Starrett, Micrometers, Multimeters & Pipettes: A Practical Guide to Precision Measurement

There's no single "best" precision instrument. That's not a dodge—it's the truth you'd hear from any metrologist.

I'm a senior application engineer at Starrett. I've handled 300+ rush orders in 12 years, including same-day turnarounds for machine shops with a line down. When customers call, the first question they usually ask is about brands: "Is this one better than that one?" But the question that actually saves time and money is: what exactly are you measuring? Get that wrong, and you'll buy a $300 micrometer when a $40 caliper would've done—or worse, trust a $40 caliper at a tolerance where a micrometer is required.

This guide follows the decision tree I use on the phone every day. It covers the four situations I see most often: measuring solid parts, measuring electricity, non-contact detection in automation, and liquid handling in the lab. Each one gets its own lane. Don't skip the last section—that's where you'll figure out which lane you're actually in.

Scenario A: You're Measuring a Solid Part's Dimensions

If the object in front of you is solid and you need to know its length, diameter, depth, or thickness, you're in mechanical measurement territory. This is Starrett's home ground, and it's where two of the most useful resources on our site get used: the Starrett catalogue and the micrometer parts diagrams.

Start with the Starrett Catalogue (Yes, the Printed One)

"I need a micrometer." I hear that nearly every day. The follow-up question is always: what kind? Outside, inside, or depth? What measurement range? Do you need a ratchet or friction thimble for consistent pressure? Inch or metric? The Starrett catalogue is organized around exactly these questions, which is why it's still faster to navigate than a search engine if you know how it's laid out. Product families come first: micrometers, calipers, dial indicators, depth gauges, height gauges. Once you find your family, you narrow by range, then by resolution. Simple, once you know the path.

Here's a story from March 2024 that changed how I think about documentation. A tool and die shop outside Cleveland called at 2 PM on a Thursday. Their QC inspector's micrometer had seized, and they were holding up a $40,000 run. Instead of guessing, I walked them through the catalogue section by section: outside micrometer, 0 to 1 inch range, mechanical with a ratchet stop. Replacement shipped that evening; they were measuring by Friday morning. No heroics—just following a structured index. That's when it clicked for me: the catalogue isn't just a price list, it's a diagnostic path.

If you don't know which tool family you need, one question settles it: which dimension matters? Outside diameter and wall thickness point to an outside micrometer. The distance between two parallel faces points to a caliper. Small surface-to-surface differences on a bench setup point to a dial indicator. The catalogue's section headers are really a decision tree written by professionals.

Troubleshooting? Read the Starrett Micrometer Parts Diagram

You'd be surprised how many experienced machinists replace a $200 micrometer because a $6 part wore out. The Starrett micrometer parts diagram exists to prevent exactly that. It's an exploded view that labels every component—spindle, thimble, frame, ratchet stop, lock lever—with a part number. Have one open while you call technical support, and the conversation changes instantly. Instead of twenty minutes of vague descriptions ("it's the... clicking wheel? no, the knurled wrap part"), it's: "I need the ratchet stop assembly, part number 436-xxxx." (I really should keep a printed diagram in my desk instead of digging through PDFs each time.)

The cost math matters too. A replacement part, a bit of solvent, and an hour of your evening routinely bring a micrometer back to spec for under $50. Compare that with a $250 replacement, plus shipping time and the hassle of establishing a calibration baseline on a new tool. Repair-first sounds counterintuitive in a throwaway culture, and for a solid micrometer, it's the right call. But don't be sentimental about it: if the frame itself is bent or the spindle threads are damaged, stop, and buy the new tool.

(And yes, I've got my own micrometer at home waiting on a new ratchet assembly. I know. The cobbler's children have no shoes.)

One more thing: any claim of accuracy—new tool, repaired tool, doesn't matter—requires calibration. If your work demands real confidence, the tool goes to calibration and you log the certificate. That's not a loophole; it's good practice.

Scenario B: You're Measuring Electricity

"What is the best multimeter for electricians?" I get this one from contractors constantly, and my honest answer includes a point some don't expect: we don't make multimeters at Starrett. There's a reason for that. Electrical testing is a different discipline with different standards, different failure modes, and a higher cost when something goes wrong. I'd rather see you pick the right tool from a specialist than a so-called "total solution" from a company that doesn't live in that world.

So here's how I help customers evaluate a multimeter, regardless of brand:

• Check the CAT rating first. An electrician working on mains distribution needs a CAT III or CAT IV rated meter. That's an arc-flash safety specification, not a marketing badge. Ignore it and you're gambling with your hands.
• Prefer true RMS. Motors, VFDs, and modern lighting loads distort waveforms. An average-responding meter converts that distortion into readings that look confident and lie with precision. True RMS reads the actual heating value.
• Drop resistance is real. Your meter will fall off a ladder eventually. The brands that dominate here—Fluke and Klein Tools are the ones our customers trust most—engineer for that. The cheap clones look identical on spec sheets and are unforgiving in real life.

And when you search "best multimeter," keep the FTC in mind. Per the Federal Trade Commission's advertising guidelines (ftc.gov/business-guidance/advertising-marketing), claims like "best" have to be truthful and backed by evidence. A top-10 list with no methodology, no test conditions, and no data is not a recommendation; it's an ad.

