Two Industrial Parts Look Identical: The 12 Axes That Decide

Two Industrial Parts Look Identical: The 12 Axes That Decide Whether They Interchange

Two industrial automation parts can look identical from the outside — same outline, same label family, sometimes even the same brand and series — and still fail to interchange on firmware, terminal assignment, electrical rating, communication protocol or certification scope. A documented substitution has to clear twelve named axes, not one or two, and on at least one of those axes the honest answer is often unknown until someone reads the original datasheet and the candidate's datasheet side by side. The mistake is treating a cross-reference as a confirmation. The right output is a table that names what was checked and what was not.

Why "they look the same" is the most expensive assumption in maintenance

A buyer searching «чем заменить <модуль>» or what replaces this module lands on a forum answer, a Chinese listing with a similar-looking photo, or a colleague's recollection that "the S7-300 and the S7-1200 look about the same." None of those is a substitute for a twelve-axis comparison, and the cost of getting it wrong is paid in three currencies at once: downtime while the wrong unit is bench-tested, rework on the cabinet, and a control program that no longer behaves the way the original did.

We see this most often with three categories. PLC modules from the same family but different function status suffixes — for example a Siemens 6ES7315-2AH14-0AB0 CPU module replaced by a candidate with a different FS suffix. Photoelectric sensors from different brands with the same housing thread and sensing range, such as a SICK WTB4 swapped for an Omron E3ZM-T81 2M. Contactors from different vendors carrying the same current rating on the data sheet. In each case the outline matches, the brand family looks right, and the substitution still fails the moment the cabinet is powered on — because a terminal is in a different place, because the firmware refuses the program, or because the EMC class is below what the surrounding equipment assumes.

The cross-reference we publish for any of those lines reports twelve axes with a verdict of matched, differs or unknown. It does not issue a compatibility certificate and it does not declare a part equivalent. We are an independent industrial automation distributor and sourcing desk — that role lets us source across the supply chain, including parts the franchised channels list as obsolete or allocation-only, and it also means we do not certify on the manufacturer's behalf. For buyers in Moscow, the regions of the Russian Federation, and the wider CIS, the same cross-reference is built on the same twelve axes; delivery is arranged under EXW, FCA, DAP or DDP as stated on the quotation, and consolidated shipments are issued with one commercial invoice and one packing list.

The twelve axes, and what each one actually checks

The list below is the order we work through when a buyer asks for a substitution. Some axes can be checked from a datasheet and a photograph. Some need the original part in hand. Some need the machine's own documentation. Each axis is reported as matched, differs or unknown; an axis left blank is an axis the comparison silently assumed, and that is the failure mode the framework exists to prevent.

1. Form, fit and outline (mounting method, module width, depth)

DIN-rail versus panel-mount versus screw-mount changes the cabinet layout immediately. Module width matters when the part sits next to other terminals or beside another module on the same rail; a 35 mm versus 45 mm body width can move the wiring plan and change how the cable duct sits above the rail. Depth behind the panel includes the room needed for the front connector and any captive strain-relief, and on retrofit work it is the dimension that decides whether the existing cabinet door still closes.

This is the axis a visual comparison can sometimes settle — a photograph with a ruler, an outline drawing from the datasheet, a CAD footprint. It is also the axis where "looks the same" most often gives a false positive, because two modules from the same series can share an outline and disagree on everything else. Verdict: usually matched on outline, sometimes matched on outline only.

2. Dimensions (panel cutout, mounting hole pitch, depth behind the panel)

Distinct from outline because two parts with the same outline can have different panel cutouts, different mounting hole positions or different rear clearances. A substitution that "fits the DIN rail" can still need a new hole pattern on a panel-mount device. Dimensions also drive the mechanical drawing the cabinet builder works from, and on a control panel that is being refurbished rather than rebuilt, the panel metalwork is fixed.

