22
\$\begingroup\$

This article Understanding Digital Multimeter DMM Specifications / Specs mention

A digital multimeter will only be able to meet its specifications when it is within a certain environment. Conditions such as temperature, humidity and the like will have impact on the performance. Also conditions such as line voltage can affect the performance. In order to ensure that the digital multimeter is able to operate within its uncertainty specification, it is necessary to ensure that the external conditions are met. Outside this range the errors will increase and the readings can no longer be guaranteed.

A further element to be considered is the calibration period of the digital multimeter. As all circuits will drift with time, the DMM will need to be periodically re-calibrated to ensure that it is operating within its specification. The calibration period will form part of the specification for the DMM. The most usual calibration period is a year, but some digital multimeter specifications may state a 90 day calibration period. The 90 day period will enable a tighter specification to be applied to the digital multimeter, allowing it to be used in more demanding applications.

When looking at the calibration period of the digital multimeter, it should be remembered that calibration will form a significant element of the cost of ownership and after some years will be significantly above that of any depreciation. A long calibration period for the digital multimeter is normally to be advised, except when particularly demanding testing is required.

Is it necessary to calibrate a digital multimeter every year? (From my understanding, only analog multimeters require calibration)

\$\endgroup\$
  • 4
    \$\begingroup\$ What do you mean by "only analog multimeter(s) require calibration"? At some level all multimeters are analog. The ones with a galvanometer and needle usually have a user calibration knob... and are not very accurate to begin with. \$\endgroup\$ – George Herold Dec 30 '14 at 4:16
  • 4
    \$\begingroup\$ Mandatory EEVblog YouTube links: What is calibration?, DIY Multimeter Calibration and How To Calibrate A Calibrator. Enjoy! \$\endgroup\$ – Lorenzo Donati supports Monica Dec 30 '14 at 13:39
  • 5
    \$\begingroup\$ Note that some but not all meters start reading wrongly with low battery levels that are still high enough to maintain an apparently good display. The cal sticker on the Fluke that gets destroyed when the battery is changed is not a mistake. \$\endgroup\$ – Russell McMahon Dec 31 '14 at 0:43
  • 1
    \$\begingroup\$ It mostly depends on how you want to use the DMM - in particular, whether you want to trust the values it gives, and what (lives) depend on you being right. \$\endgroup\$ – Floris Dec 31 '14 at 23:20
  • 1
    \$\begingroup\$ I just calibrated my 3 1/2 digit DMM after 10 years. It was showing 4.999V instead of 5V \$\endgroup\$ – PlasmaHH Feb 24 '15 at 15:07
44
\$\begingroup\$

For hobby/student DMMs, the answer is no. You don't have to calibrate it every year. Please take note of the quote: "A long calibration period for the digital multimeter is normally to be advised, except when particularly demanding testing is required.". For a 3 1/2 digit battery-powered DMM, most never get calibrated after being bought.

If you're using a 6 1/2 digit unit, and measuring microvolts to trouble-shoot medical equipment, that's another story.

It all comes down to how important absolute accuracy is to you.

However, your belief that "only analog multimeter require calibration" is dead wrong. The distinction between analog and digital in this case only applies to the display. An analog multimeter uses its conditioning / amplifier circuits to directly drive an analog meter. A digital unit uses its conditioning / amplifier to drive an A/D converter. In both cases, if the analog circuits get out of whack, the meter will give bad results. A problem is more likely to be noticed in a digital meter simply because DMMs make small errors easier to see.

\$\endgroup\$
  • 2
    \$\begingroup\$ Analog meters (especially ohmeters) sometimes include a fixed reference device of some sort and one or more "calibration modes" where the user is expected to turn a knob until the reader shows a particular value; in some cases, such "calibration" would be a routine part of normal usage. Digital meters can perform such adjustments automatically. A four-wire digital resistance meter could be constructed to perform such compensation entirely automatically, such that the all measurements would be as accurate as a single fixed reference resistor. \$\endgroup\$ – supercat Dec 30 '14 at 21:16
17
\$\begingroup\$

I work for a company that makes medical instruments. Similar to Stephen Colling's answer, UL audits our facilities once a year, so we have one of our meters calibrated a month or so before the audit. Not cheap, around $70 as I recall. The meter we have calibrated is a high-end Fluke. We also have another half-dozen or so cheaper DMM's that have a sticker on them that says "Calibration not required". So we don't have to have all of our meters calibrated. UL seems fine with that.

