How to Verify Your Precision Multimeter: A 5‑Step Quality Checklist for Keithley 2000 and Benchtop DMMs

Who This Checklist Is For

If you’re commissioning a new batch of six‑and‑a‑half‑digit multimeters—say a Keithley 2000 or a DMM6500—or you’re integrating a benchtop DMM into a production test line, this checklist saves you from the standard “trust the datasheet” trap. I’m a quality compliance manager at a mid‑size test equipment distributor. I review roughly 200 instruments every year before they go to customers, and I’ve rejected about 12 % of first deliveries in 2025 alone because specs were technically “within tolerance” but the real‑world behavior didn’t match the label.

This works best when you’re:

• Setting up a new lab station with a Keithley 2000 or similar benchtop DMM
• Evaluating refurbished units from resellers
• Comparing a precision DMM against a handheld multimeter like a Fluke 87V
• Building a test system that also involves refrigerated centrifuges or thermal camera iPhone attachments (weird combination? Actually, happens more than you’d think)

Step 1: Verify the Calibration Certificate’s “Effective” Date

Sounds obvious, but I’ve seen certificates dated way after the unit was built—someone backdated the sticker. Check the actual test date vs. the instrument serial number’s production date.

For the Keithley 2000, the standard 1‑year accuracy spec assumes calibration within 90 days of shipment. If the cert is older than that, your “new” unit is already on the clock. What I do: request the as‑found data (before adjustment) from the vendor. If the 10 V DC reading was off by more than 5 ppm at the time of calibration, that’s a red flag—the reference may be drifting.

“I don’t have hard data on how many instruments fail this check industry‑wide, but out of the last 30 Keithley 2000s I’ve processed, three had certificates that were essentially ‘cosmetic.’ The as‑found data showed they were already out of spec before adjustment.”

Step 2: Run a Simple Self‑Test & Offset Null

Every Keithley benchtop DMM has a built‑in self‑test. But that only checks internal diagnostics—it doesn’t catch input bias current or thermal EMF errors. After power‑on (wait at least 30 minutes for warm‑up), short the input terminals with a low‑thermal (<10 µV) copper short. Read the DC voltage on the lowest range (e.g., 100 mV).

If the reading is above 3 µV, you have a thermal EMF problem—maybe the input jacks are corroded, or the shorting plug is cheap brass.

I actually saw a batch of refurbished units where every reading hovered around 8 µV. The vendor said “within spec.” I rejected the entire lot. The $2,000 redo cost them more than just replacing the input connectors.

This step also applies when you’re checking instruments that live inside environmental chambers next to a refrigerated centrifuge—the temperature gradient across the terminals can easily generate 10 µV.

Step 3: Check AC Bandwidth with a Known Source (Not Your Signal Generator)

Datasheets for the Keithley 2000 claim AC bandwidth to 100 kHz, but real‑world behavior depends on source impedance and cable length. Use a certified AC calibrator (e.g., Fluke 5500A) or a precision thermal transfer standard. If that’s not available, at least compare against a second trusted DMM. I’ve found that the AC accuracy on the 100 mV range often degrades above 10 kHz due to parasitic capacitance in the input relay.

One time the numbers said the calibrator was good, but my gut said the DMM was reading low. Turns out the cable I was using had a broken shield—classic gotcha. Put another way: test the whole signal chain, not just the instrument.

Step 4: Evaluate the “Human Factor” – UI & Speed

This isn’t a spec you’ll find on a brochure, but it matters. The Keithley 2000 has a button‑heavy front panel—great for benchtop, less great if you’re programming it remotely. For automated production, I’ve seen engineers spend a day just timing the SCPI commands because the settling time on the 100 mV range was 50 ms longer than advertised.

If you’re used to a portable meter like a Fluke 87V (which many people search for “where to buy fluke multimeter”), the Keithley feels like a tank. It’s slower, bigger, but way more precise. The trade‑off is obvious once you measure a 1 µV change in a temperature sensor—your handheld Fluke would never see it.

I wish I had tracked the number of returns caused by “it’s too complicated” vs. accuracy issues. What I can say anecdotally: maybe 20 % of first‑time buyers of a Keithley 2000 would have been better off with a mid‑range handheld. But if you’re running a semiconductor lab or a metrology bench, there’s no substitute.

Step 5: Test with Real‑World Loads (Not Just Ideal Resistors)

A precision DMM looks great on a dummy load but fails when connected to a refrigerated centrifuge’s motor controller (inductive kickback) or a thermal camera iPhone module that draws microamps from the same power rail.

I keep a set of “dirty” loads: a small relay coil, a thermoelectric cooler, and a 10 µF capacitor. I measure the voltage ripple while the load is switching. If the reading jumps more than 0.01 % of range, the DMM’s input filter isn’t aggressive enough, or the guard shield is picking up noise.

For the Keithley 2000, the default reading rate is about 5 readings/s. Switch to PLC 10 to get better noise rejection on a dirty line. If you still see fluctuations, consider adding an external low‑pass filter.

Notes & Common Mistakes

Don’t assume “keithley oscilloscope” exists

Occasionally someone asks for a keithley oscilloscope. Keithley doesn’t make oscilloscopes—they focus on DC precision (SMUs, DMMs, electrometers). If you need an oscilloscope, you’re looking for Keysight, Tektronix, or Rohde & Schwarz. But if you combine a Keithley DMM with a thermal camera iPhone attachment for PCB hotspot identification, you actually get a pretty decent fault‑finding station.

Where to buy fluke multimeter vs. Keithley

If you’re searching “where to buy fluke multimeter,” you probably need a rugged handheld for field work—that’s not Keithley’s lane. Keithley shines in the lab: 6.5‑digit resolution, low‑noise, tight accuracy. For most production test, a Keithley 2000 will outperform any handheld, but it costs more and requires mains power. Pick the tool for the job.

Refrigerated centrifuge interference

I’ve seen engineers try to measure a centrifuge’s temperature using a Keithley 2000 without galvanic isolation. The compressor spikes can couple into the measurement path. Always use guarded inputs or an isolation amplifier if your DMM must be connected while the centrifuge is running.

Final thought

The $200 difference between a “barely in spec” Keithley 2000 and one that’s truly well‑calibrated shows up in customer trust. I’ve rejected units that passed datasheet specs because real‑world behavior was borderline. That’s the quality‑as‑brand‑image angle: if a car company uses your DMM to validate battery voltage, every 10 µV counts.

Bottom line: run these five steps before you put a precision DMM into production. It takes about an hour—and it’s saved me from at least three expensive field failures in the last year alone.