Hi I Thought it'd be interesting to take a look at a typical three-and-a-half digit or six thousand count multimeter and see how accurate these things are. Out of the box by getting not just one of them, but by getting a whole bunch of them and just so happen to have a whole bunch of these new Eevblog meters here in the lab and before I should mount. I Thought I would just hook them up to my reference generator here and see how close they are on various ranges across a whole spread of multimeters. Grander.

They are all part of the same shipment. They're all manufactured at the same time on the same run. They were all calibrated at the same time, But that's the interesting point. How close can a typical factory calibration on a meter like this actually be? Is it the plus/minus one least significant digit? Is it half its spec? You know this is like 0.3 percent.

DC Volts are typical, which actually gives originally wide spec in terms of counts pluses in the number of art counts as well. I've done a video on this explaining that sort of thing and how close is it to its absolute? You know, into an absolute reference standard. So I Thought we'd get like 50 odd multimeters or so and actually measure them and oh, go through. I won't bore you with all the details, but I'll actually take as many multimeters as I've got here all from the same batch and actually measure them on various ranges see if we can get the data.

Could be interesting. Let's go. hey, check this out. look amazing.

Symmetrical multimeter stacking just like the Philadelphia Mass Turbulence of 1984 aren't believable. No human could stack multimeters like this. So if we actually have a look at the specs of a typical three and a half digit multimeter like this I say three and a half digits. It's actually a 6,000 count I've done a whole video on that which I might have to link in here if you haven't seen that.

That explains counts at resolution and counts and all that sort of stuff. Anyway, typically for the DC voltage range, here for the 60 millivolts to the 6 volt we're looking at not 0.3% plus two digits there. So that's actually at naught point. Three percent of the actual reading that'll to you over here on the electrical specs is given as plus minus a percentage of the reading plus number of digits.

So if we actually have a look at that, the meter could actually read and still be within spec. Let's say you had your measuring 5 volts on the 6 volt range. For example, just with the plus minus 0.3 percent accuracy figure, it could read anywhere from 4 point, 9, 8, 5 volts to 5.0 1 5 volts and still be within spec for a precise 5 volt input. And that's not taking into account the extra the plus 2 digits here which you have to add on to the least significant digit over here and so it could display anywhere between that figure.

But how close is it actually going to be? Not just on one meter because that really doesn't tell you much. this ineffectively doesn't tell you anything. You've only got one data point, but if we can get say 50 multimeters or something then hey, that can give us a good indication. I've never actually done this on a bulk lot of meters, but of course a medal like this is actually or should be actually calibrated at the factory.

It's actually software adjusted so they will feed in the the reference that voltage generated Hindi here and actually program the exact figure in there. Or that's what they should do in. back in the old days, they'd get their tongue at the right angle and tweak a little pot inside. But modern Modi's multimeters like this one don't have pots, they're all that software configurable so when you measure 5 volts, it should be spot-on in theory of course.

but the course this changes with the temperature and you know I age and other characteristics. So if we interesting to measure a whole bunch of these and see what we get. but of course as with any measurement system like this, even if it's bang on as I like to say and we feed in five point, Zero, Zero Zero zero volts which I'm capable of doing with my lab gear here. a precisely knowing value it could still be plus minus one digit here and you always expect that from the converter that's just inherent in the converter itself.

So you know if it's five Point Zero, zero One, for example that you wouldn't worry about that into you would say that they're all bang on. When you start talking point double O2 or point double O three or something like that then you can start saying yeah, it's starting to be different to the other meters or but still, it's definitely going to be within spec I Know that for sure. So what I'm going to use is my advantage are Sixty One Forty Two programmable voltage and current generator so we can do DC voltage and DC current. I Also have an AC voltage generator I have a resistance standard so I'll try and do as much as I won't do all the ranges.

Obviously this thing can only go up to like a you know 11 volts maximum I think it is so you know. Okay, I'm not going to test like the thousand volt range, but hey anyway, I won't go quite to full scales as six thousand count meter I'll drop it back by one and just being give a nice familiar voltage reading of 5 volts here. And yes, I've let it warmed up and I've confirmed it with my Keithley meter above it and other stuff. So no drama.

