The Replacement Fan Fit the Opening but the Cabinet Still Ran Hot
An anonymised equipment integrator replaced an obsolete cooling fan with the same nominal frame size, but the cabinet temperature rose during loaded operation.
The original fan had become difficult to source. Its outline, mounting holes and connector appeared close to a readily available replacement, so the first screening decision was made from the label and the mechanical drawing. The replacement installed without forcing the frame and reached a normal-looking RPM. On the bench, its free-air airflow also appeared acceptable. The cabinet result was different: the hottest drive and power-supply region rose faster, and the controller increased fan command more often than expected.
The team initially suspected a faulty sample. A second unit showed the same behavior. The investigation then separated three questions that had been combined under the word replacement: does the fan fit, does it deliver the required airflow at the cabinet pressure, and does its electrical control interface behave like the original? Matching the outline answered only the first question. The purchasing comparison was rewritten so that each candidate carried an operating-point record, a control-interface record and a production-assembly record. That made the next supplier discussion much more precise.
Customer identity and measured acceptance values remain confidential. This case is an anonymised engineering pattern, not a claim that one replacement family fits every cabinet. A fan substitution must be checked against the actual air path, voltage, controls, environment and system acceptance method.
Freeze What the Original Fan Actually Did
A part number and outline drawing are useful starting points, but they are not a complete replacement specification.
The review record collected the original fan's frame and thickness, mounting pattern, airflow direction, connector keying, lead length, voltage, current, control type, FG/RD convention, speed range, noise condition and environmental basis. It also recorded the cabinet filter, grille, heat sinks, cable routes, service panels and the temperature locations that defined acceptance. Several of these details had never appeared on the purchasing description.
The team then photographed the installed fan and marked the inlet and outlet boundary. A fan with the same square outline can have a different impeller, motor curve, pressure capability, bearing behavior or cable exit. A different guard or gasket can reduce the free opening. The replacement request therefore included a system sketch rather than only a dimensioned rectangle.
The original operating point was treated as evidence to be reconstructed, not as a number to copy blindly. Where the old P-Q curve was unavailable, the team measured airflow and pressure in the cabinet and used the original command, current and temperature response as additional clues. The goal was to define the duty the replacement had to reproduce. The record also noted whether the original measurement was taken with a clean filter, a closed service door and the production heat load. Without that context, two apparently different fan results could simply reflect two different test boundaries.
| Compatibility layer | What to record | Why it can fail |
|---|---|---|
| Mechanical | Envelope, holes, guard, gasket, airflow direction | Frame fits but opening or outlet is restricted |
| Operating point | Airflow and pressure in the installed cabinet | Free-air CFM does not equal useful cabinet airflow |
| Electrical | Voltage, current, startup and protection | Harness or supply cannot support the replacement |
| Control | PWM, FG/RD, alarm and minimum duty | Controller reads the wrong state or hunts |
| Environment | Temperature, dust, humidity, vibration and service | Catalogue conditions do not represent the cabinet |
Find the Real Operating Point Instead of Comparing Free-Air CFM
The candidate failed the first comparison because the cabinet resistance curve was missing from the decision.
The installed path included a filter, inlet grille, fan guard, narrow clearance, heat-sink fins and an outlet opening. Each element consumed pressure. The team measured pressure at defined points and checked airflow through the heat-producing region. It did not use a single pressure number without saying whether it was across the fan, the filter, the heat sink or the complete cabinet.
The fan P-Q curve and the cabinet system curve were then considered together. At zero pressure, the replacement's free-air airflow could look similar to the original. At the pressure created by the filter and heat sink, the two candidates separated. The replacement reached a different intersection, so its normal RPM did not mean normal cooling. The pressure budget also showed which component had to change first, instead of turning every thermal complaint into a request for more motor speed.
This also explained why increasing the command gave only a partial improvement. Higher RPM could add airflow, but it increased noise and current, and the cabinet restriction still consumed much of the additional pressure. The team first corrected the operating-point comparison, then decided whether a different fan curve or an air-path change was needed.
Separate Fan Capacity Loss from Air-Path Loss
A replacement investigation can waste time if a dirty filter, blocked grille or bypass is blamed on the fan.
The cabinet was checked in clean, loaded and service-limit filter states. The team measured the filter differential pressure and inspected the seal around the frame. A gap can allow air to bypass the heat sink, while a clogged filter can move the operating point without changing the fan's own curve. Both conditions can produce a hot cabinet and a normal-looking fan signal.
