Fluke accuracy ±(2%+2) - What means the +2?

[2X]

Hello,
does anyone knows what means the +2 on 2%? If we meausure a known voltage 100V then the multimeter will show a value between 88V-102V.  At the example on the below link says 97.8 V to 102.2 V, so I suppuse the +2 is equals with 0.2 but how this arises? If the +2 means 0.2 why the didn't write just ±2.2%? I haven't see any similar declaration on other multemeter.


https://www.fluke.com/en-us/learn/blog/digital-multimeters/accuracy-precision

[voltsandjolts]

+2 of the rightmost digit on the display

[BeBuLamar]

So if the reading is 50V then the actual voltage can be 48.8 to 51.2 that is +/- 2% plus 2 more counts of the last digit. So if it reads 1V then it's like 0.8 to 1.2V. In this case the 2% is less than the 2 counts error.

[voltsandjolts]

The OP hasn't specified the actual DMM but
±(2%+2)
likely means
±(2% of reading + 2 least significant display digits)

[HKJ]

Hello,
does anyone knows what means the +2 on 2%? If we meausure a known voltage 100V then the multimeter will show a value between 88V-102V.  At the example on the below link says 97.8 V to 102.2 V, so I suppuse the +2 is equals with 0.2 but how this arises? If the +2 means 0.2 why the didn't write just ±2.2%? I haven't see any similar declaration on other multemeter.


https://www.fluke.com/en-us/learn/blog/digital-multimeters/accuracy-precision

Nearly all multimeters uses that way to specify tolerance, it is necessary.
An example: Multimeter in 100.0V range and display shows 0.1V with shorted input, what is the tolerance?
It is way above 1000%, if you only specify in %, but by saying +2, it is covered in the specifications, without affecting the % tolerance.

If have written a bit about multimeter tolerances here: https://lygte-info.dk/info/DMMTolerances%20UK.html

[alm]

HKJ explains it well. Another way to express it is to say the uncertainty is 2% of the value being read + 0.2% of the range that is used. I think it's intuitive that even if the 1V range and 1000V range are both 2%, then measuring 1V on the 1V range is going to have a better uncertainty than measuring 1V on the 1000V range. I'd say a multimeter that does not specify an offset like +2 least significant digits or + 0.2% of range should not be trusted for accuracy, because this implies that 0V should be measured with zero uncertainty.

[2X]

Hello,
does anyone knows what means the +2 on 2%? If we meausure a known voltage 100V then the multimeter will show a value between 88V-102V.  At the example on the below link says 97.8 V to 102.2 V, so I suppuse the +2 is equals with 0.2 but how this arises? If the +2 means 0.2 why the didn't write just ±2.2%? I haven't see any similar declaration on other multemeter.


https://www.fluke.com/en-us/learn/blog/digital-multimeters/accuracy-precision

Nearly all multimeters uses that way to specify tolerance, it is necessary.
An example: Multimeter in 100.0V range and display shows 0.1V with shorted input, what is the tolerance?
It is way above 1000%, if you only specify in %, but by saying +2, it is covered in the specifications, without affecting the % tolerance.

If have written a bit about multimeter tolerances here: https://lygte-info.dk/info/DMMTolerances%20UK.html

Thanks for your reply. I read your article about the tolerances (very detailed with pictures and examples) but for some reason I have stack and I can't understand how mathematically translate/arises the +number.
The 0.5% of 2.000V is 0.01 but I can't understand how the +3 is translated to 0.007 in order to take the values of 1.987 to 2.017 volt that you write in your article.

Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13, i.e. the real value is somewhere between 2.000-0.013 to 2.000+0.013 -> 1.987 to 2.017 volt.
Display 2.00V in 60.00V range: The 0.7% is 1.4 and then the +3 for a total of 4 (rounded down), i.e. the real value is somewhere between 2.00-0.04 to 2.00+0.04 -> 1.96 to 2.04 volt.

If I understand right the +number is not about the deviation between the measurement value and displayed value but how the selected range of the multimeter has a tolerance independently if we measure a value or we just short the leads of the multimeter.

[HKJ]

Thanks for your reply. I read your article about the tolerances (very detailed with pictures and examples) but for some reason I have stack and I can't understand how mathematically translate/arises the +number.
The 0.5% of 2.000V is 0.01 but I can't understand how the +3 is translated to 0.007 in order to take the values of 1.987 to 2.017 volt that you write in your article.

Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13, i.e. the real value is somewhere between 2.000-0.013 to 2.000+0.013 -> 1.987 to 2.017 volt.

Yes, the 2.017 is a typo, as it says just before it: 2.000+0.013 and for some reason I make a typo with 2.017 instead of 2.013

If I understand right the +number is not about the deviation between the measurement value and displayed value but how the selected range of the multimeter has a tolerance independently if we measure a value or we just short the leads of the multimeter.