A few years ago, a purchasing manager showed me an online marketplace caliper that cloned Starrett's dimensions and felt respectable in the hand. Forty percent cheaper, same ratings on paper. My gut said it wouldn't hold up on a shop floor. Two months later, the reading jumped about 0.0002 inch every time the battery door flexed. The spec sheet was right, and the real world was more complicated. That's also why we don't sell clones alongside the real thing.

Bottom line for Scenario B: we can help you trust the mechanical side of your work. For the electrical side, go to a specialist. That's not a limitation—it's how you get the right answer.

Scenario C: You Need Non-Contact Detection

Sometimes you don't need a measurement readout at all. You need to know whether something is present—a part sliding down a chute, a full bin, a pallet in position—and you can't touch the object. That's the world of the photoelectric sensor.

A photoelectric sensor uses light to detect an object's presence or position. Three families dominate the market:

• Through-beam: emitter and receiver are separate units pointing at each other. The object breaks the beam. Longest range and most reliable, but you have to mount both sides.
• Retroreflective: emitter and receiver share a housing, and light bounces off a reflector. Simpler installation, medium range.
• Diffuse: light reflects directly off the target back into the receiver. Shortest range and cheapest—but easily fooled by dark or shiny parts.

This isn't our field at Starrett either. When automation customers ask, I point them to the specialists—SICK, Banner, Keyence, and their peers live and breathe this. What I can add is the most common mistake: picking a diffuse sensor where through-beam is required, then spending days blaming "bad sensors" for what was a wrong sensor family all along.

Define the distance, the object's reflectivity, and the environment (dust, moisture, pointing sunlight) before you open a datasheet. The sensor family is the decision; the brand comes second.

Scenario D: You Need Your Pipette to Be Trustworthy

The most interesting precision-instrument conversation I have is about pipettes, and the search phrase "how to calibrate pipette eppendorf" is a frequent one. The honest answer is more nuanced than the YouTube videos suggest.

Pipette calibration is done gravimetrically. You dispense distilled water, weigh it on a high-resolution analytical balance, and convert mass to volume using the density of water at the measured temperature. At 20°C, water density is about 0.9982 g/mL. At 24°C, it's slightly lower. That "slightly" is exactly the problem: a pipette specified at ±1% or tighter leaves little room for sloppy inputs, and water density at the wrong temperature quietly eats your error budget.

If you want to calibrate Eppendorf pipettes in-house, you need: a balance with resolution an order of magnitude better than the pipette's tolerance, a room that holds temperature within a degree or two, and hands that dispense with machine-like consistency. One degree of temperature shift, a draft across the balance pan, or a warm palm near the tip destroys the reading. And you can't just verify one volume—an adjustable pipette behaves differently at 10 µL than at 200 µL. The traceability standard referenced by most accreditation bodies is ISO 8655-6, and it's a procedure, not a suggestion.

Now the honest part. I've watched labs "save money" by doing this in-house on a decent balance. After technician time, uncertainty budgets, and audit documentation, most discover that sending the pipettes to an accredited lab—Eppendorf's own service or any ISO 17025-accredited provider—is not only more defensible, it's often cheaper. I understand the appeal of self-sufficiency here; it feels like a basic measurement. But the consequences of a bad calibration beat the cost of a good one every time. A pharma QC lab we work with rushed a release through on an in-house check that didn't account for humidity. The out-of-tolerance investigation that followed cost them six days of release holds. Six days of expensive, clean, sitting-idle facility time. (Note to self: never forget how quiet that production floor became.)

If your pipette is due, build a calibration schedule—every 3 to 12 months depending on usage—and commit to it. The lab that treats calibration as an ongoing system, rather than a panic event, is the lab I trust.

Which Scenario Describes You?

Let's narrow it down. When I triage a rush order, I run three filters: time (how many hours until you need it), feasibility (can the tool actually hold that tolerance), and risk (what's the worst case if you choose wrong). Try those filters on these questions:

1. What are you actually measuring? A solid part's dimension puts you in Scenario A. Start with the Starrett catalogue; if a micrometer is misbehaving, pull up the parts diagram. An electrical quantity puts you in Scenario B: check CAT rating, true RMS, and a brand with a field reputation. Presence or position without contact puts you in Scenario C: pick the sensor family first, the brand second. Liquid volume accuracy puts you in Scenario D: calibrate, but let an accredited lab do it unless you're fully set up for regulated work.

2. What tolerance are you working to? If you're checking a rough cut before a finishing pass, stop overthinking and buy the practical tool. If you're holding 0.0001 inch or micro-volumes, overthinking is underrated.

3. What happens if you're wrong? A scratched part is a minor annoyance. An arc flash, a shut-down production line, or a released pharmaceutical batch that fails is a completely different scale. Let the consequence drive the budget.

4. Is the clock running? It is. It's always "the line is down" or "the audit is Friday" or "the customer moved the deadline." When that's your situation, the decision tree matters even more—it stops the panic-scrolling and puts you in the right lane fast.

And while we're on time: shipping costs are going up, and USPS rates as of January 2025 (usps.com/stamps) put a one-ounce First-Class letter at $0.73. Package rates climbed too. But nobody ever complained about courier fees on the day the part saved a production run.

There is no universal best precision tool. There's the best tool for the thing you're measuring. Figure out what that thing is first—the right instrument tends to reveal itself.