DMM: Measuring AC Volts

[nadona]

On a DMM, there is an OFFSET or NULL button.
I know this button is needed for measuring DC, but is it needed for AC as well? I can't find it in the manual.

[Wytnucls]

I don't know what meter you have, but on my DMMs, there is a 'REL' button (relative) which is mostly used to null the impedance and capacitance of the test leads for low value DUT measurements.
It shouldn't be required for voltage or current measurements (AC or DC).

[nadona]

I always use OFFSET(Fluke), REL(Keithley), and NULL(HP) after shorting out the leads to measure DC, but I haven't used the AC function other than to measure the main, and this was with a  hand held.
When I measure DC Volt, I short both of the leads and press the OFFSET button for accuracy. When I measure the 300 Volt in a freshly NIST Calibrated HP 3457A, it displays 0.29 volt whether I short the leads or not.
If I NULL this when I measure 40V AC, its effect is huge.
Anyways, the calibrator said he doesn't use it. And the manual does not go into specifics of how or when to use this function.

[EDIT] Even though manuals say "The OFFSET feature may be used in all functions." - Fluke 8840A manual page 2-10 &  "You can use the NULL operation on any measurement
function (DC voltage, AC current, and so on)." - HP3457A manual page 2-25. Obviously it shouldn't be used in case of AC because the OFFSET's purpose was to take out the leads's effect.  In case of AC V or I, shorting doesn't shrink the number. The number has something to do with inside the meter. For example offsetting 0.3V to measure 32.5V is huge.

[alm]

No, you should not null the shorted reading with a true RMS meter. A true RMS converter is non-linear near 0 (which is why ACV is usually only specified from 5-100% of the range). It's designed to be linear within the 5-100% range. If you correct the reading for the shorted leads, it will be off for the higher part of the range. See this Keithley document why this offset is insignificant, and how using REL affects your measurements.

[nadona]

Thank you for the new material to think.
In case of my HP3457A, the offset is 300 counts. So according to the document it will display 0.10005.
The paper says most DMM but I think most Keithley DMM.

One more thing to think is the 100 count in the paper. Is that number happened because they connected the leads? My HP3457 's number is there whenever I choose AC V or I and doesn't change when I put the leads into the terminal and shorted.

Any HP3457A owners to see what number they got at AC V or I setting? my HP3456A has 150 count, my Fluke 8840A and 8842A have 40 to 50 count.

[alm]

The paper says most DMM but I think most Keithley DMM.

Which part of this do you think is not applicable to other brand DMMs? I think this is applicable to all DMMs in this class.

In case of my HP3457A, the offset is 300 counts. So according to the document it will display 0.10005.

The 3V range is specified from 10% of the range, or from 0.3 V as 0.13% + 116 counts in 5.5 digit mode. To meet its 90 days spec, the reading for 0.3 V should be between 0.29845 V and 0.30155 V. The error induced by the offset would be 0.300015 V, so well within spec.

One more thing to think is the 100 count in the paper. Is that number happened because they connected the leads? My HP3457 's number is there whenever I choose AC V or I and doesn't change when I put the leads into the terminal and shorted.

The open circuit value is meaningless and may depend on input impedance, leakage and EMI. They measured this with shorted leads.

[nadona]

1. I think we're misunderstanding each other. I meant that the 100 count doesn't apply to other brand DMMs. That's why I uploaded the photo. I'm not saying you are wrong, my reasoning is based off of the document itself.

2. At least we found out the manual is wrong. It should have said the offset feature shouldn't be used with AC. The technical writer and the design engineer are different people, so that's a possible reason for the error.

3. If the offset is big for all 3457As and not just mine, then this seems like a serious issue. In comparison to Fluke DMMs, the offset for HP is too big to be overlooked for decades.

[alm]

The paper doesn't claim the 100 counts applies to most DMMs, only to the one particular model they use as an example. The offset feature can be used with AC, you just shouldn't use it to null the reading with shorted leads. Why are you so worried about an offset that will cause at most 15 counts of error over the specified range? It's allowed more than ten times that error, even if it's been freshly calibrated.