The +number is a offset, it can be due to errors around zero, sampling errors or rounding errors. The important part is that it is not really depend on the input value or the display value, i.e. it cannot be specified as a % of reading.
Note: AC ranges may have an extra error at small values in addition to this, this is typically specified as a minimum percent of range where the reading/tolerance is valid, i.e. anything below 5% or 10% of full range may be invalid (or at least out of tolerance).

[TimFox]

Go back to elementary analytic geometry (graphing).
If the relationship between the input voltage V and the display value Y is truly linear (ignoring the discrete values restriction), then its equation is
Y = m V + b
where m is the slope and b is the offset.
For that ideal case, there can still be errors in m, which can be specified in terms of fraction of reading, and b, which would be given in terms of Y, either in counts displayed or fraction of full scale.
Of course, in the practical case there are still unwanted variations about the best-fit straight line, which will increase the specification values.

[2X]

Thanks for your reply. I read your article about the tolerances (very detailed with pictures and examples) but for some reason I have stack and I can't understand how mathematically translate/arises the +number.
The 0.5% of 2.000V is 0.01 but I can't understand how the +3 is translated to 0.007 in order to take the values of 1.987 to 2.017 volt that you write in your article.

Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13, i.e. the real value is somewhere between 2.000-0.013 to 2.000+0.013 -> 1.987 to 2.017 volt.

Yes, the 2.017 is a typo, as it says just before it: 2.000+0.013 and for some reason I make a typo with 2.017 instead of 2.013

If I understand right the +number is not about the deviation between the measurement value and displayed value but how the selected range of the multimeter has a tolerance independently if we measure a value or we just short the leads of the multimeter.

The +number is a offset, it can be due to errors around zero, sampling errors or rounding errors. The important part is that it is not really depend on the input value or the display value, i.e. it cannot be specified as a % of reading.
Note: AC ranges may have an extra error at small values in addition to this, this is typically specified as a minimum percent of range where the reading/tolerance is valid, i.e. anything below 5% or 10% of full range may be invalid (or at least out of tolerance).

Maybe it is obvious but could you explain me what do you mean by "Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13" or "Display 2.00V in 60.00V range: The 0.7% is 1.4 and then the +3 for a total of 4 (rounded down)". How arises the 10 from 0.5 and 1.4 from 0.7? In fluke's example the "+2 number" equals to 0.2 value but in yours the "+3 number" equals to 0.003 value. This value of the +Number is where mathematically I can't understand in what value corresponds. I beleive is something simple but for some reason I can't catch it. Thanks for your replies.

[2X]

Go back to elementary analytic geometry (graphing).
If the relationship between the input voltage V and the display value Y is truly linear (ignoring the discrete values restriction), then its equation is
Y = m V + b
where m is the slope and b is the offset.
For that ideal case, there can still be errors in ^m, which can be specified in terms of fraction of reading, and b, which would be given in terms of Y, either in counts displayed or fraction of full scale.
Of course, in the practical case there are still unwanted variations about the best-fit straight line, which will increase the specification values.

Thanks for the nice example.

[HKJ]

Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13, i.e. the real value is somewhere between 2.000-0.013 to 2.000+0.013 -> 1.987 to 2.017 volt.

Maybe it is obvious but could you explain me what do you mean by "Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13" or "Display 2.00V in 60.00V range: The 0.7% is 1.4 and then the +3 for a total of 4 (rounded down)". How arises the 10 from 0.5 and 1.4 from 0.7? In fluke's example the "+2 number" equals to 0.2 value but in yours the "+3 number" equals to 0.003 value. This value of the +Number is where mathematically I can't understand in what value corresponds. I beleive is something simple but for some reason I can't catch it. Thanks for your replies.

2.000V in 6.000V means 2V in 6V range.
0.5% is 0.5% of 2000 that is 10 or 0.010V

The +number depends on the actual resolution on the display, it is always on the last digit. i.e. if the display is 2.000 then +3 is 0.003, if the display is 2.0000 then +3 is 0.0003

[J-R]

Side note, calibration data is more useful in my opinion than the published accuracy specs.  Many of my Fluke DMMs exceed their spec by a factor of 10 or more, so the fact that the published specs are worse than some other brand doesn't translate to reality in many cases.

[2X]

Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13, i.e. the real value is somewhere between 2.000-0.013 to 2.000+0.013 -> 1.987 to 2.017 volt.

Maybe it is obvious but could you explain me what do you mean by "Display 2.000V in 6.000V range: The 0.5% is 10 and then the +3 for a total of 13" or "Display 2.00V in 60.00V range: The 0.7% is 1.4 and then the +3 for a total of 4 (rounded down)". How arises the 10 from 0.5 and 1.4 from 0.7? In fluke's example the "+2 number" equals to 0.2 value but in yours the "+3 number" equals to 0.003 value. This value of the +Number is where mathematically I can't understand in what value corresponds. I beleive is something simple but for some reason I can't catch it. Thanks for your replies.

