Cross-Reference Charts and Their Errors

Why this matters

A cross-reference chart turns "I need the part for this unit" into a part number in thirty seconds, and most of the time it is right. The trouble is that it is right often enough that techs stop checking, and the errors it does make are the quiet kind: a motor one size too small, a rotation stated from the other end, a valve for the wrong gas. None of these stop the part from fitting. They show up as a callback, a short-lived replacement, or a heat exchanger running hotter than it should. Knowing how these charts are built tells you where they are likely to be wrong, which is where the thirty seconds of checking should go.

How cross-reference tools get built

Every chart or lookup app matches one part to another on a subset of attributes, chosen by whoever built it. The subset is where the errors come from.

Manufacturer interchange lists. The appliance or component maker publishes which of its part numbers replace which. These are the most reliable source for that maker's own parts, because the maker knows its revisions. They are weakest at the edges: discontinued lines, other makers' parts, and revisions added after the list was last updated.

Distributor and aftermarket databases. A universal-part maker or a distributor maps OEM part numbers to its own catalog. Someone keyed the OEM part's attributes, often from a spec sheet or a sample, and matched them to the nearest catalog part. Accuracy depends on how many attributes were keyed and how carefully.

Model-to-part lookups. Enter the equipment model, get the part. These usually key on the model family, not the serial number, which is where revision changes get lost: a model built for ten years may have used two or three different motors or controls.

Aggregated "fits" data. Some sources combine lists from several of the above, sometimes scraped, with no indication of which entries came from where. Treat these as a starting suggestion only.

Supersession, where a part number has been replaced by a new one, is its own problem with its own card (See related: Supersessions and the Part Number That Changed).

The errors that recur

Voltage. A chart matches a motor or a coil by function and frame and lists one voltage when the original came in two versions. 115 V and 230 V versions of the same motor, or 24 V and 120 V coils on a contactor or relay, look identical in a photo.

Rotation stated without a viewpoint. Motor rotation can be given as viewed from the shaft end or from the lead end, and the same spin reads opposite from the two ends. A chart that says "CW" without saying which end is half a specification.

  SAME MOTOR, SAME SPIN, TWO VIEWPOINTS

   from the LEAD end          from the SHAFT end
       ---->                      <----
     /       \                  /       \
    |  leads  |                |  shaft  |
     \       /                  \   o   /
       <----                      ---->
   reads CLOCKWISE            reads COUNTERCLOCKWISE

A rating rounded to the family. A chart keyed on model family lists one horsepower, one microfarad value, one limit temperature, or one gas type for every unit in the family, when a revision changed it partway through production.

Mechanical attributes dropped. Shaft length, flange pattern, thread type, mounting stud spacing, connector style. The part is electrically right and does not fit, or fits only with an adapter that changes what it was designed to do (a fan blade set at a different height, a valve on a reducer that restricts flow).

Safety ratings treated as interchangeable. A chart may list a "universal" replacement for a limit, a pressure switch or a relief valve because it fits the mount. These are the parts where only the specified rating will do (See related: Universal Replacement Parts and When They Are Right).

Transcription. A number entered wrong once, copied everywhere the database is used. These are rare and impossible to predict, which is why the verification does not depend on guessing which entries are suspect.

Worked example: a furnace blower motor from a model lookup

A gas furnace's direct-drive blower motor has failed with a seized bearing. The tech enters the furnace model number into a distributor's lookup app. It returns a universal motor: 1/2 hp, 115 V, 1075 rpm, 4 speeds, rotation "CCW."

The safety step first. The furnace's service switch and its breaker are both turned off, the breaker locked and tagged under OSHA's 29 CFR 1910.333(b)(2); the blower door switch is not relied on as isolation. The tech proves the blower circuit dead at the motor leads with a meter checked on a live source before and after (the live-dead-live test of NFPA 70E Article 120 (120.5 in the 2021 edition, renumbered in 2024)), and discharges the motor's run capacitor through a resistor, confirming zero volts.

Verification, line by line, against the failed motor itself. The tech reads the nameplate on the failed motor, not the furnace's model label.

• Voltage: 115 V. Matches.
• Speed: 1075 rpm, 4 speeds. Matches.
• Horsepower: 3/4 hp. The chart said 1/2.
• Rotation: "CCW LE," counterclockwise viewed from the lead end.
• Capacitor: 10 uF, 370 V.

