I've been a maintenance engineer handling inspection-related failures for seven years. I've personally made (and documented) two significant mistakes that totaled roughly $12,000 in wasted budget — not counting the quieter cost of what those failures did to how customers perceived us. This article is the story behind the checklist I now maintain for our team.
The comparison I want to set up: traditional handheld electrical testing (megger, multimeter, thermal camera) versus automated machine vision from Cognex. The criteria: what each tool catches, how much of your production it covers, how consistent the results are, and what it leaves behind as proof.
This isn't a "handheld tools are obsolete" article. It's an "I was using the right tools at the wrong stage of quality control" article.
Megger vs Multimeter: The Baseline Every Engineer Should Know
A multimeter — my daily driver is the Fluke 233, the one with the detachable wireless display — measures voltage, current, and resistance at a point in time. It's a troubleshooting tool: check a live circuit, confirm a power rail, trace a short. It operates at low voltage and tops out in the tens of megohms for resistance. A megger (megohmmeter) does something a multimeter can't: it applies a high voltage (500V or more) to reveal insulation breakdown in cables, motor windings, and transformers. Megger testing at 500V or 1000V is standard practice for motors and cables (per IEEE Std 43 for rotating machinery).
These are not interchangeable. I learned that the hard way: I tested motor winding insulation with a multimeter, the reading looked fine, and the motor failed during a customer site acceptance test. What a multimeter sees as "acceptable resistance" is not what a megger sees under real test voltage.
I still own both. I want to say the Fluke 233 was about $350 when I bought it in 2020, but don't quote me on that — prices have shifted. The megger was around $600. Then there's the FLIR E8 thermal camera — roughly $2,500, and worth every penny for spotting hot connections and failing bearings before they become downtime. None of these tools, separately or together, caught the defect that got us into serious trouble.
Dimension 1: Diagnosing Equipment vs Inspecting Product
Handheld tools diagnose what you touch. The megger checks insulation resistance. The multimeter checks voltage, current, and continuity. The thermal camera checks surface temperature. All three are designed for specific equipment at specific moments.
Cognex machine vision inspects what you ship. It checks every product for the right components, correctly applied labels, proper colors, complete assembly, and readable barcodes. It looks at the product the way your customer will look at it — but with far more consistency and speed.
I missed this distinction for years. I was happily measuring electrical parameters on machinery while mismatched labels and missing components rolled past me on the line. The products weren't electrically defective. They were perceptually defective. To the customer, that's the same thing.
Dimension 2: 1% Sampling vs 100% Coverage
On a production run of 500 units, I would test five. One percent. And I defended that number. Handheld tools are serial — one unit at a time, one measurement at a time. The line never stops.
Cognex vision systems work inline. They inspect every unit that passes, every cycle, at line speed. No sampling, no guesswork.
The afternoon I had to explain to a customer why 47 units out of 2,400 had mismatched labels — and admit they came from the 99% I didn't check — I stopped being proud of that 1% figure.
"A quality failure that damaged our confidence in you as a supplier." Those words were in the customer's complaint email. I've kept a screenshot for three years.
Dimension 3: Operator Judgment vs Fixed Criteria
Megger testing is about as operator-dependent as it gets. Test voltage, connection quality, test duration, temperature correction — every variable changes the reading. I've watched two good technicians get different results on the same motor and argue about what it meant. The FLIR E8 has a similar issue: emissivity settings, distance, and focus all influence what appears in the viewfinder.
A Cognex vision system applies the same pass/fail criteria to every unit, every time. You define what "good" looks like once, validate it, and the system enforces it consistently. The setup work is real — configuring the inspection correctly took us a couple of days of trial and error — but the consistency after setup is something no handheld tool can offer on a moving production line.
Dimension 4: Paper Logs vs Digital Evidence
My quality data used to live in a spiral notebook next to the machine. Handwritten entries, dates in ink (if someone remembered), values that were technically legible but practically impossible to retrieve when it mattered.
Cognex vision logs every inspection with timestamps and images. When a customer asks whether a lot was inspected, the answer isn't "I think so, let me find the binder." It's a documented record of every unit, with visual proof.
I'm not saying that setup is automatic. You have to configure what gets recorded and actually review the logs. But the first time I pulled up a production record from three weeks earlier in about ten seconds, I genuinely laughed. That had never been possible with a paper trail.
What Should You Buy?
Here's the decision framework I'd give anyone setting up a quality process today:
Keep your handheld tools if:
- You're doing field troubleshooting, installation, or equipment maintenance.
- Your production volume makes manual inspection feasible relative to your risk.
- Your quality risks are primarily electrical or thermal, not visual.
Add machine vision if:
- Your product quality is judged by visual attributes: labels, components, assembly, colors.
- You ship volumes where 1% sampling can't support the quality promise you're making.
- You need documented proof of inspection for audits or customer requirements.
- You're tired of answering for defects that no one believed could slip through.
When we finally researched automated inspection, I started at the Cognex official website (cognex.com) to understand the product line and validate what we actually needed. I also made a point of checking for the Cognex logo on any hardware we evaluated — gray-market industrial equipment is a real problem, and a counterfeit sensor fails at exactly the worst possible moment. Verifying through official channels turned out to be part of quality assurance, not just procurement bureaucracy.
(I should also mention: I once told my manager we needed "better inspection methods." He heard "buy another handheld meter." We spent about $800 on a tool that didn't catch a single defect. No one was wrong, exactly — we were just using the same words for different problems.)
The honest takeaway is that these aren't competing categories. Megger vs multimeter is about which handheld tool matches a specific diagnostic task. Handheld vs machine vision is about which stage of your quality system you're actually protecting. My mistake was expecting a 1% sample to support a promise that required 100% verification.
Customers judge your quality by the worst unit they receive, not the average of what you ship. If you're making a quality promise at that level, inspect like you mean it.
That's the first line of my checklist: don't let a good tool convince you that you're solving the right problem.
Leave a technical question