Let's plug this puppy in and see what we get. But I won't bore you with all the details of all the rest of them. bang 5.000 But as I said, there we go. Point: Double-a one right? not concerned.

plus minus one least significant digit. So I'm going to see how close they all are I pretty much expect them all to be within a couple of least significant digits because that's effectively what this R Plus digit on the end is saying they're You know, fairly confident that it's going to be the exact figure they calibrated it at plus two digits. Basically, so you know I expect it to be within that. it's 24 degrees know? sorry, 20 sorry, 22 degrees here in the lab and you know, would have been similar temperature to what they calibrated that.

So I expect them all to be pretty darn close like that. It could actually be a really incredibly boring result. sorry ahead of time. In fact, I think it will be.

Hmm, Well, that's a few multimeters 40 to be precise I Can't explain it, but there's something very therapeutic about doing this. Oh yeah, so I've got my handy banana plug lead you here I Go! Ah goodness the things I do Now let's repeat that same thing. but we thought: 50 millivolts I've changed to the millivolt range here and we are. You know 50 millivolts is quite low, so you know not all multimeters have a 50 millivolt range.

But hey, let's have a look. this one's bang on 50.00 Ah, 50.00 Geez. like in the millivolt range. Anyway, I Am recording all these.

Actually, the millivolt range so far is ridiculously ridiculously spot-on So I think what they must be doing is calibrating the millivolt range and relying on the resistor divider ah to get the other ranges all the way up. it's getting. It's got it dropped down and a little bit of excess charge there. So these are all spot-on 50.00 This is insane.

So yeah, this is just. this is crazy business. so that's what they must be doing. That's why we saw some variation on the readings before: $4.99 for nine, seven.

from basically for 4.99 seven up to 4.5 basically. so just some tolerance on the resistive divider. but these are all cheese. that's ridiculously bang on.

That's just crazy. Ah, oh, is that one one least significant digit out. Oh no. And now for the resistance standard.

I'm using my weak on Precision reference standard. If you have to ask the price, you can't afford it. So yeah, I don't think we're going to question the accuracy 9.99 Let's go now. of course.

one of the problems with a 10k standard on a 60 K range is that. Well, you know you only got two decimal places here, so you know I expect it to be like and it is. It's like 9 9, 9, 9, 9, 8. At worst they're all going to be up there.

We go: 10. Well that came in not too bad from 99 eight up to why 1000? Now it's time for current. So I'm going to use up 50 milliamps and well, this one's bang on 50. So let's uh, let's start there.

Here we go: Oh goodness. I'm going this way and then back that way for those who are playing along at home. $49.99 Here we go: $49.99 Oh boy. Well, that current is a very crazy tight spec.

The only one was that nine nine, nine, four, nine, nine, eight down here. So but only a I - I'm sorry. three least significant digit spread. they're very tight on current.

Okay, one last thing I'm going to check is our AC because that's not typically our tested. So I've actually got an AC voltage standard here in EDC just like my EDC DC reference standard up here, which is the best standard I had in the lab. I Didn't need to use that for today's purposes because we're not only looking at a point three percent class instrument. Anyway, this a sieve already standard I Forget the spec, but it's more than good enough.

I Set it for five point Zero Zero Zero Zero zero volts at 60 Hertz So that's I Believe smack bang in the middle of it's a most accurate range. so at 5 volts, let's give it a ball. Currently getting five Point Double O2 and I have actually confirmed it up here and I've given it a little bit of a tweak to be closer there. I've turned some by averaging on and I won't bore you with the time lapse of all these, but Dyke's probably expect enough.

Yeah, helps if I forget to go select all the ranges. I'm not going to use the the VFD which is the variable frequency drive which is the key which is the filter in there. As you can see, it doesn't get the resolution so hey, no one's bang on. So white boy with another time-lapse Well I was actually quite surprised by the AC spec here.

It was really tight. I mean the highest was five double O three up there and there, but most of them were pretty almost bang on. You know, five double O one or five thousand spot-on And the spec by the way, for AC range fifty to sixty Hertz is the tightest response. It's a nought point seven percent plus three digits, so it's easily well, it's within.

