-
Let's say that I have a freshly calibrated 34401A with a complete calibration certificate showing readings for 90% of full scale of all ranges.
The calibration lab, while measuring a 10k resistor (assume it's 10k precisely, it's all hypothetical) reported the reading on my dmm to be 10.000,86 ohms, with shorted inputs the reading is 0. The reading is within the 1 year accuracy of 0.01%
With this information, I can extrapolate a simple line equation (y = 1.000086x) and use it to adjust all my readings to a more accurate result which takes in consideration the calibration data.
So if i want to measure an unknown 5k resistor and measure 5.024,77 ohms i can divide this reading by 1.000086 and get a more accurate 5024.33 ohms.
The question is: can I do this, at least for a short term period after calibration, or am I missing something?
Of course the assumption is that the ADC is linear, and in integrating ADC usually it's very linear, otherwise single or two points calibration would have little meaning -
Yes, these are faithful reasoning. A good multimeter will calculate the amendment when you tell him what is actually equal to this artifact.
You must take into account in the calculations of uncertainty not only the uncertainty of the device (0.01%) but also the calibration uncertainty with which you measured 10.000.86 OHMS. -
Yes, these are faithful reasoning. A good multimeter will calculate the amendment when you tell him what is actually equal to this artifact.
Thank you!
Now i have another question, which may be very silly.
If I recieve the DMM right after calibration and measure a very stable resistor like a 10k 0.01% vishay bulk foil, can I use the calibration data and the reading for that resistor to actually adjust the DMM?
Say, i try to measure the Vishay resistor and the reading is 9.999,96. According to what i said before I can adjust this value by diving it for the coefficient I found through the calibration report, so the new accurate value will be 9.999,10 ohms.
Now I have an accurate value and I know the dmm hasn't drifted because I am using it right after calibration.
I could now adjust the DMM telling it that the value it should read is 9.999,10 ohms, so i now have an adjusted dmm with maximum accuracy and I don't need to adjust all the readings manually.
Of course we must take in consideration the uncertainty of the calibration but for now just assume that the calibration uncertainty is much lower than that of the DMM, right now i just want to understand if my reasoning is correct.
I'm asking all of this because here in Italy it's literally impossible finding someone who calibrates intruments for hobbysts, let alone adjust them, so i'm really desperate for some method to do this all by myself.
I may only have a 6 1/2 DMM for now, but I crave accuracy, and already saving for a 7 1/2 DMM -
I stopped understanding you ... was a resistor 10.000.86 and now it is already 9.999.10.
The idea that you describe is called artifact calibration. Implemented for example in HP3458A.
But you must be very attentive. The resistance of the resistor during calibration is measured for a specific temperature. You must work at the same temperature. -
I know how the 3458A is calibrated. But what I am describing is not quite the same thing.
For the 3458A you use a known resistor value (so a calibrated resistor) and tell the DMM what exactly it should read.
What I am describing here is the use of an unknown resistor measured with the same DMM you want to adjust.
You take the unknown resistor, measure it, adjust the measurement through the data on the DMM calibration report and then use this new value to adjust the DMM telling it that the value it should read is the adjusted one.
It may be a bit convoluted, but I don't think it's so hard to understand.
Obviously we are talking about 6 1/2 DMMs, not metrology grade instrument calibration and adjustment
-
You cannot increase the accuracy unless the value of the resistor is known. You can change the readings on the screen so that they show the value multiplied by the coefficient. But this will add an error.
But you don't need a resistor for this. The multimeter correction factor is calculated from the multimeter calibration data.
Try to write it in formulas and you will find that the resistance of the resistor is reduced. -
I feel we are talking about two different things here.
When i say calibration, i refer to the act of comparing the reading of the DMM to a known value and writing down the difference, nothing more.
When i say adjustment, i mean changing the internal coefficient of the DMM to make it more accurate.
When I send a multimeter to a calibration lab, the calibration consists on writing down the readings of the DMM and their differences with respect to a known value. They do not perform any adjustment.
HOWEVER, since now we know what the DMM reads and what it should read and we are doing this little experiment right after calibration (so no significant drift has occurred), we should be able to perform adjustment at home using an unknown STABLE reference like i described before -
Your desktop DMM comes with a specification, which is :
- PPM of reading
- PPM of range
For different time periods.
You take the PPM of reading, and you multiply it with your reading.
Same with the range.
If you recently calibrated your meter, you could use the 30 days or 90 days accuracy specification. If you have a transfer standard in your lab, typically close in value to your measured value, you can use that for the 24H specification.
Otherwise it is not a simple extrapolation, because you don't know if the error is offset or gain, or nonlinearity or thermocouple effect or the dozen different error sources in a measurement. -
If we are talking about change the internal coefficient of the multimeter, your method will work but introduce an additional error.
Therefore, this coefficient is usually used in Excel if necessary
-
If I recieve the DMM right after calibration and measure a very stable resistor like a 10k 0.01% vishay bulk foil, can I use the calibration data and the reading for that resistor to actually adjust the DMM?
What? If you receive the DMM right after calibrating then you could use the report to adjust the DMM (gaining more accurate reading, but making its drift-history a bit more complicated). What is that resistor for? It's nice that it's very stable (how determined?), but that's not going to help, if the true value isn't known. If it were known, then you could do your own calibration (and depending on how well the value of the resister is known and how well you can measure it, you could achieve lower uncertainty, i.e. a better calibration then the lab. I mean hypothetically speaking).
[..] -
What was the uncertainty of the lab that calibrated your unit?
I'm going to take a stab at the 1yr spec uncertainty you have (excluding the imported from the cal lab) is about 115 μΩ/Ω + 0.12 Ω any reading you do on a resistor you need to take this into account as any reading done using your meter could wander about that area. For a rough idea 24hr would be 23 μΩ/Ω + 0.06 Ω, 90 Days is 93 μΩ/Ω + 0.12 Ω.
You measured, 5.02477 Ω but there is a 95% probability that the true reading is between 5.02549 kΩ and 5.02405 kΩ which is also ±0.00072 kΩ
If you had a meter that was out by 0.5 kΩ and its a linear error then you might be able to adjust it out but once you are into the uncertainty/spec of the unit it is just not worth it.
What I suggest you do is have a few select resistors, measure them each month record the readings, if it starts to drift off then either the resistor or the meter is drifting off so its worth getting it recalibrated. But you may well find, weather, leads, and butterfly in Brazil will all affect your readings so you need to play and learn the meter and develop the methods you use to measure stuff, by then you will have a 8.5digit meter and you will be looking at ppb.