Calibration

[Dane Bear]

How often do good multimeters (such as Fluke 289 or 87 V) need to be sent in for calibration?

[SeanB]

Depends, how it is used and environment. At least every 6 months for one tossed in a tool bag and dropped on occasion, to every 2 years for a lab unit.

[Bored@Work]

Depends, how it is used and environment. At least every 6 months for one tossed in a tool bag and dropped on occasion, to every 2 years for a lab unit.

And never for one sitting on a hobbyist's bench when the hobbyist is happy with it as it is.

It comes down to how certain you need to be that the meter is still within specs. In a professional environment it might also include liabilities when something went wrong.

[M. András]

when i bought mine from a fluke distributor the guy told me i should take it to calibration within a 6-8 years from now on but he bets it will be within specs. i could have asked them there to hook it up for  calibrator and check it on the spot cos the cal lab was on the other side of the floor

[EEVblog]

It comes down to how certain you need to be that the meter is still within specs. In a professional environment it might also include liabilities when something went wrong.

You can never be "certain" that a meter on it's own is within spec, even if it was calibrated yesterday.
The only way to do that is to double check with a second meter, and some companies will actually have procedures that demand this for certain critical measurements.
Periodic calibration checking is about building up "confidence" in the meter over time, so you can fairly confident when you use it that it hasn't drifted out of spec since the last calibration.
For example, Agilent recently increased their recommended calibration interval on their X series scopes from 1 year too 2 years, because they have done further testing and now have greater confidence in the stability of them.
Companies might do the same with their procedures. For example, a new meter might get calibrated every 3 or 6 months at the start to build up confidence in how much it drifts. And then this period might be increased to say 12 months after you gain more confidence.

How often do good multimeters (such as Fluke 289 or 87 V) need to be sent in for calibration?

That entirely depends on what your procedures say!

Dave.

[Dane Bear]

Good info. Thanks guys!

[Conrad Hoffman]

I have calibration sources readily at hand and can check my meters whenever I want. For meters that don't get banged around, my guess is "dang near forever". I bought my Fluke 77 new when they first came out, probably 20 years ago. It has never required the slightest adjustment and is spot on. My old HP 3455A stays within a few PPM pretty much forever. Ditto the HP 3478. OTOH, I've had older Keithleys that needed periodic touch up and whenever something fails and needs repair it also needs to be checked. I doubt the typical hobbyist would ever know or care if the reading was ever so slightly out of spec.

[G7PSK]

If you get your own voltage/amps standard how can you be sure that it is true. For that matter how can one be sure any thing is true and accurate.

[cybergibbons]

Calibration isn't about certainty, it's about the confidence that a device can measure properly.

[Rerouter]

how can one be sure any thing is true and accurate.

If you follow the trail back to your national standards, they have very extensive experiments to confirm what they know to be the standard volt and amp, as for an ebay standard, its similar to buying a second multimeter, it only tells you that either one is getting out of calibration at a faster rate, or at the same rate, though the degree with which this occurs is variable,

[saturation]

That question is the basis of metrology: while every device calibrated can be traced to something else, in this case a standard or reference, who does the standard refer too?  The only thing one can say for references is that the probability of an error in measurement can be specified as ever decreasing values, but is not zero, and is dependent on time.  This is obtained by continuously measuring using a device and tracking all those measurements with a log.  Likewise, the item being measured can also vary over the same period.  This creates a historical time line, the longer the better.  As this history is built, a probability can continuously be calculated, and if the historical device measures its values true over time, over weather, over aging, and other non-controlled variables, then one can say within X years, months or days, the likelihood it will read in error is in ppm or ppb etc., over that period.  When a device is designed and tested, logs are made, and whenever the largest change in accuracy occurs over a population of devices in the same model line over time occurs, that is roughly the period specified as the calibration interval.

Back to DMMs, by using your DMM, even without a log, you get a similar feel on how true and accurate it is over time.  This tend ties to the model, and maker, and how a reputation for confidence is built.  You also pay for this reputation and confidence; many DMMs can easily meet or beat a Fluke 87V specs, but do they have this type of history too?