The inlet and outlet were also checked for recirculation. Warm discharge air could return through a side gap and make the replacement appear thermally weak. The review used the production cover, cable bundle, blanking pieces and service panels. A temporary open-door test was kept separate from the closed-cabinet acceptance condition.
The result was a list of fan-side and system-side actions: select more pressure at the required airflow, increase filter area, repair a seal, clear a grille, change a baffle or revise the outlet. The integrator retained responsibility for cabinet thermal architecture; the fan supplier's task was to make the fan evidence match the defined operating point.
Check Voltage, Startup Current and Control Compatibility
Two fans can share a connector shape and still behave differently when the cabinet controller starts or modulates them.
The replacement was tested at the voltage available at the fan connector, including harness drop and the lowest specified supply. Startup current, time to valid rotation and steady current were logged. If several fans start together, a different inrush profile can trip the supply or cause a brief controller reset even when the steady-state current looks acceptable.
The control review identified whether the original used voltage control, PWM, FG, RD or a simple alarm contact. The replacement's logic level, reference, pulse convention and minimum stable duty were checked. A controller may interpret a tach pulse count differently, or it may treat a missing signal as a fault during a period when the replacement is still starting.
The team also looked for control hunting. If the replacement produces less airflow at the same command, the cabinet temperature can cross the controller threshold more often. The controller then changes speed repeatedly, increasing noise and current without establishing a stable thermal state. The replacement curve and the control curve had to be reviewed together.
| Check | Failure mode | Evidence |
|---|---|---|
| Connector voltage | Startup fails at the real harness voltage | Voltage, current and start trace |
| Inrush current | Several fans trip the supply | Synchronized start measurement |
| PWM behavior | Duty does not map to the expected RPM | Reference, frequency, duty and RPM |
| FG/RD convention | Healthy fan becomes a false alarm or vice versa | Signal and controller response |
| Minimum duty | Fan stalls or controller hunts | Low-speed stability and airflow |
Use Noise and Vibration as Compatibility Evidence
A mechanically compatible replacement can still create a new tonal peak, panel resonance or mounting load.
The replacement was measured at the same position and operating states as the original. Overall sound level was not considered sufficient because a narrow tonal peak can be more noticeable than a broadband change. The team correlated RPM, PWM command, pressure, airflow and the acoustic spectrum. This showed whether a tone followed the motor speed, the control frequency, the grille or the cabinet panel. The microphone position and bandwidth were kept unchanged for the comparison.
Mounting preload and frame support were checked because a different frame stiffness or fastener response can transmit more vibration. The cable exit was inspected for contact with an isolator or cover. The team also checked whether the replacement's airflow direction and guard created a different turbulence pattern at the grille.
A quiet result was not accepted if it came from lower useful airflow. Conversely, a higher sound level did not automatically mean the fan was unsuitable if the system could change the operating band without losing thermal margin. The decision used a defined tradeoff between temperature, pressure, current, noise and service conditions.
Validate Temperature, Filter Loading and the Worst Duty
The replacement was evaluated in the condition that defined cabinet acceptance, not only during a comfortable room-temperature run.
The test sequence included the highest heat load, warm inlet air, clean filter, loaded filter and a recovery period. Sensors remained at the original acceptance locations so the comparison would not change its boundary. Fan voltage, command, RPM, current, pressure and airflow were logged with the cabinet temperatures.
The team separated average cabinet temperature from the limiting component. A replacement may reduce the average while leaving a local heat sink hot, or it may cool one area while causing recirculation elsewhere. The location and response time of the sensor mattered as much as the final number.
Filter service was treated as part of the replacement decision. If the original fan had enough pressure margin when clean but the replacement lost most of its airflow after loading, the correct answer could be a different pressure curve or a maintenance trigger. The fan was not accepted on a clean-filter result alone.
Review Service, Life and Change Control Before Release
A replacement is only useful when the next technician can install it without changing the tested cooling boundary.
The pilot work instruction showed airflow direction, connector orientation, cable routing, frame clearance, fastener method, gasket position and filter seating. Service access was checked with the fan installed. A replacement that fits during engineering assembly may still be impractical when a technician must remove a cover or work around a harness.
The life discussion used the real temperature, speed, pressure, voltage, dust and humidity assumptions. A headline L10 value without its test condition was not treated as proof of cabinet life. The team recorded whether the requirement meant continuous operating life, a planned replacement interval or a qualification target for one product generation. It also stated whether service staff would replace the fan by condition, by calendar interval or only after a fault, because each choice changes the evidence needed from the sample.