The calibrated Fluke is then used for any critical tests, for example where we need to record values in a log. The lesser DMM's are just used by engineers for routine tests, do we have an expected voltage on this pin, or what is the value of a resistor. If we get unexpected readings with one of the cheaper DMM's, we can use the Fluke to double-check it. (Theoretically, we could use the Fluke as a reference to calibrate the cheaper DMM's, like using the NIST but locally instead, but we don't really have a way to calibrate them.)

\$\endgroup\$
14
\$\begingroup\$

Calibration means that someone who knows what they're doing has certified that an instrument is performing within its specifications. If that's important to you then you need to get it done. Typical hobbyists don't need that; if their meter is a little off it's okay. Note that calibration does not mean tweaking anything in the meter; it's only checking the meter's response against known sources to see whether it's doing what it's supposed to do. If a meter passed calibration a year ago and it passed again today, you can be pretty sure that the measurements you made during the year are good; if it didn't pass its latest calibration and you've used it to adjust equipment for yourself or your customers since its last calibration you'll have to go back and redo those adjustments after you get the meter fixed and calibrated.

\$\endgroup\$
8
\$\begingroup\$

Digital multimeters can definitely be calibrated. However, the only time I've heard of meters getting calibrated is for paperwork purposes, not because the meters are actually wrong. For example, when my company has something UL listed, the tests must be performed with meters that have been calibrated within the last year. The same for some of our customers' runoff test requirements. I suspect it's a trust issue; they don't want to risk someone gaming the test by altering the behavior of the meters.

There are services that calibrate meters, but it's sometimes cheaper just to buy a new one! We have twenty-year-old digital meters that still read the same as our newly calibrated ones. So unless you have specific reason to think a meter is out of cal, or unless you really need someone to blame if the meter turns out to be out of cal, I wouldn't worry about calibration. Low batteries are more likely by far to give you wrong readings.

Of course, one of the unfortunate aspects of Fluke multimeter designs is that you have to break the calibration sticker to change the battery...

\$\endgroup\$
  • 1
    \$\begingroup\$ The meters aren't ever wrong because they are calibrated more frequently than they need to, as that is far less expensive than hunting down some bizarre issue that turns out to have been due to a bad measurement. \$\endgroup\$ – whatsisname Dec 30 '14 at 15:53
  • \$\begingroup\$ @whatsisname: Is there any mechanism by which a meter that would never be off by more than 0.1% if calibrated yearly, could, if calibrated less frequently, drift by more than 0.1%/year? If a meter "goes bad" such that drift accelerates to 0.1%/month, it may be better to discover it after 12 months than after 12 years, but if one needs measurements within 5%, would there be any benefit to yearly calibration to within 0.1% versus twice-yearly sanity-checks? \$\endgroup\$ – supercat Apr 15 '15 at 19:27
2
\$\begingroup\$

Typically there are two reasons to get devices proffessionally calibrated.

One is when you need a level of accuracy that is hard to obtain and requires periodic checking and adjustments to maintain it. This is likely pointless on a 3.5 digit handheld but potentially more important on a 5-6 digit bench meter.

The other is when you are depending on the equipment to give a legal/safety paper trail and you want independent verification that it remains sufficiently accurate.

Calibration paperwork should indicate the calibration status both when the instrument arrived at the cal lab and when it left the cal lab. The "incoming" data is important as it provides information on the validity of measurements taken with the device since it was last calibrated.

As I understand it usually calibration service for high precision instruments includes adjustments but not repairs. For low precision instruments there may be nothing to adjust.

\$\endgroup\$
1
\$\begingroup\$

Calibration is also used to check that the meter is accurate and that it is within the accuracy as specified by the supplier's manual (e.g a shop cannot weigh a product as 100g if it is actually 105g on another scale from the same supplier). Scales are never 100% accurate so they need periodic checking or recalibration.

\$\endgroup\$
1
\$\begingroup\$

I work at a company where we make RF antenna's and Filters

we only use calibrated equipment for important checks

we use uncalibrated equipment for production (checking as components are installed). and calibrated equipment at specific points in production (mostly to check s-params such as VSWR).

if you get unexpected readings on a uncalibrated machine you can check on the calibrated one (or if you get bad readings on one calibrated machine, you can check the second)

as was mentioned by others, Calibration for DMM's does not involve changing the measurement scale to account for drift (I believe in weigh scales this is the opposite). calibration is just used to certify the accuracy of a device at specific reference measurements. it can give you a good idea of how long your equipment will be good for. it's for keeping track of the consistency in the change in measurements over time (drift)

Continuity check is unaffected, and even uncalibrated and known bad measuring DMM can give readings for voltage, resistance etc. that if not of by a factor of 10+ can still be useful. (checking if a line is live for example, or a motor is shorted)

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.