Personally, as per Conrad, I have Flukes dating back to late 1980s, the first model 85s.  I have 4 in that series.  All have not required calibration or adjustment.  The only time I've ever seen a Fluke require some form of calibration is after a disaster, a fall off a roof not just a table and onto concrete, or if the 10A fuse blows during an accident.  Its as much a check the meter is working, as it is to insure it measures true.

If you get your own voltage/amps standard how can you be sure that it is true. For that matter how can one be sure any thing is true and accurate.

[Shock]

How often do good multimeters (such as Fluke 289 or 87 V) need to be sent in for calibration?

When your companies standards specify.
When your instrument has fallen outside the tolerance when compared to a reference measurement or (calibrated) instrument.
Those two are give ins.

There is calibration but there is also certification.  You "could" have your instrument calibrated (or matched as close as possible) to any old instrument.

For calibration certification you would be taking it to a professional that has their calibration instrumentation (normally but not always) calibrated regularly to a higher tolerance.  They use this as a benchmark to see if your instrument is drifted or defective.  When they calibrate your instrument they should be following a documented standard/process and provide you a certificate of the results (which could be as simple as in spec).  The testing procedure may include to a certain temp/humidity/warmup time and anything else their standard defines.

The certificate is two fold first as posted previously, confidence in the instrument.  An independent tester offers high confidence as they are trained in calibration and pay big money to have their testing equipment as accurate as possible. The other is if your instrument falls outside its accuracy and this impacts your business or a customer you're some what covered as you can show the steps you had in place to prevent it.  Some customers may even request your calibration process.

So its for accuracy, peace of mind and well to cover your ass (blame the instrument not the engineer).

As Dave mentioned equipment can fall out of calibration overnight.  This is why in some labs where accuracy is critical calibration checks may be required before and after the equipment is used.  Not to be confused with certification however.

Aside from the obvious benefits standards and certification is also good marketing spin.
For me if I was in business yearly certification would be the absolute minimum.  More frequently if the impact warranted it.
For hobbyist I'm satisfied as a calibration leech.  I'll seek someone out who had their gear recently certified and do a comparison.

Make sure you follow your and state/govt regulations, trade practices, customer requirements along with your guilty conscience and you cant go wrong.

Source: Used to be calibration tech.

[G7PSK]

The thing is if you follow the trail all the way back to its origins how can you be sure that is correct, my point is you can only take it on trust that better brains than yours go it right, there is only two things in life that are certain taxes and death.

[jucole]

I'm going to take my Solartron 7150 bench meter for calibration this Friday!  I can't wait!

[Shock]

The thing is if you follow the trail all the way back to its origins how can you be sure that is correct, my point is you can only take it on trust that better brains than yours got it right, there is only two things in life that are certain taxes and death.

Sure if you get a backyard calibration its not worth the sticker its written on.
But normally its a couple of steps back to a government testing laboratory.
This of course depends what country you are in.  But they should be a certified lab and have paperwork. 

There is the human factor and labs tend to have a reference they can use to check their own calibration.

If your keen you could always get into the chemistry side of it and make your own Weston Cell. 

[ecat]

The thing is if you follow the trail all the way back to its origins how can you be sure that is correct

Science*

my point is you can only take it on trust that better brains than yours go it right

Which is why we have Science*

there is only two things in life that are certain taxes

Ah, now, taxes are a creation of the human mind as is the society in which they exist...
Let me double check...
<double checks>
Nope, no certainty there. Consensus, sure - all be it on threat of incarceration or death, but no certainty.

and death.

Only assuming death is the default state and life a statistically insignificant blip... You're probably right on this one


*Science is empirical. Knowing the answer means nothing, testing your knowledge means everything (Lawrence Krauss).

I'm going to take my Solartron 7150 bench meter for calibration this Friday!  I can't wait!

Oh, let me know how you get on. I have a 7150 Plus here and maybe someday I'll splash the cash.

[SeanB]

That is why I had a scale teardown. I check them, and this one had developed a slight nonlinearity at the 10kg point, where it was more than 50g out. At 5 and full scale it was fine, just at this point it was going nonlinear. As it was obsoleted 10 years before it was replaced with one that will hopefully be obsolete only in the future. At least the new one does not need a plugged in dongle to get it into calibration mode, you just use the right magic keypresses and then you can run through the setup with only needing the right masspieces.

