Nebie question about counts and acuracy

[001]

Hi! Sorry for stupid question

But what the fuck with Uni-T DMMs?

UT61E is 22000 counts and UT71D is 40000 counts
And their basic acuracy is only 0.1%
It measns what I can`t  trust all five digits, isn`t it? I think that few digits unusable in this DMM.

Is Uni-T high counts for marketing trick only or I misuderstood something?

Thanx!

[MosherIV]

Hi

Try Dave's explaination:

[VintageNut]

Basic accuracy (uncertainty) is worst-case for every production model.

If you want to know the best uncertainty, then you have to do some work or have a cal lab do some work for you.

Uncertainty is comprised of an offset and noise. You can, in many cases, characterize the offset and noise and then take steps to try to reduce uncertainty.

Offset is relatively easy to eliminate once you know what the offset is.

As a first step, try to characterize the offset for 100% of range, 50% of range and for zero.

The offset at zero is the easiest to characterize for voltage measure and for current measure.

For voltage measure, short the input and record the reading.

For current measure, open the input and record the reading.

Some DMMs have a REL function. You can use the REL function to eliminate the offset after you measure it.

[001]

Thank You!

But I`m not understand english speakers. Can I find text version?

[IconicPCB]

Dave does not speak English...
Dave speaks Strine..excited Strine

[BFX]

Dave does not speak English...
Dave speaks Strine..excited Strine

I'm not native speaker and my English is poor from my point of view but I like Dave's English "Bob's your uncle"   

[martinr33]

The digits are not useless if you are looking to measure the change in something. The absolute accuracy doesn't make too much difference in that case. These meters are good at spotting sub-microvolt changes in 10V signals.

For example, HP's 3458a is specified at 8 ppm per year. If one drifts that far, that's 80uV, or almost all of the bottom three digits. That does mean that you get great accuracy at 5.5 digits, across time and temperature.

On the other hand, if you are measuring the noise on a 10V precision reference, you can get that even if you are a few microvolts off on the average reading.

Where the argument falls apart is when the noise in digit 9 in the meter is significant. Then, it doesn't matter what you do because there's no useful information in the noise.

[David Hess]

Many meters have a number of counts out of proportional to their accuracy however there are some advantages.  More counts allow using a more sensitive range for some measurements reducing (often dividing by 10) the "+ digits" component of the error.  This also produces higher resolution for these measurements and less range switching.  But the greater number of counts can also be misleading by implying a higher level of accuracy than is actually present.