When the original fan stops being buyable
A replacement fan project rarely starts as a design choice. It starts when the original fan is discontinued, placed on allocation, repriced, or quoted with a lead time the production plan cannot absorb.
The trigger is usually commercial, but the risk is thermal. The fan in a running product is not a generic part: it was selected, knowingly or by trial, against one enclosure, one filter state, one control signal and one noise target. Replacing it with the first fan that shares the frame size transfers all of that hidden engineering into an unplanned experiment on your own product.
This guide gives a working method for the replacement decision. It is written for teams that buy fans in small and medium volumes, including startups and small manufacturers that cannot keep a thermal lab busy full time. The method has three steps: capture the duty point of the installed fan, shortlist candidates from the fan curve rather than the label, and validate samples on your own equipment before any volume commitment.
MEGA Tech supports this exact workflow. We manufacture custom DC axial fans and backward curved centrifugal fans, we select by operating condition rather than by part number, and we validate samples on the customer unit. What we need from you is listed in the MEGA Tech section near the end, together with the boundary of what we do not supply.
| Trigger | What it breaks | First move |
|---|---|---|
| End of life notice on the original fan | The qualified configuration loses its supply | Freeze the installed duty point data before the last units ship |
| Lead time grows past the production plan | Line stops or expensive buffer stock | Start sample validation in parallel with the buffer order |
| Minimum order quantity rises | Small teams cannot buy the original part at all | Ask a fan manufacturer for a sample based selection route |
| Price step on a sole source part | Unit cost and margin of the whole product | Compare total cost, not fan price alone |
| Field failures traced to the fan | Warranty cost and reputation | Treat the replacement as a reliability change, not a purchase |
The outline size is not the operating point
Two fans can share the same frame size, the same nominal voltage and a similar free air rating, and still deliver very different airflow once they are bolted into a real enclosure. The enclosure, the guard, the filter and the board layout form a system resistance curve. The fan delivers the airflow where its own pressure and airflow curve crosses that system curve. That crossing is the operating point, and it is the only honest basis for a replacement decision. The underlying method is the same one set out in the fan selection guide and the guide to airflow and static pressure curves, applied here to a replacement decision.
Free air numbers are measured with no resistance and describe a condition your product never sees. A candidate fan with a higher free air rating but a weaker pressure curve can deliver less installed airflow than the original. The reverse also happens: a fan with a modest free air rating and a stronger pressure curve holds the duty point better behind a loaded filter.
This is why MEGA Tech does not endorse the claim that the same outline dimensions make two fans equivalent. Outline, mounting holes and connector are necessary checks. They are not sufficient. Equivalence is a statement about the installed duty point, the control behaviour and the validated sample, in that order.
| Data point | On the label | In the installed product |
|---|---|---|
| Airflow | Free air maximum | Set by the crossing of the fan curve and the system curve |
| Static pressure | Maximum at zero airflow | The value at the duty point, often 20 to 60 percent of maximum |
| Noise | Free air, ideal inlet | Rises with inlet obstruction, guard and filter state |
| Current | At free air or at rated point | Changes with the duty point and the supply tolerance |
| Life | At a stated temperature and orientation | Depends on bearing type, orientation and real air temperature |
Step 1: Capture the duty point of the installed fan
Before you search for candidates, record what the original fan actually does in the product. If the original fan is still available, run this capture on a production unit. If it is already gone, run it on the best remaining unit and mark the data as reference only.
You do not need a wind tunnel for a useful capture. You need the temperatures that matter, the electrical conditions at the fan connector, and the mechanical facts a replacement must match. Record the data with the filter in its normal state and repeat the temperature record with a loaded filter if the product has a service interval, because the replacement must hold the duty point at the end of the interval, not only on day one. The installed airflow measurement guide gives repeatable ways to record that state on a real enclosure.
- Component temperatures: the hottest part the fan protects, at worst case load and highest rated ambient temperature.
- Air temperature at the fan inlet and outlet, so the capture separates fan performance from room conditions.
