Buying Keithley Gear and Lab Equipment: A Procurement FAQ from the Person Who Signs the PO

I've been handling procurement for a 120-person dental equipment manufacturer since 2020. That means I buy the test gear for our electronics lab, the microscopes for our QC department, and—somehow—the coffee for the break room. On average I process 60–80 orders a year across maybe eight or nine vendors. Here are the questions I actually get asked, with the answers I've earned the hard way.

What should I budget for a Keithley 2000 multimeter—really?

When I first started buying test equipment, I assumed the lowest quote was always the best choice. Three budget overruns later, I learned about total cost of ownership.

The Keithley 2000 is a 6.5-digit bench multimeter. A new unit typically lists in the $1,200–$1,500 range from authorized distributors (as of Q1 2025—verify current pricing before you budget). But that's the number on the invoice, not the number in your budget.

Add a calibration certificate if your lab needs traceability—$150 to $300 annually depending on turnaround. Add the rack-mount kit if it's going in a test rack, around $100. Add spare fuses, low-thermal cables, and the small stuff you forget about until the day you need it.

I now budget 1.3x to 1.5x the unit price for any bench instrument. If a quote comes in way lower than everyone else's, ask what's excluded. There's always something.

Do we really need the Keithley 2000 multimeter software?

Depends on whether you're logging readings by hand or by the hour.

Honestly, the free route works fine if you've got an engineer who likes tinkering. A basic IVI driver plus a Python script can handle most benchtop automation. But I watched our lab manager spend two weeks—actually, closer to two and a half by the time he finished—building a workaround that a licensed software package would have handled in an afternoon.

That's a TCO question, not a price question. If his time is worth $60 an hour, two and a half weeks of half-time work is around $3,000. Suddenly the software license doesn't look so expensive. Plus the licensed packages come with support. When something breaks the night before a customer demo—and it will—"check the user forum" is not a plan.

Time-pressure version of this story: last fall we had a customer audit in two days and our data logging wasn't set up. I had about four hours to authorize a purchase. Normally I'd get three quotes and sleep on it. Instead I went with the licensed software from our existing distributor, paid a rush fee, and it worked. Would I do it again? Yes. Do I want to rely on that again? No.

Why does everyone care about the Keithley logo on the front panel?

This one surprised me. I used to think a meter was a meter.

Then we bought a "Keithley" 2000 off an auction site for what looked like a great deal. Saved about $400 versus a distributor-refurb price. The unit arrived and wouldn't hold calibration. The logo on the front was printed—not etched, printed. Turned out to be a rebuilt chassis with aftermarket internals.

Ever since Keithley became part of Tektronix—that was 2010, if you're curious—the authorized channel has gotten stricter about gray-market gear. If you're buying used, get the serial number and verify it with Tektronix before money moves. I now do this for anything over $500, no exceptions.

That $400 in "savings" cost us roughly $800 in return shipping, restocking fees, and a rush replacement order. Lesson learned the expensive way.

Is a 77 multimeter still worth having when we've got bench meters?

Yeah, actually. They solve different problems.

The Fluke 77 is a handheld workhorse—drop-resistant, battery-powered, good for crawling under a machine or checking a live circuit without dragging a bench meter over. The Keithley 2000 is a precision bench instrument. You're not going to clip it to your belt.

Our electronics team has both. The 77 gets used constantly for sanity checks and quick diagnostics. The Keithley handles anything that goes into a test report or a customer deliverable.

If you're building a lab from scratch and can only buy one first: get the handheld. You'll outgrow it, but you'll never stop using it. That said, budgets being what they are, sometimes you just buy whichever the project actually needs. I get it.

When does "true RMS" actually matter?

True RMS matters when your waveform isn't a clean sine wave. Which, in the real world, is most of the time.

An averaging meter is calibrated for sine waves. Feed it a distorted waveform—a switching power supply output, a motor drive, anything with harmonics—and you'll get a number that's confidently wrong. Not off by a little. Off by enough to matter. Depending on the waveform shape and crest factor, published comparison data from instrument manufacturers shows accuracy differences of 10–40% between averaging and true RMS meters on non-sinusoidal signals (verify against your specific instrument's datasheet).

A true RMS multimeter calculates the root-mean-square value directly, so it reads correctly regardless of waveform shape. For our dental equipment—which has motor drives and switching supplies—this isn't optional. If you're only checking mains AC in a facility closet, an averaging meter is fine. If you're debugging a board, get true RMS.

Someone asked me about Zeiss vs Global dental microscopes—what's the procurement angle?

I didn't expect this question either, but our QC manager asked, and the framework is the same.

Zeiss (the German optics company) and Global (the American surgical and dental microscope brand) both make excellent instruments. Both also have the same hidden cost categories: service contracts, calibration, replacement illumination modules, and lead time on parts.

My advice to our QC manager was the same advice I'd give anyone buying precision instruments: get the 5-year TCO from each vendor, in writing. Ask about service turnaround. Ask what happens when the unit is out of warranty. Then compare.

The other thing I've learned: lead time is a cost. Zeiss tends to have more North American service locations, which usually translates to shorter downtime. Global tends to be more configurable, which means you can spec exactly what you need instead of paying for features you won't use. Which one wins depends on whether your bottleneck is uptime or overspec. Ask your lab what they actually lose sleep over before you make the call.

One thing I'd add

If you take nothing else from this: the unit price is a starting point, not a decision. I've seen a $600 quote beat a $500 quote more times than I can count once you add in shipping, setup, training, and the time it takes to fix what the cheap vendor got wrong.

Ask what's included. Ask what happens after the sale. Do the math on your own time. That's the whole framework. Everything else is just specific numbers.