3. Terminal and interface assignment

This axis catches more bad substitutions than any other. Two parts can carry identical-looking terminal blocks and disagree on the order of L+ and M, on the pin assignment of a multi-pin connector, or on whether pin 8 is an input or an output. The same applies to the front connector on a PLC module: the connector may be the same physical part, but the pin-to-channel map can be rotated or rearranged between revisions.

For a SICK WTB4 versus an Omron E3ZM-T61 2M, for example, the M12 connector and the four-pin assignment are commonly the same — but a PNP versus NPN output flips what each wire does, and that is not visible from the outside. Verdict: matched on paper is not the same as matched when wired.

4. Electrical parameters

Supply voltage and tolerance, channel current, output type (sourcing versus sinking), surge and short-circuit protection, power consumption. A 24 VDC module that accepts 19.2 to 28.8 V is not the same as one that accepts 20.4 to 28.8 V. A relay output rated 2 A is not the same as a transistor output rated 0.5 A even when both are called "digital output." Lower voltage tolerance can pass bench test and fail in the field when the supply dips at motor start.

5. Function and I/O specification

Channel count, signal type (digital, analogue, temperature, counter, high-speed, PTO, PID), resolution on analogue channels, isolation (channel-to-channel, group-to-group), diagnostic features. A 16-channel digital input is not the same as a 16-channel module where two channels are configured as counters — even if both are described as "16 DI" on a one-line product description. Where the original design uses a feature the candidate does not have, the substitution is not equivalent; the right response is to look for a module that retains the feature, not to assume the feature was unused.

6. Firmware, hardware version and engineering-software compatibility

The axis that most often ends up at unknown. For a Siemens S7-300 module, the firmware version and the hardware revision together decide which engineering software can configure it, which function blocks will compile, and whether the existing program loads at all. Function status (FS) suffixes encode behaviour that earlier datasheets do not document; FS04 versus FS05 versus FS06 is not a marketing label, it is a behavioural change.

For most non-current-production parts the firmware relationship between original and candidate is not published in a form a buyer can verify without the original engineering software in hand. We report this axis as unknown when the datasheet does not state it, and the verification step is then passed back to your engineering team. Verdict: this is the axis where 'looks the same' fails most often.

7. Communication and protocol

PROFINET versus PROFIBUS, EtherNet/IP versus Modbus TCP, RS-485 versus RS-232, the supported baud rate, the maximum number of nodes on the segment, and the network configuration file (GSD, GSDML, EDS, ESI). A module that "speaks Ethernet" does not speak PROFINET, and a PROFINET device that supports only conformance class A does not interoperate with a controller expecting class B. Substituting a remote-I/O head on a live network without confirming the device description file is one of the more expensive ways to find out that unknown should have been the published verdict.

8. Mechanical parameters

Terminal torque range, allowable wire cross-section, mechanical endurance (operations, mating cycles), locking and latching. A relay with a 100,000-cycle mechanical life is not equivalent to a relay with a 10,000,000-cycle mechanical life if the application is a high-cycle interlock. We report this from the datasheet where published, and flag it where the figure is not stated by the manufacturer.

9. Environmental class and EMC

Operating temperature for the cabinet interior versus the cabinet exterior, humidity, vibration and shock resistance, EMC immunity and emission class, pollution degree, installation altitude. An office-grade sensor does not survive a factory floor with high-voltage drives nearby, and the difference is in the EMC class, not the price tag. Where the original was specified to industrial environment and the candidate is specified only to commercial, the substitution cannot be made without changing the environmental claim.

10. Ingress protection and temperature range

IP20 versus IP65 versus IP67 — and whether the IP rating applies to the device as installed behind the panel, to the device body, or to the front face only. A sensor that is IP67 on its own is not IP67 once it has been mounted in a panel that does not seal around it. Temperature range too: an industrial-rated –25 to +70 °C operating window is not the same as 0 to +50 °C. Cabinet interior sensors and cabinet exterior sensors sit in different thermal worlds.