- Voltage at the fan connector during start, during normal run and during the lowest supply state the product allows.
- Control behaviour: DC voltage control range, or the PWM duty range, the PWM frequency and the FG output or RD output the controller expects.
- Mechanical facts: frame size, mounting hole pattern, airflow direction, rotation direction, lead wire length, connector type and cable exit side.
- Mounting orientation: horizontal, vertical or shaft vertical, because bearing life depends on it.
- Noise record: a repeatable phone or meter reading at a fixed distance is enough to compare candidates later.
Step 2: Shortlist candidates from the fan curve
With the duty point recorded, the shortlist is a curve comparison, not a catalogue search. Ask each candidate supplier for the pressure and airflow curve of the exact model, and read the curve at your duty point. A candidate that cannot show a curve is not a candidate for a validated replacement.
Read four values at the duty point: airflow, static pressure, current and speed in RPM. Then check the margins. The candidate should hold the required airflow with the filter loaded, start at the lowest voltage your product can present, and stay inside the current budget of the connector and the driver. Check the bearing type against the mounting orientation and the air temperature at the fan, and check the rated life at that temperature rather than at room temperature.
Control details decide many replacements. A fan with PWM control expects a defined PWM frequency range and a defined behaviour at zero duty: full speed, stopped, or minimum speed are all in the market, and they are not interchangeable. The FG output gives a number of pulses per revolution that the controller may use for a locked rotor alarm; a candidate with a different pulse count can make a healthy fan look faulty to your firmware.
What changes inside the enclosure when the curve moves
The schematic shows the same enclosure with two candidate fans. In the left panel the candidate curve crosses the system curve at the required airflow, the particle stream reaches the load, and the exhaust air carries the heat out. In the right panel the weaker pressure curve crosses earlier: less air enters, part of the stream recirculates inside the enclosure, and the load sits in a hot zone even though the fan is spinning and the frame size is identical.
This is the failure pattern behind most bad replacements. Nothing looks wrong at switch on. The product passes a short bench test, then drifts into thermal throttling, nuisance alarms or early bearing wear in the field, weeks after the purchasing decision that caused it.
Step 3: Validate samples on your own equipment
Samples are where a replacement becomes engineering instead of hope. Order a small sample set of the leading candidate, fit it to a production representative unit, and run the checks in the table. The sample quantity a small team needs is modest: enough units for one thermal soak, one spare for a repeat, and one for the file.
Run the thermal soak at the worst case the product allows: highest load, highest rated ambient you can arrange, filter in the loaded state if that is the service condition. Record the same temperatures you captured from the original fan, in the same positions, so the comparison is direct. A replacement that holds component temperatures within the original record, starts cleanly at low voltage and behaves correctly on the control signal has earned the next step: a pilot batch with change control.
| Check | Method | Pass condition |
|---|---|---|
| Fit and direction | Install the sample with the production guard and filter | Mounts without force, airflow direction correct, cable reaches the connector |
| Start behaviour | Apply the lowest voltage the product can present, ten starts | Starts ten of ten, no stall, no reverse run |
| Duty point | Compare inlet and outlet air temperatures against the original record | Air temperature rise within the agreed band of the original fan |
| Component temperature | Thermal soak of two hours at worst case load | Hottest component at or below the original record, with margin to its limit |
| Control and alarm | Sweep the PWM duty or control voltage, block the rotor briefly | Speed follows the control, FG output or RD output matches the firmware expectation, alarm triggers |
| Noise | Repeat the original noise record at the same distance and state | No new tonal peak, overall level within the agreed band |
| Restart after a power dip | Interrupt the supply for one second during the soak | Fan restarts and the product recovers without a manual reset |
The small differences that cause field returns
Most replacement projects that fail in the field do not fail on airflow. They fail on a small difference nobody wrote down. The list below is the review we run with customers before a sample is approved, and it is the list to attach to your own change record.
- FG output pulses per revolution: one pulse and two pulses per revolution both exist; the firmware alarm threshold must match the fan.