2.000V in 6.000V means 2V in 6V range.
0.5% is 0.5% of 2000 that is 10 or 0.010V

The +number depends on the actual resolution on the display, it is always on the last digit. i.e. if the display is 2.000 then +3 is 0.003, if the display is 2.0000 then +3 is 0.0003

This example make it very clear. Many thanks for your time and for your help.


So If theoretically we have a multimeter with specs ±(0%+4) in range of 60.00V & in range of 100.0V:

 if we measure a 10V in scale of 60.00V the offest will be 0.04 it will display a value between 9.96V - 10.04V
 if we measure a 30V in scale of 60.00V the offest will be 0.04 it will display a value between 29.96V - 30.04V

 if we measure a 10V in scale of 100.0V the offest will be 0.4 it will display a value between 9.6V - 10.4V
 if we measure a 30V in scale of 100.0V the offest will be 0.4 it will display a value between 29.6V - 30.4V

[HKJ]

So If theoretically we have a multimeter with specs ±(0%+4) in range of 60.00V & in range of 100.0V:

 if we measure a 10V in scale of 60.00V the offest will be 0.04 it will display a value between 9.96V - 10.04V
 if we measure a 30V in scale of 60.00V the offest will be 0.04 it will display a value between 29.96V - 30.04V

 if we measure a 10V in scale of 100.0V the offest will be 0.4 it will display a value between 9.6V - 10.4V
 if we measure a 30V in scale of 100.0V the offest will be 0.4 it will display a value between 29.6V - 30.4V

Correct.

A "fun" fact: I have seen a meter where the +number specification was worse than the % specification, even at full scale. I believe that this is sort of misleading. It was a ACV specification where the % was very good, but when adding the +number it was at best average (I do not remember the name of the meter).

[boggis the cat]

A "fun" fact: I have seen a meter where the +number specification was worse than the % specification, even at full scale. I believe that this is sort of misleading. It was a ACV specification where the % was very good, but when adding the +number it was at best average (I do not remember the name of the meter).

That isn't uncommon for the very cheap instruments.  Sometimes you will see a change from +/- counts to +/- a value in order to try to be extra squirrelly.  Higher priced instruments will occasionally have this crop up in the margins, where they are presenting accuracy data (sometimes only 'typical') for end users that insist on using equipment beyond its real capabilities.

Beware of cut-down specification comparisons that only quote the % values (or other ratio).

Then there is the entire discussion over how realistic some accuracy claims may be.  (In general, the bigger name brands even amongst the cheap offerings are a lot more reliable.  No-name stuff is a bit of a lottery.)

[2X]

A "fun" fact: I have seen a meter where the +number specification was worse than the % specification, even at full scale. I believe that this is sort of misleading. It was a ACV specification where the % was very good, but when adding the +number it was at best average (I do not remember the name of the meter).

That isn't uncommon for the very cheap instruments.  Sometimes you will see a change from +/- counts to +/- a value in order to try to be extra squirrelly.  Higher priced instruments will occasionally have this crop up in the margins, where they are presenting accuracy data (sometimes only 'typical') for end users that insist on using equipment beyond its real capabilities.

Beware of cut-down specification comparisons that only quote the % values (or other ratio).

Then there is the entire discussion over how realistic some accuracy claims may be.  (In general, the bigger name brands even amongst the cheap offerings are a lot more reliable.  No-name stuff is a bit of a lottery.)

With this "a change from +/- counts to +/- a value" you mean that they refer only the measurement scale for instance 0-1.999 0-19.98V, 0-199.7V and the don't refer 2000counts?

[boggis the cat]

With this "a change from +/- counts to +/- a value" you mean that they refer only the measurement scale for instance 0-1.999 0-19.98V, 0-199.7V and the don't refer 2000counts?

Specifications are generally given in a +/-( [ratio] + [floor] ) form; often with additional notes concerning 'edge conditions' (such as minimum valid reading per range, any over-range capabilities, etc.).

Sometimes a floor given in counts may look bad.  For example given a range of 100.000 V, '0.5% + 2000 counts' may look worse than '0.5% + 2 V' at a glance.  You may see the % component and not do the mental arithmetic for the 2 V while the 2000 counts stands out (due to being a number of digits).

Ideally a specification should be as clear as possible, but if your product has a poor specification you may be tempted to try to obscure this.  (Outright lying is not that common, but don't be surprised if a cheap instrument has an accuracy that sits very close to the maximum calculated specification, whereas an expensive instrument will typically be around 30% to 40% of this or less.)