The Emergency Repair Toolkit: Starrett Micrometers, FLIR Thermal Cameras, and Multimeters

In my role coordinating field service for a mid-size manufacturing company, I've handled more than 200 rush repair jobs over the past decade. Last quarter alone, we processed 47 rush orders with a 95% on-time rate. The 5% we missed? Almost every one traces back to a measurement error—not a speed problem, not a supplier problem, but a "grab the wrong tool" or "trust a bad reading" problem.

This checklist is what I use when the clock is ticking. It won't teach you metrology from scratch, but it will keep you from making the expensive mistakes I've already made for you. Six steps, plus the details most people skip.

Step 1: Diagnose before you measure

It's tempting to think you can just grab a caliper and start checking dimensions. That shortcut costs more time than it saves. The first question has to be: what am I actually looking for? Mechanical wear, overheating, or an electrical fault? Each needs a different tool, and using the wrong one can send you chasing a problem that doesn't exist. I remember a job where a technician spent twenty minutes measuring a shaft diameter while a bearing was cooking twenty feet away. The shaft was fine. The bearing was not. Diagnosis first, always.

Step 2: Match the tool to the tolerance

Once you know the measurement type, match the tool to the required tolerance. This is where value over price hits home. A $15 digital caliper is fine for rough checks, but if the drawing calls for ±0.0005 inches, you need a real micrometer. The Starrett anvil micrometer is my go-to for external dimensions. The anvil is ground flat to prevent rocking, and the ratchet applies consistent pressure, so your reading doesn't depend on how hard you squeeze. For micrometers, the ASME B89.1.13 standard defines accuracy grades. A reputable brand should state compliance; a cheap one often doesn't. I've used cheaper micrometers that felt okay at first, but their readings drifted as the parts warmed up. A quality micrometer costs more upfront, but the total cost of ownership includes fewer reworks and longer tool life. When you're measuring a hundred parts against a deadline, repeatability saves a ton of rework.

Checkpoint before you start: wipe the anvil and spindle faces clean, then close them on a standard to confirm zero.

Step 3: Use the Starrett anvil micrometer the right way

This routine sounds boring, but it's the difference between a real number and a wild guess. Clean the part and the anvil faces first. Even a speck of coolant changes your reading. Then close with the ratchet, not by feel—the ratchet is consistent; your fingers are not. Finally, take three readings at different positions along the part. The average gives you the dimension; the spread you get tells you about taper or out-of-round.

Most people skip the three-reading part. They grab one number and move on. In a rush, that's exactly when errors happen. I learned this the hard way in March 2024, when I trusted a single reading and shipped a batch that came back for rework 18 hours later. That mistake cost us $600 in extra machining. So glad I now treat the three-reading rule as non-negotiable.

Step 4: The Starrett ID micrometer—the step everyone forgets

Internal measurement is a different game. The Starrett ID micrometer is the right tool for holes, but only if you center it properly. What most people don't realize is that a digital readout can look stable even when the contact points aren't at the true diameter. You have to rock the tool slightly—forward, backward, left, right—and look for the minimum reading. That minimum is your actual bore size. On a depth micrometer, the logic flips: you're looking for the maximum consistent reading.

Another thing: blind holes. Check the tool's depth range before you start. If the range is less than the hole depth, you're measuring the sidewall, not the bottom. This is one of those details that doesn't show up in a spec sheet, and it's a classic cause of bad ID measurements.

Step 5: How to use a FLIR thermal camera without chasing ghosts

Thermal imaging is the fastest way to find an overheating component, but it's not quite point-and-shoot. Here's what works for me with FLIR thermal cameras:

• Set the emissivity first. Painted metal is around 0.95; bare, shiny metal is much lower. Leave the default on an unpainted copper busbar and you'll read a wildly wrong temperature.
• Look for relative differences, not just absolute numbers. A suspect part that's hotter than a known-good neighbor under the same load means more than any absolute temperature.
• Adjust the level and span so the target fills the screen. A wide span can hide a serious hot spot.
• Verify with a contact thermometer before condemning a part. Infrared measures surface temperature, and emissivity errors are common.

For fast fault-finding on electronics, an IR thermal camera—what many techs call an IC thermal camera when they're hunting a shorted chip on a board—is the quickest tool I own. Just learn the emissivity setting before you need it. A surprise hot spot is a lead, not a conclusion.

Step 6: What makes the best multimeter for electronics troubleshooting

If the fault is electrical, you need a multimeter you can trust. The best multimeter for electronics troubleshooting isn't necessarily the most expensive, but it should have three things:

• True RMS AC reading, so you get accurate voltages on variable-speed drives and other non-sinewave outputs.
• A low-impedance mode to avoid ghost voltages that fool you into chasing a problem that doesn't exist.
• Proper input protection and visibly rated fuses. A short on a 400 V bus can be dangerous, and a cheap meter might not protect you.

I've seen a $25 meter show 120 V when the actual circuit was delivering 90 V under load. The tech replaced a perfectly good motor based on that reading. The motor cost $1,200, and the problem stayed. The real expense of buying cheap isn't the tool; it's the decisions that tool leads you into. In my opinion, a quality meter—Fluke or equivalent—is worth it, but only if it spec's out for your actual work. Brand alone doesn't fix a wrong range setting.

And before you probe anything, inspect the test leads. Frayed leads cause false readings and personal risk. I swap mine out twice a year, no matter what.

Two last checks before you close the toolbox

First, verify calibration. If your micrometer or multimeter hasn't been calibrated in the last year, don't trust it for emergency work. A calibration sticker costs around $50. Reworking a $5,000 part because of a bad reading costs way more. Second, write down what you measured and when. A clean log turns a one-off failure into a pattern you can actually fix.

If you've ever had to explain why an urgent part came back wrong, you know that conversation is no fun. Most of those calls are preventable with the right checklist and tools that behave consistently under pressure. Next time you hear "we need this tonight", run through this list before you pick up a tool.