- Behaviour at zero PWM duty: full speed, minimum speed and stopped are different products from the controller point of view.
- Lead wire length and connector: a shorter lead or a mirrored connector can force a harness change that costs more than the fan.
- Bearing type and orientation: a sleeve bearing fan mounted outside its orientation range can lose most of its rated life.
- Operating temperature range: the air at the fan inside a loaded enclosure can exceed the room rating by 20 K or more.
- Locked rotor protection and automatic restart: the candidate must protect itself and restart in the way the product expects.
- Ingress protection and coatings: a dusty or humid application needs the protection decision repeated, not inherited by assumption.
The real cost of a replacement fan
A replacement decision made on unit price alone usually moves cost rather than removing it. The fan is a small part of the product price, but it guards the parts that are not small: the power stage, the battery, the optics, the compute board. One field return with shipping, labour and a replacement unit can erase the saving from a large quantity of slightly cheaper fans.
The cost effective route is boring on purpose. Capture the duty point once, validate samples once, document the change, and then buy a stable configuration from a manufacturer that keeps the configuration stable. For a small team the same route has a second benefit: the validation package you build for the replacement becomes the specification for the next product, so the work is spent once and reused.
Judge a candidate supplier on the evidence they can put on the table: a real fan curve, sample units for your own test, a stated life basis, and change notification for the parts that matter. Those four items protect your product more than any single price line.
How MEGA Tech can support the review
MEGA Tech manufactures custom cooling fans and has done so since 2008, with an office in Guangzhou and a factory in Shenzhen, under ISO 9001, with CE and RoHS on the product range. Our published catalogue is DC axial fans, and we also build backward curved centrifugal fans for enclosure work. For replacement projects we work in the order of this guide: operating condition first, curve second, sample third.
Small teams and startups are a normal customer shape for us. A replacement project can start with a small sample set for your own validation, with our engineers reviewing the duty point data with you before the samples ship. As your volume grows, the validated configuration is held under change control, so the fan you qualified is the fan you keep receiving. That is what growing together means in practice: the engineering file built at sample stage carries the production stage.
To start a review, send us the installed fan data you captured in step 1: frame size and mounting, voltage and control signal, FG output or RD output expectation, connector and lead length, the component temperature record, airflow or static pressure target if you have it, quantity per order and the environment the fan works in. We answer with a curve based shortlist and a sample proposal. We do not supply coils, liquid loops, controllers or cabinet safety design, and we will say so early if your problem sits in that territory.
Engineer FAQ
Can a replacement fan with the same frame size and a higher free air rating still cool less? Yes. Installed airflow is set by the crossing of the fan curve and the system curve. A weaker pressure curve can cross at a lower airflow even when the free air number is higher.
How many samples does a small team need for validation? A practical set is three to five units: one for the thermal soak, one for control and alarm checks, one spare for a repeat, and one kept as the approved reference sample.
What if the original fan curve is not available? Work from the installed record instead: component temperatures, inlet and outlet air temperatures, voltage at the connector and the control behaviour. A candidate can be validated against that record on your own equipment.
Should the replacement match the original bearing type? Match the requirement, not the label. The bearing must suit the mounting orientation and the air temperature at the fan, with a rated life at that temperature that covers the service plan.
When is a replacement a design change rather than a purchase? Whenever the curve, the control behaviour, the FG output or the protection behaviour differs from the original. That is most replacements, which is why the validation record belongs in the product file.
Questions for a Design Review
Bring these answers to the review and the replacement decision becomes a record instead of an opinion.
- What is the installed duty point, and which measurement proves it?
- Which candidate holds that duty point with the filter in the loaded state?
- Does the candidate start at the lowest voltage the product can present?
- Do the PWM behaviour, the FG output and the alarm thresholds match the firmware?
- Does the bearing suit the mounting orientation and the air temperature at the fan?
- Which sample checks were run, on which unit, and where is the record filed?
- What change notification will the supplier give for the parts that matter?
