One afternoon, an engineer put a purchase request on my desk with the subject line: Tester. That was all. No voltage range, no channel count, no sensor types, no pass/fail criteria. The request mentioned two possible sources: National Instruments and Broadcom. In the notes there was also a reference to an old National Instruments FP 1000 module found in storage.
I sat there and thought: this is exactly how cost overruns get born.
I am not a salesperson, and this is not a pitch. I manage test equipment procurement for a 40-person engineering services company. Over the last six years, I have tracked about 60 instrument purchase orders and analyzed roughly $500,000 in cumulative spend. When I see a vague request, my job is to translate it into real costs before anyone signs anything.
The real question hidden behind 'National Instruments vs Broadcom'
A search phrase like National Instruments vs Broadcom looks like a brand comparison. In most cases it is not. National Instruments designs modular test and measurement platforms. Broadcom designs semiconductors. If your goal is to test a chip made by Broadcom, those two vendors are not necessarily substitutes. The chip may be the device under test and the NI system may be the tester that measures it. They can be complements, not competitors.
When I first started this job, I thought the safest move was to compare vendors before looking at application requirements. That assumption was wrong. I only believed it was wrong after we approved a low quote for a modular DAQ system and spent weeks adding cables, software licenses, and signal conditioning. The quote was not deceptive. It was simply incomplete. The missing line items were not visible in any brand-level comparison.
Why the product name is only the tip of the iceberg
Under every tester there are layers. There is the measurement module, but also the connector, the signal conditioning, the software driver, the calibration plan, the power supply, the documentation, and the person who has to integrate all of it. A product name is shorthand for all those layers. The problem is that shorthand gets mistaken for the complete system.
The phrase National Instruments FP 1000 is a good example. The FP 1000 was part of the older FieldPoint distributed I/O family. If you already have a complete FieldPoint system, a spare FP 1000 can be a sensible buy. If you don't, the cheap part becomes expensive very quickly. It may need legacy modules, a specific network interface, software drivers, and power hardware nobody listed in the initial search. That is not an argument against legacy equipment. It is an argument for costing the whole system.
The National Instruments myDAQ sits at the other end. It is a small, affordable device and a great starting point for bench work and engineering education. But when a project plan says 'myDAQ National Instruments' on a production tester line, I stop and ask what happens after the prototype. The purchase price looks like a win. The external signal conditioning, isolation, extra channels, enclosure, and installation labor do not. Those costs have a way of showing up in a second budget request, which is always harder to explain than the first one.
Then there is the N93 tester. N93 is an asset tag in our lab, not a model number. When someone submitted a replacement request for the N93 tester, I asked what the N93 tester did. Nobody could give me a complete answer. We had a fixture, a DAQ module, a LabVIEW program, and a retired technician who had understood the whole thing. The retired technician's knowledge was a real line item, but it was not on any purchase order. Rebuilding the test from scratch cost more in engineering time than the replacement hardware would have cost in the first place.
The most frustrating part is that these are not exotic failures. They happen whenever a product name replaces a measurement requirement. You can hire any vendor and still miss the driver, the connector, the installation panel, or the documentation that the project depends on. The device shows up on time. The project shows up late.
The fix: make the total cost visible before comparing quotes
The solution is not to stop searching for vendor comparisons. The solution is to make the comparison happen after the system is defined, not before.
- Write a one-line test objective before contacting anyone. Something like: measure four thermocouple channels at 100 S/s with an accuracy of plus or minus 0.5 degrees C. It sounds basic, but it filters out most of the confusion.
- Ask the vendor what is not included. I have learned to ask that before I ask for the price. The answer tells me exactly where the hidden work will live.
- Put software, calibration, cables, and engineering time into the same budget line. If they live in different budgets, they are invisible during the purchase and painful during the audit.
If you cannot write the test objective in one paragraph, no vendor quote will save you.
I have nothing against National Instruments, Broadcom, or any of the other names that end up in the same search box. But I have learned to stop treating vendor names as interchangeable options when they are usually different layers of the same problem. A transparent quote might look higher on paper. It is often the cheaper one after the project actually runs.
Leave a Comment