Well actually yes it's within no four digits within the four digits, let alone the point seven percent. so balls in it. So there you go. That's the final data table for this thing.

Just some spot checking on various ranges here and it is. As you can see, it was well within the spec of this thing which was our by the way of their resistance some spec. There was typically not point three percent there we got. yet actually on the 10k range it up to 0.5 percent plus three digits.

We can see. you know it did pretty damn well because they calibrate these things may not necessarily on every range as I said. I Actually haven't asked our broman what they actually how they actually calibrate this thing but as you can see, the the best range by far was the 50 millivolt one and also it would be the same for the 500 millivolt one two. I would think because what they're doing is they're calibrating it on the millivolt scale and that's very typical and then actually relying on the resistor divider to do the rest.

So that's why you know when start getting out to the 5 volt range early. a little bit out. I was very impressed by the current actually that was that really spread on that was really good and the AC very surprising as well. I Would have expected a larger spread on the AC but didn't see it at all.

so I'm actually quite impressed by this. but this is typical of even you know relatively low cost meters like this one because they do have software art calibration and I'm too lazy to go in there and you know, put this data into a spreadsheet and you know, maybe get some been in and things like that it there's not enough spread in there really to get a huge amount as useful data. All we want is the min max spread on each one. You can calculate the percentage.

It's much better than the specs. That's all we care about. So they can actually software calibrate these things I believe. but I'm not sure they're exact our procedure: I Can't find a cow like a some meters have like a cowl menu.

When you write, you know you hold a button down and turn them on for example. So I'm not sure the procedure for this one or whether or not they're just relying upon a the voltage reference which I believes inside the chipset for this one and the and the precision of the resistor networks that they plug into these things. whether or not they do individual cow on each one I Don't know. as like manufacturers and different manufacturers going to be different, different models are going to be different.

all that sort of stuff on how they calibrate them. Actually, that wasn't a really fair test with the 10k resistor here because it was. as I mentioned, it was much lower down in the range so the error becomes a greater percentage because purely from the ADC count the resolution. So what I've got is my IET Decade resistance box here.

It's not a standard as such, but I'm able to select 50 K so we can use the same order as before and Bam I've tweaked it until it's good enough. so I'll repeat that and there we go. We actually got a bigger spread on the actual number of digits we got. not a differential of nine between the highest and the lowest number nine least significant digits, but because it's much higher in the range.

When you calculate the percentage, it's a lower percentage or, ie, you know, a tighter tolerance than what we got for the 10k one here. purely because we're closer to full scale on the 60 K range than we were with the 10k resistor also on the same 60 K range. and just for kicks for all your Unity fanboys out there, I'll just try myut 61 E and hmm. forty-nine Point eight for K and not that terrific is it Is that within spec? Hmm no point three, two percent.

Hmm yeah. I think it? I Actually, it probably is just within spec is it? And it's saying 50 millivolt range will bang on here with before whoo I might have drifted by point one milli volts. Um, yeah, we're a little bit low there, but still within spec I believe. But yeah, significantly different.

Anyway, a sample size of one and I would show you my fluke 17 B but I kid you not, it has failed. and no, it's not the contact on the banana plugs I've tried the probes which come with it. Short it out. it does nothing.

what the. and it doesn't even read the millivolts either. It's absolutely useless. it's dead, maybe a repair video And just for kicks, there's the older.

Bremen BM Two, Five Seven. Oh look, Alpha B's D cup. And of course, the venerable Fluke 87 is bang on well. I'm telling you what.

here's another unity and hmm. not too good at all. But as you can see well within the quoted spec and that's not unusual these days. It was different back in the day with the tweaking the pots in the things.

But you know, usually even your cheaper multimeters are pretty good these days. So anyway, hope you found that interesting. I Just thought I had these meters I've never really had a chance to do this, so hey, why the hell not? Anyway, if you enjoyed it, please give it a big thumbs up. Discuss it down below all that sort of jazz you want to support me.

Patreon links down below: I've got a new Ee-vie blog newsletter you can sign up to. By the way, if you wanted to know about this meter, people on the newsletter and Patreon and supporters, they actually found about this and it's some currently practically sold out. So sorry there's not going to be another batch until April Is it Yes. Hmm.

Catch you next time you.