11. Materials and construction

Contact plating (silver-tin oxide versus gold-flash versus silver-nickel), housing material, sealing material (NBR versus FKM versus silicone), PCB conformal coating, RoHS and REACH status. A contactor with silver-nickel contacts is rated for a different switching regime than the same form factor with silver-tin oxide, and low-level signal relays often require gold flash on the contacts to remain reliable at small switching currents. We do not infer material equivalence from a photograph; this axis is reported from the datasheet or marked not stated by manufacturer.

12. Certifications and lifecycle status

CE, UL, CSA, RCM, EAC, functional safety (SIL/PL), ATEX — the actual certificate numbers where the manufacturer publishes them, and the lifecycle status of the part (current, phased out, end-of-life). A phased-out part with a valid certificate is still certified for its declared ratings; a current part with a lapsed certificate is not. We do not hold certifications ourselves and we do not issue them; this axis is reported as the manufacturer states it, with the file numbers where the manufacturer publishes them.

The hardest axis to verify, and why it ends up at "unknown"

For a substitution involving any non-current part — a discontinued PLC module, a sensor that has been replaced by a newer housing, an I/O card whose function status has changed — the firmware/engineering-software axis is the one that most often returns unknown. The original is no longer in production, the candidate's firmware is current, and the migration guide either does not exist or does not cover the specific module-to-module pair.

The honest answer is to write unknown in that cell of the comparison table, and to refuse any drop-in framing. The recovery path is on the engineering side: open the original program in the candidate's engineering software, check that the project compiles, and validate the I/O configuration against the wiring diagram. That work is not something we can perform from a desk in Shenzhen — it is a step your engineering team owns, and the comparison we publish is the input to that step, not a substitute for it.

When substitution is the wrong answer

Three situations where the right response to a discontinued or hard-to-find part is keep sourcing the original rather than substitute:

The part participates in operator-protection circuitry. Stop circuits, guard interlocks, two-hand controls, light curtains — substitution here changes the operator-protection case. The independent desk finds the original; the assessment is owned by your machine builder or notified body.
The part sits inside a certified assembly whose certification scope names the part. A substitution, even one that checks every axis, changes what the certificate covers. Keep the original, or treat the change as a re-certification project.
The machine's own documentation pins the design to a specific part number. A spare-parts list, a bill of materials with revision letters, a validated BOM — when the machine builder has documented the part, the substitution is a design change, not a maintenance decision.

We treat those three as a hard line. If a comparison ends at "but the operator-protection case depends on this exact part," the answer is to source the exact part, not to find a "close enough" substitute.

What a written cross-reference actually contains

A comparison we publish for a substitution question contains: the two identifiers (original and candidate, by part number or series), the twelve axes with verdicts in matched, differs or unknown, the unverified items listed explicitly, at least one situation where the substitution should not be made, and the verification line — verify against the original manufacturer datasheet and your own qualification process. It does not contain a compatibility certificate, because an independent desk cannot certify on the manufacturer's behalf.

What it does contain, when the line is sourced through us, is the per-line condition (new, new surplus, refurbished or used), the markings on the actual unit, and the pre-dispatch photographs. Those photographs are the operational equivalent of "I checked the part is the part"; they do not certify interchangeability, they document identity.

For a comparison to be worth the paper it is printed on, every cell has to be filled with one of the three verdicts. An empty cell is a silent assumption, and silent assumptions are what make a substitution fail in the cabinet rather than at the desk.

How to start the work — send us your BOM

For a comparison on a specific line, send your BOM to our sourcing desk with the part numbers, the markings on the housing, the machine the part is fitted to, and any acceptable-substitute note your engineering team has already agreed. A typical request asks for a cross-reference on three to ten lines drawn from a working bill of materials; we return the comparison line by line, with the unverified axes named and a verification instruction at the end. Lead time and price for the cross-reference report itself are stated on the quote, with the lines underneath quoted per line as usual.

