If you're responsible for approving test equipment purchases, you've probably run into the same questions I hear every week. I review every instrument order before it reaches the lab floor—roughly 300 line items a year—and I've rejected my share of deliveries for simple spec mismatches. Here are the questions I wish someone had answered for me earlier.
1. Should I get a Keithley Model 2000, 2010, or DMM6500?
I get asked this a lot. The Model 2000 is a good general-purpose 6.5-digit bench digital multimeter. The Model 2010 gives you lower current and resistance ranges, which matters if you test sensors, semiconductor devices, or components with high output impedance. The DMM6500 adds a touchscreen, more memory, and scanning capability. But here's the thing—or rather, the thing I check first—the calibration interval and the range coverage. A used Keithley 2010 with a fresh calibration can serve you better than a new entry-level unit with missing ranges. In 2024, I sent back two instruments because the requested range was not installed.
2. What accuracy spec matters in a Keithley bench digital multimeter?
Accuracy is never one number. On a Keithley 2000, per the datasheet, the DC voltage spec is stated as a percentage of reading plus a percentage of range. People fixate on the first number and ignore the second. For example, on the 10 V range, a 1 V reading might give you something like 0.002% of reading plus 0.0006% of range. The range term changes with the range you select. So if someone tells you a meter is 'accurate to 0.002%,' ask them for the full expression and the range.
Also, always verify the calibration certificate is NIST-traceable and, ideally, ISO/IEC 17025. I've rejected three 'calibrated' DMMs since Q1 2024 because the cert was only a test report, not an accredited calibration.
3. Is 6.5 digits really necessary, or is 5.5 enough?
That depends on what you measure. If your work is pass/fail on 12 V rails and digital logic levels, a 5.5-digit meter is probably enough. For low-level sensor outputs, leakage currents, or drift measurements, the extra digit helps you see trends before they become failures. In one product evaluation, our 5.5-digit DMM could not resolve a 10 µV offset that the spec required. We switched to a Keithley 6.5-digit bench digital multimeter and re-tested 40 boards. The offset was there all along—we simply couldn't see it.
I'm not saying everyone needs 6.5 digits. I'm saying the decision should be based on the smallest increment you need to resolve, not on the brochure.
4. When should a QA lab add one thermal imaging camera?
One thermal imaging camera is enough to start if you choose it carefully. For PCB failure analysis, it helps you locate hot components before they degrade. For incoming inspection of power supplies and connectors, it catches loose crimps and high-resistance joints that a DMM cannot see without applying a load. What to look for: macro focus for small targets, adjustable emissivity, and spot temperature markers. If the camera cannot measure a 2 mm component accurately, the high pixel count does not matter.
I'll be direct: if you are doing thermal validation, a small IR thermometer is not a substitute. I've compared both side by side on a suspect MOSFET, and the camera showed a 30°C hotspot while the IR thermometer gave us an average that looked normal.
5. Where to buy FLIR thermal cameras without getting a gray-market unit?
Buy from FLIR's authorized channel—either their industrial sales team or a regional distributor. That's where you get the full warranty, firmware updates, and calibration documents. If you search for a model and find a price 40% below every other listing, stop and ask for the serial number. I've seen gray-market thermal cameras come with non-English firmware, no valid calibration certificate, and no North American warranty. The savings disappear when you need a repair or an NIST cert.
Where to buy FLIR thermal cameras really comes down to your service region. For our lab, we use a FLIR E-series for routine board work and a T-series for more detailed analysis. Both were ordered through an authorized distributor, and that made the calibration scheduling straightforward.
6. How do I spec a Safeline metal detector for a production line?
Safeline metal detectors work on a different principle than electrical test instruments, but the buying mistake is the same: people focus on a single sensitivity number. A detector rated at 0.8 mm ferrous in air might only hit 2.0 mm through a wet, salted product. Before you request a quote, know your product temperature, moisture content, packaging film type, and line speed. Otherwise, you'll get a machine that works in the seller's demo but not in your plant.
I also ask for test sticks that match the exact reject spec. In one audit, the system passed its daily check because the test stick was larger than the actual contract specification. That's the kind of detail that causes recalls. For the Safeline metal detector, define the critical limit in the HACCP plan and verify the test piece is actually traceable to a reference.
7. Can one instrument replace a bench DMM, a thermal camera, and a metal detector?
No. I get versions of this question more than I expected. A Keithley bench digital multimeter measures electrical signals, one thermal imaging camera measures temperature distribution, and a Safeline metal detector guards against physical contaminants. They answer completely different questions. You might be able to combine functions within a DMM or a thermal camera, but you won't replace a metal detector with a multimeter, and you won't find a thermal camera that measures millivolts.
What you can do is avoid overlapping purchases. Choose a bench digital multimeter with the ranges your products actually need, get one thermal imaging camera with a macro lens option, and buy a metal detector only when your process risk assessment says you need one. That approach has saved us from buying redundant test equipment three times in the last year.
