Plenty of used test equipment is listed as "calibrated", "cal'd" or "NIST traceable". Some of those claims are backed by real paperwork, and some of them mean a sticker. Here's what calibration is, what a listing has to show before the word means anything, how much of it a home bench needs, and which older instruments can lose their calibration sitting on a shelf.
What calibration is
Calibration is a comparison. A lab feeds your instrument known values from a reference standard and records how far its readings are from the reference. Strictly, the word covers only that comparison. If the instrument is out of tolerance, the lab may also adjust or repair it and then measure again. Tektronix's certificates call the two sets of readings As Found (taken before any adjustment or repair) and As Left (taken after). As-found data tells you how well the instrument held up since its last calibration. As-left data tells you where it ended up.
Traceability, NIST and ISO/IEC 17025
A traceable calibration links a result to a reference through a documented, unbroken chain of calibrations, each one adding to the measurement uncertainty. In the US, that chain runs back to NIST. NIST is clear that it doesn't certify anyone else's traceability. Supporting the claim is the job of whoever provides the measurement. So "NIST traceable" in a listing is a claim some calibration lab made, and it only means something with that lab's paperwork attached.
ISO/IEC 17025 is the standard for the competence of testing and calibration laboratories. A lab accredited to it has been assessed by an accreditation body. In the US the main ones include A2LA and ANAB. Tektronix spells out the difference on paper. A standard certificate says the instrument was calibrated with traceable standards, and a full data report is optional. An accredited certificate meets ISO/IEC 17025, carries the accrediting body's logo and reports measurement data and uncertainty for each parameter tested.
What a sticker or a "calibrated" title proves
On its own, very little. A sticker says someone, at some point, signed off on the instrument. It doesn't say what was measured, whether the instrument passed as found or whether it was adjusted, and anyone can print a sticker. A cal-due date that's long gone doesn't mean the instrument now reads wrong, and a recent one doesn't mean it reads right.
A calibration claim means something when the seller can show a dated certificate from a named lab, with data, for that serial number. Tektronix notes that each certificate carries a serial number tying one calibration to one instrument, along with the dates, the environmental conditions and the standards used. Check the serial number on the certificate against the one on the instrument in the photos. If they don't match, the paperwork belongs to a different unit.
Calibration intervals
Makers publish accuracy specs for a set period after calibration. The HP 34401A datasheet gives its accuracy at 24 hours (at 23 ± 1 °C), 90 days and 1 year (both at 23 ± 5 °C), relative to calibration standards, and the allowed error grows as the period gets longer. Fluke specifies the Fluke 87V's accuracy for a period of one year after calibration. That's why one year is the common interval for DMMs. It's where the published spec stops.
The interval is a policy, set by people. Tektronix says there's no one-size-fits-all calibration interval. In practice the maker suggests one, and the owner sets the real interval from how the instrument is used and how well it has held its readings. An instrument past its due date has no current paperwork behind it. A quality-controlled lab can't use it for that reason. A hobbyist often can.
What a home bench needs
Most hobby work needs readings that are close and that repeat. It doesn't need a certificate. You get there with a known reference. Measure the same stable source with two meters, ideally of different makes and designs, and see whether they agree within both meters' specs. A small voltage reference, checked once against a recently calibrated meter, gives you a fixed point to come back to. If a 34401A and a Keithley 2000 agree on that reference, both are very likely fine for anything you'll build at home. If one drifts away from the other over the months, you know which to send out.
Pay for a real calibration when your numbers have to satisfy someone else, such as a customer, an employer or a published result. Then you want the accredited certificate with data.
Calibration memory and battery-backed RAM
Many instruments from the 1980s and 1990s store their calibration as numbers in memory. Those numbers have to survive between calibrations. On some instruments they live in RAM kept alive by a battery, and when the battery runs flat, the calibration goes with it.
The Tektronix 2465B keeps its calibration values, its last front-panel setup and its running-hours count in a Dallas DS1225, a RAM module with its own lithium cell, soldered to the A5 board without a socket. Dallas guaranteed it for ten years. Hugo Holden, who wrote up the repair, has seen modules fail at 13 years and others still working at 20. When one fails, the scope needs a full calibration. The usual fix is to fit a socket and a new or emulated module, then recalibrate. Holden suggests reading the module out to a file named for the scope's serial number after calibration, so a later failure only means programming a fresh module.
The HP 3457A stores its calibration coefficients in RAM backed by an on-board lithium battery, and xdevs.com notes the calibration is lost if the battery discharges. Some instruments tell you when their stored calibration is bad. The Fluke 87V shows CAL Err for invalid calibration data. Others may give you no message at all, so a reading checked against a known reference is the real test.
When you're looking at one of these, ask the seller when the battery or memory module was last replaced and whether the instrument has been calibrated since. A good answer names a date and a lab. If nobody knows, plan on a calibration before you trust its readings. For the rest of what to check on a listing, see how to check used test equipment before you buy.
