Output voltage: 10.000V DC
Sensing: internal or remote sensing to compensate for output cable voltage drop
Basic accuracy: ±0.03% (±3mV) without adjustment, ±0.002% after optional trim adjustment and calibration
Long term drift: <15ppm per 1000 hours, mostly in first year of operation
Temperature coefficient: 3ppm/°C between -25°C and +85°C
Maximum output current: 10mA
Noise on output: less than 6mV RMS
Load regulation: less than ±50μV/mA for loads up to 10mA
Supply line regulation: less than 200μV/V
Power supply: 12V AC, current drain <60mA
Most of us don't ever get our DMMs calibrated, though they do drift out of calibration over years of use. If you are using them in your occupation, they should be checked every year or so – otherwise how can you trust the readings?
Fig.1: block diagram of the AD588 voltage reference. It contains four op amps (A1-A4) plus a "buried" zener and its current source to provide the voltage reference cell.
But it can cost quite a bit to send a DMM away to a standards lab for calibration – more than many DMMs are worth. So generally we either hope for the best or simply buy a new DMM if we suspect that our existing meter has drifted too far out of calibration.
Back in the 1970s when DMMs first became available, the only practical source of an accurately known DC voltage was the Weston cell, a wet chemical “primary cell” which had been developed in 1893 and had become the international standard for EMF/voltage in 1911 (see panel). It produced an accurate 1.0183V reference which could be used to calibrate DMMs and other instruments.
Unfortunately, Weston cells were fairly expensive and few people had direct access to one for meter calibration. As a result, most people tended to use a reasonably fresh mercury cell as a “poor man’s” voltage reference. Fresh mercury cells have a terminal voltage very close to 1.3566V at 20°C and this falls quite slowly to about 1.3524V after a year or so. Silver oxide cells can also used for the same purpose, having a stable terminal voltage very close to 1.55V.
Of course, batteries of any kind have a tendency to obey Murphy’s Law and usually turn out to have quietly expired before you need them. And although mercury and silver oxide cells have quite a long life, especially if you use them purely as a voltage reference, they certainly aren’t immune to this problem. Which means that these batteries make a pretty flaky voltage reference, at best.