The horsepower. The furnace's rating plate shows a serial number that falls in a later production range. The appliance maker's own parts list, looked up by serial number on the maker's site, shows the furnace family switched from a 1/2 hp to a 3/4 hp blower motor partway through production, for a larger-capacity cabinet. The distributor's lookup keyed on the model family and returned the earlier revision. A 1/2 hp motor in this furnace would move less air than the furnace was certified with. The heat exchanger would run hotter, the temperature rise would climb toward or past the range on the rating plate, and the high limit would start cycling the burner, which the homeowner would describe as "it blows cold air sometimes."

The rotation. The universal motor's box says "Rotation: CW, shaft end; reversible." The chart's "CCW" was the original's lead-end value carried over without its viewpoint. Counterclockwise from the lead end is clockwise from the shaft end, as the diagram shows, so the universal motor's default rotation is already correct. A tech who saw "CCW" on the chart and "CW" on the box, and reversed the motor to make them agree, would have installed a blower running backward: it moves some air, much less than it should, and the symptoms look like a dirty filter.

The fix. The tech orders a 3/4 hp universal motor rated 115 V, 1075 rpm, 4 speeds, CW from the shaft end by default, and checks its label for its own capacitor requirement: 10 uF at 370 V or higher. The existing capacitor reads within its printed tolerance on the meter and is reused.

Resuming, with the check that makes it safe. Blower door on, lock removed, breaker and service switch on. A personal CO monitor is clipped on before the furnace fires. The tech confirms airflow direction at a supply register, measures the motor's current on its lead with a clamp meter in the PPE the shop requires for energized testing and finds it under the new motor's full-load amps, then measures supply and return air temperatures after 10 minutes of steady firing: a 48 F rise against a rating-plate range of 35 to 65 F. The rise is the number that proves the airflow is right, because it is the one the heat exchanger actually experiences.

What the chart cost and what the check bought. The chart's two errors, one wrong and one ambiguous, were both caught by reading the failed motor's nameplate and the appliance maker's serial-specific parts list. That took about ten minutes. Skipping it would have produced a furnace that ran, tripped its limit intermittently for months, and drew a second visit and probably a third.

The verification that catches these

Start from the failed part's nameplate, not the equipment model. The part in your hand is the only source that is certain to describe this unit's revision. If its nameplate is unreadable, the appliance maker's parts list by serial number is next.

Pull the replacement's own label or spec sheet, not the chart's summary. The chart is a pointer. The replacement's label is the claim the part's maker stands behind.

Compare every attribute, including the ones the chart left out. Voltage, frequency, phase, horsepower or rating, speed, rotation with its viewpoint, capacitor requirement, mechanical fit, enclosure. A chart that does not list an attribute has not checked it.

Check the chart's notes and serial ranges. Many lists carry footnotes ("serial numbers after...", "except LP models") that do not appear in an app's one-line result.

When two sources disagree, the appliance maker's serial-specific parts list governs over a distributor or aftermarket chart, and the part's own nameplate governs over both.

Write down what you crossed and how you checked it. On the job record: the chart's suggestion, the attribute that differed, and the part actually installed. If a chart is wrong once, the next tech who meets that furnace family benefits, and the distributor can correct its database if you tell them.

When the answer changes

The original is a safety control. Do not cross by chart at all. Use the appliance maker's specified part.

No nameplate and no serial-specific parts list exists. On very old equipment, measure what you can (shaft, mount, current draw before failure if it was recorded), choose conservatively, and treat the first run as a test with the resume checks above.

The chart is the appliance maker's own interchange list for its own part. It is the most reliable source available, but serial-range notes still apply.

References

• Appliance manufacturers' serial-specific parts lists and component manufacturers' labels and spec sheets
• NFPA 70E Article 120 (120.5 in the 2021 edition, renumbered in 2024), and 29 CFR 1910.333(b)(2) for isolating equipment before work
• Furnace rating plates and installation instructions for temperature-rise range
• See related: Universal Replacement Parts and When They Are Right; Supersessions and the Part Number That Changed; The Wrong-Spec Substitute Part That Works Until It Doesn't; Handling Discontinued Parts and Finding Substitutes