Indeed, the stability of high-quality multimeters can be much better than specced in the datasheet, the value in the datasheet is probably something like the three-sigma value under worst-case conditions, most meters are not worse-case. Especially the long-term drift is often quite low. It's not unusual for a meter to still meets it specs after many years without calibration. The trick is how to know if this is the case. This obviously won't cut it in a commercial environment where it's used for production testing, you want impressive official documents to impress your managers and customers, and to claim that your process that produces equipment that dies after two years has a constant quality, and will always die between two years and two and a half years.
Calibration is all about 'probably accurate' anyway, the minute the meter is put in a box and tossed in the UPS van, it might have drifted (although modern meters without trimmers are much less susceptible to shocks). The manufacturer has designed it so it probably won't drift more than X within a year, and you can use your own historical data to back it up. Sometimes they perform worse than manufacturers specs, I've heard about a metrologist that had some meters on a three month cal cycle, and they still barely met their 1-year specs.
Even with real calibration, all you have is a certificate that it was within factory specs at some time in the past, a datasheet from the manufacturer that claims that it won't drift more than X in the calibration interval, and if you're lucky some historical stats. None of this guarantees that the measurement you take today is accurate. I think Dave touched on this in a past blog.
Home
Help
Search
Login
Register