[G7PSK]

The point I am trying to make is you cannot be sure your measurement is absolute it never can be. The only thing that is really important is consistent reproducible measurement, it really does not matter if your meter reads potatoes or moths but it has got to reproduce the same reading on a consistent basis. I have seen some people on other video blogs tweak their meter to comply with a so called reference in either just one or two spots. What is there to guarantee a consistency across the whole range.
Science is not a god in its own right and there is no certainty that what is right today or has been right for the past hundred years will be right tomorrow. I do actually have a Weston standard cell some where which I got at auction of ex Cambridge university equipment many years ago unfortunately it either does not work or does not supply enough current to work an AVO which was the only meter I had at the time.       

[Conrad Hoffman]

Naturally there are tolerances on everything, but you can be quite certain within those tolerances. What they call "maintenance of the volt" is actually a fascinating topic. I maintain 3 Fluke voltage standards. The three are inter-compared on a regular basis.

Every so often a single one is sent to a lab for NIST traceable calibration. Actually I only have them read it, never adjust it, because history is very important in this endeavor. Before it goes out I know the relationship between the three standards. After it comes back I inter-compare again. That tells me if anything has changed in transport. If not, I know the value of the calibrated unit, and thus the value of the other two. Now we motor along doing the occasional inter-comparison, confident that the voltage is, in my case, good to a few PPM, until it's time to send a unit out again.

It turns out that one can stay withing very close limits for many years with this method, but for anything requiring paperwork you still send one out every year. BTW, the whole thing relies on being able to compare two voltages with extreme accuracy, but this isn't very difficult.

Weston cells are nothing but trouble! They don't have the same voltage, depending on the purity of the chemicals, and now you have the problem of perfect temperature control. Note that the more common un-saturated cells are more temperature stable, but have a limited life. IMO, all the common unsaturated cells made by Eppley are junk now. Environmental regulations put them out of that business more than 15-20 years ago, and those cells all go unstable in that time. A few years ago I took all my standard cells to the yearly hazardous waste collection event and said good riddance. Even though it's against eBay policy, you see them there quite often. Don't buy because they're all useless.

[SeanB]

Weston cells are good to 3 decimals, but time and technology has overtaken them, now a cheap reference can outperform them They were the duck's guts a century ago.

[jucole]

I'm going to take my Solartron 7150 bench meter for calibration this Friday!  I can't wait!

Oh, let me know how you get on. I have a 7150 Plus here and maybe someday I'll splash the cash.

The results for my 7150

[chrome]

You sure you have a 7150? mine doesn't have a 10kR range (or 100k, 1M or 10M).

[jucole]

You sure you have a 7150? mine doesn't have a 10kR range (or 100k, 1M or 10M).

Maybe it's the 0S1L model and I've just got it upside down ;-)

[Spawn]

About calibrating, I got around 25 multimeters : 11 Flukes some of them are bench type, 4 Gossen Metrawatts, 2 Voltcrafts, 2 Philips amm’s, a Thrulby bench mm and some pocket dmm’s.

 If I send them out for calibration I would be poor by now, but the plus side is I can compare them together, which I do quite often, every 3 months I check all the batteries in my multimeters and do some measurements on same source at current voltage and resistance. I don’t have any calibration reference but I am assuming (maybe I am wrong with assuming but) if the meters show same values they are good and I am happy with it.

[ttp]

About calibrating, I got around 25 multimeters : 11 Flukes some of them are bench type, 4 Gossen Metrawatts, 2 Voltcrafts, 2 Philips amm’s, a Thrulby bench mm and some pocket dmm’s.

 If I send them out for calibration I would be poor by now, but the plus side is I can compare them together, which I do quite often, every 3 months I check all the batteries in my multimeters and do some measurements on same source at current voltage and resistance. I don’t have any calibration reference but I am assuming (maybe I am wrong with assuming but) if the meters show same values they are good and I am happy with it.

You could choose one of your better multimeters and calibrate just this one then use it to check the rest. Cheaper than calibrating all 25 of them.