Here's a question I hear several times a week: "Can I use a cheaper alternative for this part?"
The part is usually a proximity sensor, a PLC input output module, a Bently Nevada cable, or a PanelView 800 HMI. The person asking wants a yes or no answer.
I can't give them one. The right decision depends entirely on what they're doing, which is why I'm going to break this down by scenario instead of pretending there's a universal rule. There isn't one—and anyone who claims otherwise probably hasn't dealt with the consequences of a mismatched part at 1 a.m. on startup day.
There are three scenarios I see over and over: replacing parts in a running system, designing a new panel, and building something temporary. Each one has a different priority, and each one deserves a different sourcing strategy.
Scenario A: Replacing a Part in a System That's Already Running
The turbine is vibrating. The conveyor is miscounting. The HMI screen is dead. You need a replacement now, and a supplier is offering an "equivalent" at 40% less.
My advice: don't take that deal. And I don't say that because I'm an original-parts-only purist. Our shelves hold both genuine and alternative components, so this isn't about our margin. It's about what I've seen on the test bench.
A Bently Nevada cable is the clearest example I can give you. The 3300 XL monitoring system—probe, extension cable, and proximitor—is calibrated as a matched set. The cable's capacitance and shielding are part of the measurement chain. In the first quarter of 2024, a customer returned a Bently Nevada cable to us marked "defective." A multimeter showed continuity. The cable looked identical to the original. But the shield drain wire configuration was different, and the vibration signal was unusable. That small difference cost them two days of downtime and a $2,800 return.
If the system is running and the replacement has to work with what's already there, the genuine part is the cheaper option—even at twice the price.
In our Q1 2024 audit, 6% of incoming deliveries were rejected on specification grounds. In four years of reviewing 200+ line items weekly, the most common offenders have been proximity sensors and I/O modules—parts that look the same and test the same until you push them past the tolerance line.
The same logic applies to a failed HMI. If your panel runs a PanelView 800 with a View Designer application, the replacement needs to load the same runtime and connect to the same PLC tags over EtherNet/IP. A gray-market unit can look identical, but if the firmware behaves differently, you'll be the one writing the incident report.
Scenario B: Designing a New Panel or Machine From Scratch
When you're starting with a blank page, the calculation changes completely. Nothing is constrained by what's already there, which means you need to think about architecture, not just components.
This is where I see engineers make the same mistake I made early in my career: defaulting to the traditional layout. One central cabinet. A processor in the middle. A stack of PLC input output module cards in the same rack. And dozens of individual sensor cables running the length of the machine.
That layout works. I don't dispute that. But it is rarely the cheapest option when you count the full cost of installation.
Here's the counter-intuitive part: the conventional, cheaper-looking design is usually more expensive by the time you add wire, cable tray, conduit, and labor. On a machine with 40 to 60 field points, the wiring cost can exceed the component price difference between central and distributed I/O.
A set of EtherNet/IP I/O modules placed near the machine reduces all of that to a single network drop and a 24V feed. Installation drops from days to hours. When I compared two proposals for the same machine a few years back—one remote, one central—the distributed design cost more on the bill of materials but about 18% less in total installed cost. That changed my default recommendation.
One caveat: EtherNet/IP is maintained by ODVA. Most conformant third-party modules will exchange real-time data with a CompactLogix or ControlLogix scanner without custom drivers, but verify conformance before you commit to a low-cost brand—not after.
For proximity sensors in a new design, the first question should not be the brand. It should be the electrical spec. The most important detail on any input output module of PLC is the input type: sink or source.
Check whether your PLC input output module is sinking or sourcing. A 3-wire PNP sensor switches the positive side; a 3-wire NPN sensor switches the negative. Get that backwards and every channel reads wrong. Also check the leakage current of 2-wire sensors against the input card's off-state threshold, or you'll spend a week chasing phantom counts that only appear when the line voltage dips at night. IEC 61131-2 establishes the input classes that most industrial PLCs follow—that standard is your reference for what "high" and "low" signals are allowed to be.
I ran a blind comparison not long ago between a budget proximity sensor and a name-brand one. Both were M18, 24VDC, PNP, with the same sensing distance on paper. At 5 mm, they performed identically. At 6.2 mm—inside the advertised range—the budget unit started dropping targets intermittently. Individually, I wouldn't have caught it. Side by side, the inconsistency was obvious. In my opinion, that's the real reason to buy a known brand for field devices: not because the cheap one never works, but because the cheap one doesn't work the same way every time.
Scenario C: Prototyping, Training, or Temporary Rigs
This is the scenario nobody writes articles about, but it's extremely common. You're building a training rig, a college lab, or a proof-of-concept prototype. The machine isn't making money, and a failure costs a lesson, not a deadline.
For that situation, buying the cheaper alternative is a reasonable choice. I get why teams do it—budgets are real, and "will this $40 proximity sensor work for our test stand?" is a fair question.
To be fair, several alternative manufacturers have improved significantly over the last decade. The old rule that third-party parts are automatically unreliable comes from an era when they often were. That has changed for certain categories.
But even in Scenario C, I keep three disciplines:
- Document everything. If your prototype becomes a production machine—it usually does—you'll need exact part numbers and specs to make the right final purchases.
- Don't mix brands on a single input module. Mismatched switching characteristics cause intermittent faults that are almost impossible to trace.
- Run a 48-hour burn-in test. Keep the system running continuously for two full days before you trust it for anything serious.
Alternative components earn their place here. Some reveal problems at 1 a.m. on day three. Others work flawlessly for a decade. That's why the test matters.
How to Know Which Scenario You're In
Still uncertain? Here's the three-question test I use on sales calls:
- Is this system currently in production? If yes, you're in Scenario A. Buy the genuine part.
- Is the specification already fixed by what's in the panel? If yes, you're in Scenario A, regardless of how good the alternative deal looks.
- What does one hour of downtime cost? If you can name a number with more than two digits, you just answered the question yourself.
One $300 cable choice cost a plant $22,000 in lost output and overnight freight. Nobody planned for that. It happened because the part was ordered on price, not on specification.
If you answered no to all three questions, then look at the project itself. Is it permanent? Use Scenario B thinking. Is it temporary or experimental? Scenario C is fine.
One Final Thought: Work With Someone Who Knows Their Limits
The most important lesson I've learned in this job is that the supplier who tells you what they don't do well is the supplier you want to keep.
We sell both genuine and alternative automation parts. If you call us about a Bently Nevada cable for a critical turbine protection system, we will tell you to buy the genuine part. We aren't going to pretend an alternative is equivalent when the price of being wrong is a machine asset worth more than most house deposits.
That might sound like a supplier pushing work away. In reality, it's exactly the opposite. After four years of reviewing parts shipments—checking specifications, rejecting mismatches, explaining to customers why the cheap option is the expensive one—I've learned that trust is the only license to do business.
A vendor who says "this isn't our strength—here's who does it better" earns your confidence for everything else. A vendor who claims they can do everything is usually telling you that their quality record doesn't.
So buy according to your scenario. Match the risk to the part. And if a supplier dodges the question about what the part is actually for, that's your red flag.
Leave a Reply