Multimeter Accuracy

[MWS]

Is there a definitive way to check accuracy of a multimeter? For example, my car battery is showing 13.26v and I checked a AA (1.5v) and showed 1.72v. I am asking because I recently exchanged an old (very dead battery) that I showed at only 1.87v, and the store tested at 4.32v. I borrowed two meters from my neighbor to check the battery that I read at 13.26v, and the other 2 meters showed 12.87v and 12.14v. Just to clarify, the battery in question was under no load and disconnected. Also, on additional readings, each meter showed approximately the same as previous reading (+/- .02v).

I know I may just be picky, but numbers mean things, especially if the reading is showing to 2 decimal points.

Thanks in advance!

[Burg Boxster]

Accuracy is always more important than precision... What good is having two precise decimal points when neither is accurate

Yes, you can check and often times calibrate the accuracy of a multimeter (usually not cheap throw away ones though...). To do so though you need a known power source with measurable reliability...

Good luck

[thstone]

First, check the spec sheet from the manufacturer of your multimeter. This will contain the specified accuracy in +/- X.xx%.

If you assume that the multimeter is working properly, this is all you need to calculate the error range.

For example, if the multimeter is rated at +/-3% accuracy, then a 12.00 vdc battery could read anywhere between 11.64 vdc and 12.36 vdc.

This could roughly explain the range of values that you're seeing from different multimeters.

[thstone]

If you actually want to test the accuracy, you'll need a verified voltage source like a DC power supply which also has a manufacturer's accuracy.

Set the DC power supply to 12.00 vdc and take several readings using your multimeter. Then work the numbers.

For example, let assume that the DC power supply is rated at +/- 2% or +/- 0.24 vdc at 12vdc (2%-3% error is typical of a $300 power supply). This means that the actual voltage of the power supply could be anywhere between 11.76 vdc and 12.24 vdc.

Now, add on the error in the multimeter. Let's assume that the multimeter is also accurate to 2%. This means that the meter could read as low as 11.52 vdc or as high as 12.48 vdc.

Thus, in the example above, the total uncertainty of the measurement is on the order of 1 vdc. This doesn't mean that affordable multimeters aren't useful, it just means that they aren't accurate enough for the measurements that you're tying to take.

Affordable instruments are still useful in a lot of practical applications like determining whether the battery voltage is 12 - 13 vdc or 8-9 vdc. Its just not going to tell you conclusively whether the voltage is exactly 12.50 vdc.

Of course, better instruments have lower errors. For example, if you spend $1,500 for a multimeter with a 0.25% error, you'd be able to read the 12.00 vdc to within a range of 11.97 - 12.03 vdc or an uncertainty of 0.06 vdc. This would be highly accurate.

[RobertF]

http://support.fluke.com/calibration-sales/Download/Asset/2547797_6200_ENG_A_W.PDF

[coreseller]

I own a pretty decent Fluke and last fall I had a car battery crap out so I decided to do a comparison. I read the battery with the Fluke and then tried several of the "Freebie" Harbor Freight (https://www.harborfreight.com/7-function-multimeter-98025.html) ones that I have accumulated over the years for comparison.

To my amazement they all were within 2 to 3% of each other's readings.

As an aside, the only real way to accurately test a car battery is with a load tester.

[MWS]

Grrr. I guess I should have titled this thread "Do you trust your multimeter? - One of ten questions that you should never ask yourself. (Also on list, "Do you trust your wife?")"

What is really driving me mad is that given manufactured specification accuracy, standard deviation and calibration degradation, the displayed output could be +/- volts, however in that range is the ACTUAL voltage, and given my readings are always just about the same, it is entirely possible that I am in possession of the "atomic clock standard" of multimeters or, just as likely, a piece of crap....and I WILL NEVER KNOW. Sigh...

If anyone out there knows of something I might have around the house that has an ABSOLUTE constant known voltage, I would love to know... Would allow me to move on to one of the other 10 questions I should never ask.

[jcslocum]

Take it to a local calibration shop and have it set and calibrated. Any business in the area that is ISO rated has to have it done for all measuring devices. I'm certain there are shops you can send it to as well.

Calibration Services

[coreseller]

A new thing out there......Google

Youtube Also, They can be Helpful if actually Utilized:

https://www.youtube.com/watch?v=nDHA7cg-yoU

https://www.youtube.com/watch?v=-izNaaLWG1k

[MWS]

Well here is what I ended up doing... To find a known voltage value, I went to Duracell's web site and found the technical product information section https://www.duracell.com/en-us/techlibrary/product-technical-data-sheets/ and selected the Quantum AA battery and found the voltage to be listed at about 1.65v; knowing this, I purchased a multipack and selected 10 for testing and averaged the data. I found my meter showed apx 1.72v, and now with a benchmark, all I needed to do was adjust the internal potentiometer. Additionally, having a known benchmark readily available (and cheap), I can quickly verify accuracy at anytime in the future.

Problem solved.

What I learned is don't assume your meter is right, check it.

[RobertF]

https://www.amazon.com/Management-Development-Precision-Reference-Breakout/dp/B00SK8P51S/ref=sr_1_4?ie=UTF8&qid=1500731658&sr=8-4&keywords=voltage+reference