Hi. This is your humble multimeter and you're used to it just on Dc volts here. Reading zero volts when you're just got your probe sitting there on the bench, it's reading nothing and you might be, uh, familiar. Of course, if you switch it over to millivolts, you know it might pick up a little bit of noise there.

And if you put your hands on there, you know you might get a few like tens or even hundreds of millivolts noise. But generally it's around about, you know, zero volts something like that. Especially if you put it on the voltage range like this. Well, what happens if you take the same leads and you plug them into like, a high end six and a half digit or seven and a half digit multimeter like this.

Well, uh, Beulah Bueller? That's um, one and a half volts. What's going on there? Something's a bit weird. Why are we getting like a volt and a half? Well, let's go over to the keysight meter up here and let's plug in the exact same leads. Oh, we're still getting a vault.

I can hear you saying, dave, I know what's going on here. Bench meters are famous for having like a high input impedance on like the millivolt range and even up into uh, several, over one, or several of the voltage ranges. Well, if we go into there, you'll see that. No, we're 10 meg Ohms input impedance and if we go auto input z Yeah, it goes a bit higher and stuff like that and it might charge up because you've effectively got, uh, infinite input impedance.

But um, and if we like manually range it like this, look on the one volt range. We're getting overload. Overload. overload.

Wha? What's going on? And let's try an older school, uh bench meter in this case a um, old Phillips uh, six and a half digit jobby And if you listen very carefully, you might be able to hear something. You can hear a relay in there range switching because it's just going crazy. It doesn't know whether it's volts, millivolts, or whatever. You can see the m uh flash up there very briefly.

It's just going berserk. and this is actually, uh, very real stuff. Look, if we go into the uh trend chart over here, we can see. look at that.

I mean, that's we can auto scale that. Look at that. I mean, there's real stuff on there at like almost two volts plus minus two volts p to peak. It's enormous.

Why? As I said, we've got 10 meg ohms input impedance Exactly the same as your regular multimeter. What's going on? Why does this show zero? And these higher end meters show like a couple of volts and we can even choose to like, like slowly data log this as well. Look at this. I'm doing like one sample per second.

One? it's like minus one volt. Um, it's all over the place look, but you might see it is actually counting down. That's interesting. And if we do a trend chart of uh, this slow ones once per second data logging.

hmm. there you go. I left it for a bit and this is what we're getting. It's kinda sorta sinusoidal.

Not really, but something's going on. Look at that. Oh, there's a bit of bitter wiggle wiggle wiggle yeah going on down the bottom there and it'll probably go back up shortly. You watch.

Come on, come on. I'm betting it. will. You betcha.

Here you go. Go. you little beauty Up it goes all the way. You can do it well, add a little jaggy there.

you can see this is like real interesting stuff. This is real logged data. once again, with the probe sitting in exactly the same position we had for the other meters. It's interesting it's picking up something, but of course everyone knows what it's picking up.

It's just picking up mains and crap right? And sure enough, if we short the probes, it is zero and we can go back to our number display there and zero volts and we take our hand off that. And there we go. It's going up point Three point, Five point seven. Once again, this is ten mega ohms input impedance.

Exactly the same as our 10 mega ohm input impedance multimeter here. So why the difference? Well, I'm glad you asked. It has to do with the number of power line cycles or the integration time of the multimeter. Your typical handheld multimeters are like these: these: are relatively, uh, slow.

You used to. you know, like, a really fast one will get five, or even some on the market. might be like seven times a second or something like that. They're really quite slow, but these actually have built in 50 60 Hertz.

Sometimes it's selectable, sometimes it's not um, filters in them. So they're actually filtering and out the power line frequencies. Because in any sort of lab or environment where you're measuring stuff, 50 Hertz is going to be or 60 Hertz for you Yanks, Um is going to be like one of the predominant interference sources in a typical environment, office, or lab environment. So your handheld multimeters are being very nice to you.

and they actually have that uh, integration time set so that it takes samples long enough that it actually effectively filters out 50 Hertz or 60 Hertz uh, interference frequencies. But your higher end multimeters like this, they may or may not do it by default. You've actually got to go into the menu and check out the number of Powerline cycles. So if we go into Dc Volts here, you'll see I'm at 0.2 Plc or Powerline cycles, right? So I'll just stop my data, log in here and we'll go back to a continuous display.