If only a photograph is available, send it with whatever marking is legible — identification is part of the work, not a prerequisite for it. From there, the comparison is built line by line on the twelve axes, and the RFQ is built on top of the result. We screen end users and end uses, classify before quoting, and decline transactions that cannot be screened.

FAQ

Is a cross-reference the same as a guaranteed replacement?

No. A cross-reference is a documented comparison of twelve named axes with their verdicts. It is not a warranty of interchangeability, and it is not a certification. Two parts can match on every axis we can verify from datasheets and still fail on firmware, because that axis often returns unknown until your engineering team opens the program in the candidate's software.

What are the axes you actually compare?

Twelve: outline and mounting, dimensions, terminal and interface assignment, electrical parameters, function and I/O specification, firmware and hardware version, communication protocol, mechanical parameters, environmental class and EMC, ingress protection and temperature, materials and construction, and certifications and lifecycle status. Each is reported as matched, differs or unknown. Omitting an axis means the comparison assumed it, and silent assumptions are the failure mode the framework exists to prevent.

Why can you not just tell me whether two parts are compatible?

Because the honest answer usually contains the word unknown on at least one axis. A supplier can confirm outline, mounting and terminal layout from a datasheet. It cannot confirm that your control program will run on a different controller family, that your certification scope still holds, or whether the substitution survives the specific environment your machine runs in. That judgement is yours, and it is the reason the comparison ends with a verification instruction rather than a conclusion.

Can a different PLC family run my existing program?

That is a question for your engineering team, not for a supplier. Program portability depends on the programming environment, the instruction set you actually use, the I/O configuration and any vendor-specific function blocks. We can describe the hardware and the migration documentation that exists; we cannot tell you the program will run unmodified — and any supplier who says otherwise is guessing.

Do same-brand, same-series parts interchange?

Not necessarily, and this is one of the most expensive assumptions in industrial maintenance. Within one family, suffixes encode function status, I/O configuration, supply voltage and firmware generation. Two modules from the same series can share an outline and disagree on everything else that matters to the application. The comparison has to be done line by line, not series by series.

What does the comparison not include?

A compatibility certificate. A functional-safety conclusion. A statement that the candidate is suitable for a medical application. A guarantee of any specific lead time, transit time, or delivery term. Those limits are deliberate: we are an independent distributor and sourcing desk, and an independent desk is not in a position to certify on a manufacturer's behalf, to make a functional-safety judgement, or to make regulatory calls about a medical device.

What should I send you to start a substitution comparison?

The original part number or series, the markings visible on the housing and the label, the machine the part is fitted to, and any acceptable-substitute note your engineering team has already agreed. If you have only a photograph, send it with whatever marking is legible — identification is part of the work, not a prerequisite for it. From there, the comparison is built line by line on the twelve axes, with the unverified ones named, and the verification step left where it belongs.

The takeaway

Send us your BOM with the part numbers and the markings. The substitution comparison is built on the twelve axes, with what was checked and what was not, and a verification step your engineering team owns. The framework exists so that "looks the same" never reaches the cabinet without at least one axis marked unknown and a clear path to close it.

Data Notes

Data through: September 2026. The framework above is the same one our sourcing desk uses internally when a buyer asks «как проверить, что аналог действительно подходит» or how do I verify a substitute actually fits. The MPNs named in the body (6ES7315-2AH14-0AB0, E3ZM-T81 2M, E3ZM-T61 2M) are real catalog identifiers from our Siemens and Omron lines, included so the article is anchored to inventory we actually quote, not abstract examples. Where a figure is not stated by the manufacturer, the axis is reported as not stated by manufacturer rather than inferred. The framework is published under the brand disclaimer that we are an independent industrial automation distributor and sourcing desk — not an authorized distributor for any of the manufacturers named.

By the aoctrl sourcing desk, independent industrial automation distributor.