Change triggers were listed: new filter, grille, heat sink, cover, controller, voltage variant, connector, cable route, bracket or ambient range. Any of these can change the operating point. The same-size rule was therefore replaced by a controlled sample decision whenever the tested boundary changed.
| Release detail | Why it matters | Record |
|---|---|---|
| Airflow direction | Wrong direction defeats the heat path | Drawing and assembly mark |
| Filter state | Service condition changes pressure | Clean, loaded and trigger basis |
| Mount and gasket | Leak or preload changes useful airflow | Photos, drawing and work instruction |
| Control interface | Feedback or startup can change | Pinout, logic and response |
| Change trigger | New boundary needs new evidence | Review list before substitution |
Result: The Replacement Was Selected from Evidence, Not Appearance
The integrator could approve the candidate only after mechanical, operating-point, electrical, acoustic and service evidence agreed.
The final record included the original and replacement envelope, voltage, current, startup behavior, control signals, installed pressure, airflow measurement point, filter state, local temperatures, noise position, mounting and service method. It stated which values came from catalogue data and which came from the cabinet sample. This made the comparison useful for procurement and for the next engineering change.
The replacement did not need to copy every visible label value. It needed to deliver the required airflow at the cabinet pressure, start at the available voltage, communicate correctly with the controller, survive the defined environment and preserve the service boundary. If a customized sample was needed to meet those requirements, the frame size remained only one input to the design before any production substitution was approved.
The transferable lesson is simple: a same-size replacement is a candidate, not an approval. The operating point, control interface and installed system evidence decide whether it is equivalent for the customer's application.
How MEGA Tech Supports Fan Replacement and Customization
MEGA Tech helps engineering teams turn an obsolete or underperforming fan into a measured replacement or customized sample.
MEGA Tech develops and manufactures custom cooling fans, covering fan selection, air-path review, sample validation and volume production. We can compare DC axial fans and centrifugal blower candidates by installed airflow and static pressure, then review voltage, PWM, FG/RD, connector, mounting, noise, vibration and environmental requirements. The same outline size is not treated as proof of equivalence.
For a first review, send the current fan or label, frame and thickness, voltage, control pinout, airflow direction, cabinet photos or drawing, filter and grille, heat load, airflow and static pressure target, noise target, environment, quantity and life assumption. If the original curve is missing, send measured pressure, airflow, temperature, command and current data from the cabinet.
MEGA Tech is the custom cooling-fan manufacturer. Complete cabinets, drives, UPS systems, heat exchangers, coils, liquid loops, controllers, electrical safety and system qualification remain with the equipment developer and system integrators. Start with the fan selection tool, review the DC axial fan range, or request a replacement sample on WhatsApp.
Engineer FAQ
These questions usually decide whether a same-size replacement is ready for production review.
Q: If the fan fits the hole, is it equivalent? A: No. Check installed airflow, static pressure, control, voltage, noise, environment and thermal response.
Q: Is the catalogue CFM enough? A: No. Compare both candidates at the cabinet operating pressure and defined measurement plane.
Q: Can normal RPM prove the replacement is working? A: No. A fan can rotate normally while delivering less useful airflow through a loaded filter or restrictive heat sink.
Q: Should the controller use the old PWM duty? A: Only after the new fan's voltage, frequency, minimum duty, RPM and feedback behavior are checked.
Q: What should I send for a custom replacement sample? A: Send the current fan, size, voltage, control, airflow direction, cabinet path, airflow, pressure, heat load, environment and quantity.
Questions for a Replacement-Fan Design Review
Use this checklist with thermal, electrical, mechanical, controls, procurement and service teams.
- What airflow and static pressure does the cabinet require at the actual filter, grille and heat-sink condition?
- Where are pressure and airflow measured, and are those boundaries stated in the comparison?
- Does the replacement start at the lowest voltage available at the connector?
- What are the startup current, steady current, PWM, FG/RD and alarm conventions?
- Does the replacement preserve airflow direction, cable exit, gasket and service access?
- What happens to the operating point after filter loading or grille contamination?
- Does a tone or vibration peak follow RPM, PWM, mounting, guard or cabinet panels?
- Which sensor and heat source define thermal acceptance?
- What change triggers a new sample: fan, filter, grille, heat sink, cover, controller, cable or mount?
- What exact information should accompany the sample request: size, voltage, airflow, pressure, quantity and environment?
