Accuracy of 3.5 digit multimeters

Wow, that will take some time to absorb but it's clearly very comprehensive!!

We're currently purchasing some more bench and handheld multimeters and so it will be fun to follow your post.

If you have time to compare the accuracy specs, you should be able to get an idea of what they are capable of. The fun comes when you start comparing values ... but admittedly, with 3.5 digit multimeters being so inaccurate, it's very easy to discount any error contribution from the 5.5 digit meter used as reference because the error magnitude is much greater than the 5.5 digit meter! But it also shows that we shouldn't necessarily be putting that much faith in the last digit (almost regardless of meter)!

This is a nice looking study and a good application for the Keithly. I enjoyed looking it over.

John

Nice work Gough

I've tended to stick with Fluke mainly because they differentiate between the ac and dc component of mixed waveforms much better than some others.

However if you're after a cheap meter to give you some readings then element14's Tenma range start from NZD$11.85.

Mark

Fluke is pretty much the gold standard - now I have an incentive to purchase one . That being said, I only decided to give it a test because I thought it was worthwhile knowing - once you know the limits, you know what's meaningful and what's not. And it puts the Keithley to another good use aside from testing power banks (many ... many ... Power Banks).

That being said, Tenma is a "rebadge" operation - remember the DSE Q1459 I have in my testing, well that was a Uni-Trend UT33D. As it turns out, this Tenma is too ... http://au.element14.com/tenma/72-7770/multimeter-digital-hand-held-1999/dp/7430566

So ... uh ... I'd have to be an advocate of looking for something with a proper error spec sheet, and probably something a bit more resolution (digits, premium cost) if you want to have more faith in your readings. I suppose half of it is knowing just how trustworthy your instruments are - gone are the days I'd try to differentiate 5.00 and 5.05v on those meters . You do get what you pay for, but I don't think it's necessarily value paying $50 upwards for a questionable 3.75 digit Tenma meter when for about three times more, you can have a basic Fluke. That's just my opinion anyhow.

I don't think it's necessarily value paying $50 upwards for a questionable 3.75 digit Tenma meter when for about three times more, you can have a basic Fluke

I'd support that view, and it was proven by a guy I worked with figuring out exactly that, some of the cheaper option flukes give better accuracy for not much more money.

While that is the upper end, paying peanuts for a meter that gives you anything with 5-10% accuracy is better than nothing.

Mark

I think this depends on if you have a 1% application or a0.1% or better application. I have 3 pretty cheap meters and don't have an expensive one. I did calibrate them some significant time ago and they are still a fraction of a % of one another.  They are great for testing that power rails aren't too vastly off  but I wouldn't use them to set a precision reference or anything .

Still any reason for a posh Fluke or similar!

I was just doing some gawking at the element14 catalogue - looks like the Flukes are 0.5% +/- 2 digits or thereabouts for their basic models, but some of the equivalent Keysight (ex. Agilent) are boasting 0.09% +/- 5 digits or thereabouts for the same price. So I suppose part of the whole thing with Fluke is indeed the "posh" brand. The improvement in accuracy is probably nice for setting up variable voltage regulators, checking for small voltage drops along low voltage DC wiring (some of their voltage tolerances are pretty small), but I agree. Many applications just need some indication - and old fashioned analog meters were lucky to be read back at less than 2% of error! Aren't we all spoiled!

- Gough

Your testing has shown some of the fineprint in those 'cheap' meters.

I think the heading was "Best Accuracy", but didn't say it was across the whole range

Tenma basic

Tenma advanced

Fluke Basic (113)

So it seems that 2% is not gone and we aren't spoiled at all.

mark

Very well spotted. I always assumed "Best Accuracy" was a translation error on Uni-Trend's behalf. I suppose, instead, they might mean that's where the bottom of the "curved" error percentage slope is guaranteed to hit instead, in which case it's a fairly useless spec. It's like saying that my kitchen scales have a best error of 0gm!

Alas, the Fluke 113 specs seem a little disappointing, but also, a lot more truthful. I suspect one must spend even more than that ...

At the basic end of Agilent Technologies/Keysight Technologies range is the U1231A, listed at AU$116 (http://au.element14.com/keysight-technologies/u1231a/multimeter-trms-6600-count-4-digit/dp/1903369):

So, it looks like it's a 0.5% + 2 digits for the "basic" voltage. Surprising the difference in spec between the Fluke and this one - but it seems that's where the 3.5 digit meters "wanted to be" but weren't.

Spending a little more, we might be able to improve the accuracy further - Agilent Technologies/Keysight Technologies U1241B, listed at AU$241 (http://au.element14.com/keysight-technologies/u1241b/multimeter-trms-10000-count-4-digit/dp/1780325):

Considering that both units are 4-digit, this one says it's 0.09% + 5 digits. So while its percentage error is much less, the error number of digits in the least significant digit is higher.

How would this affect a range, I wondered ... so I did a quick Excel calculation using the 1000mV range as an example:

It seems that ~75mV into the 1000mV scale, we are close to the break-even point where the increased error of 0.5% overtakes the +5 digits of uncertainty. So the more expensive meter is likely to show more errors for readings below ~75mV on the 1000mV scale (i.e. 0.75% scale deflection). I suppose this can be an issue if the ranges are strict and feature no overlap - so say measuring 1.002v on the 10v scale ...

Still rather fascinating and head-scratch inducing specs .