" is not wide, which is why the check belongs before startup rather than after the temperature alarm.

Following the Path Instead of Reading the Gauge

The useful diagnostic posture for a forced-lubrication gearbox is to treat the system as a sequence of segments and verify each one, rather than to accept a single pressure reading as a summary of the whole.

Segment one, the reservoir and pickup. Oil level inside the operating band, oil temperature in a range where the pump can actually draw, strainer clear. A pump that is cavitating because the oil is too cold and too viscous will show erratic pressure, and the downstream result is the same as a restricted line.

Segment two, pump and drive. Coupling intact, rotation direction correct, pump displacement as specified. A reversed pump on a unidirectional unit delivers nothing while sounding completely normal.

Segment three, the main line and its valves. Temperature control valve in the position the oil temperature calls for, overpressure bypass not stuck open, filter differential within its limit. A bypass that is passing oil continuously will hold system pressure while sending a large share of the flow straight back to the tank.

Segment four, the distribution manifold. This is the segment that usually carries the least instrumentation. Every outlet accounted for, every drilling as drawn, no open ports, no added drains, no crossed lines to points that do not exist on the drawing.

Segment five, the individual feed points. Each bearing and each mesh fed from a dedicated line, with the restrictor or orifice the drawing specifies. A missing orifice sends too much oil to one point and starves the rest; an extra one starves the point it serves.

Segment six, return and venting. Oil has to leave as well as arrive. A blocked breather raises internal pressure, which both disturbs return flow and loads the shaft seals from inside.

The point of walking the segments is one of diagnostic order rather than of failure statistics. Pump and distribution faults may affect pressure readings differently, depending on the pump characteristics, circuit arrangement and gauge location. A single pressure reading cannot reliably identify the failed section or confirm lubricant delivery to every required point.

The segmented view is also useful when reviewing a gearbox datasheet before replacement or installation. The lubrication arrangement, feed points, and instrumentation are configuration details to confirm for the specific unit. For an application review, see Mingye's industrial gearboxes range.

What a Startup Permit Should Actually Contain

Industry practice, as reflected in marine engine room procedures and in most gearbox commissioning documents, is to start the auxiliary lubrication system before the main drive, and to verify the alarm and interlock chain while the drive is still stopped. The logic is simple, because protective devices can only be proven when there is nothing to protect against yet.

A startup permit built on that principle has four parts.

Screening evidence, not proof of flow. A temperature comparison between bearing housings on the same manifold can identify an unusual pattern at comparable operating conditions, but it cannot verify oil flow to a bearing. Use only the limits and operating conditions specified for the particular unit. Confirm required flow at designated points using the manufacturer-approved circulation-test procedure and qualified personnel.

Verification of oil delivery. Inspect the lubrication circuit with the unit stopped, isolated, and depressurized. Verify actual oil delivery separately under the manufacturer's approved circulation-test conditions, using designated measuring points and qualified personnel. A static inspection alone cannot confirm operating flow.

Proof that the interlocks act where they are supposed to. Pressure switches are not self-verifying. A protective chain can have the right components and the wrong sampling geometry: an impulse line that reads pressure at a point upstream of where the pump actually delivers, or a tapping taken from a volume that stays pressurised after the pump slows, will trip the protection at a pressure the bearings never see. That class of fault is invisible to a component test, because every component in the chain is working. The way to catch it is to prove the interlock by the manufacturer's functional test procedure, using a simulated input or a test circuit where one is provided, rather than by creating a real oil interruption on a running machine. Proving it in the correct sequence, with the drive still stopped, is what makes the test meaningful.

Documented thresholds from the manufacturer. The specific pressure and flow values that constitute a permit to start are properties of the gearbox model, its oil grade, its operating temperature, and its piping. They are not transferable between units. A number quoted from another installation is a starting assumption, not a setting.

Why This Matters More on a Running Plant Than on a New One

New gearboxes are usually commissioned with a procedure and a witness. Gearboxes that have been in service for years are the ones that drift.

A manifold gets modified during a repair because a line was in the way. An orifice is left out because the correct one was not in stores. A bypass valve is left cracked open because the temperature control valve was sticking and someone wanted to keep running. A breather is painted over during a repaint. These changes may affect oil pressure, distribution or return flow in different ways. Some may change the gauge reading while others may not. Their effect on lubricant delivery must be established from the actual circuit rather than assumed.

The cumulative effect is a gearbox that runs a few degrees hot for a year and then fails in a way that gets attributed to the gear set. The temperature record, if anyone looks at it, shows the drift. The pressure gauge shows a healthy system the whole time.

The Check That Happens Before the Drive Is Released

The four parts above are a permit. What follows is the shorter version, for a unit coming back from overhaul or straight after work on its oil system, when the question is whether something was left in a state nobody documented.

Start with the manifold and the drawing side by side. This one is done with the unit stopped and locked out, and it needs no instrument at all: every outlet accounted for, every restrictor present and the right size, no added drains, no open ports, no lines running to points that do not exist on the drawing. It is the only check in this article that finds a diverted oil path directly rather than by inference, and it takes less time than filling out the permit the check belongs to.

A temperature comparison is screening evidence only; it does not confirm operating flow. Any observation under operation must follow the manufacturer-approved procedure for that model. Confirm required flow separately at designated test points using the approved circulation test and qualified personnel. Do not infer flow from a static inspection or a temperature difference.

The manifold inspection and bearing-temperature comparison are useful screening checks, but neither independently confirms the required flow at every lubrication point. Final acceptance must follow the manufacturer-approved lubrication verification procedure.

Where the path itself is being reviewed as part of a replacement or a retrofit, the review needs four things: the accurate model, the lubrication schematic, the location of the flow and pressure measuring points, and the current startup interlock scheme. With those four it is possible to say whether the oil route drawn on the schematic is the one actually plumbed into the housing, which is the same question the first check above asks with the unit stopped. Contact details are here if that review is the next thing on the list.