Right and there we go, We're in Volts and number of Powerline cycles. This actually determines the accuracy of your meter as well. So Nplc is the acronym for it and you can also do time as well. Um, in milliseconds, They're effectively like essentially the same thing.

The number of Powerline cycles means it'll do an integration measurement over one 50 or 60 Hertz power line cycle. So if Mplc is set to one and then you can do that in milliseconds as well. I mean, you know 50 Hertz would be 20 milliseconds of course, so you can see if I've actually got that value very low. I don't get many, uh, significant digits there, and I also get quite a lot of noise here.

And if I go to 0.02 power line cycles 0.06 we're still getting um, you know, like volts of noise right and point Two, We're still getting quite a lot of noise. But watch what happens when I go to one Powerline cycle. Ta-da It's magically vanished because it's doing at least one full, uh integration of the 50 or 60 Hertz power line cycle. So you're reducing the noise.

And you can see of course that we've got more significant digits now. So if we go back, of course, we still have the same number of significant digits there. It hasn't changed, but because the integration time is not long enough to do any effectively like averaging, so to speak. Even though it's integration, I won't go into the differences.

But, but anyway, if we go to there, and then if we go to 10 powerline cycles, watch Ta-da We get an extra digit of resolution here. and of course we're getting our zero volts there. Once again, if I touch those leads, right? yeah, I can get you know, tens of millivolts basically um, equivalent to what we get on our um, the handheld multimeter here. and I can show you how the uh, smoothing or you know, every mathematical averaging doesn't do the same thing.

It's actually to do with the measurement integration, not the post measurement, um, smoothing or something like that. So let's go down to say point to uh, Powerline cycles here and then we'll go into math up here. And where are we? We've got smoothing filter there. we go.

if we turn the smoothing filter on. Ah, it doesn't really do anything. so it's doing and the response. also.

Um, you know 10 readings, 50 readings of smoothing. It doesn't help. So doing post sample, uh, averaging and smoothing does not help the situation. It's all to do with the how the Adc works and these are integrating Adc's You might have heard of dual slope integration.

I've probably done a video on dual slope or multi-slope integration. The keysight have their multi-slope integration and there's dual slope. And there's single slope and all sorts of things. But that's basically how your high-end multimeters.

well, even your handheld multimeters as well. Like even your low end ones they use like dual slope integration. So it's you know if you don't have your uh integration time of your measurement uh set to actually uh, take into account and average out in the measurement the 50 or 60 Hertz noise pickup, then yeah, you're going to come a gutsy like this and you're going to measure volts and you can get the meter to do weird auto ranging stuff we saw on that Phillips one and the Keithley one down here. Exactly the same uh thing.

Like I've got the smooth that smoothing field is actually on right? The smoothing filter doesn't do anything, it doesn't help your cause at all. And check it out, it's just going me auto range in there. Oh look, it's even got like 10 volts one volt, right? It just doesn't know what to do. It's just absolutely nuts.

And you turn on the smoothing filter and well, it's still. It's a little bit slower of course, but the those high voltages are still there. it's not getting rid of them. and once again, we're still 10 meg ohms input impedance.

But you'll see that we're 0.1 power line cycles. So I'll turn off the filter here and we'll change that to one power line cycle. Bingo! it's gone away because we're doing at least what an integration over one full 50 or 60 Hertz power line cycle. Nice.

And as I said, uh, meters will typically have like a setup in there for 50 or 60 hertz. and just to show you the actual waveform that we are picking up here, What I've got is I've replaced the multimeter leads with just, uh, some banana plug leads flapping around in the breeze there and I've got a uh 10 to 1 probe directly coax connected across there. So we've effectively got a 5 mm input impedance now total. But we're going to be.

You know, that's still quite high enough to pick up the noise and stuff. So if we go in here and we have a look at our trend chart, you can see that we're getting like plus minus a volt There Does that correlate with the oscilloscope? Yep, it does. Check it out there you go. Plus minus a volt there.

So yeah, no worries and I was getting before, but I'm not now. Unfortunately I was getting like large um, spikes on there so something was switching in here. I don't know what it's gone now, of course it is as soon as I hit record white coat syndrome and of course if I touch those leads there, you can see yeah it just changes. If you twist them, it's going to change if you you know it depends where you've got.

This will change from lab to lab, whether or not you're holding them, it'll change from one part of your lab to another. It'll like, just vary all the time. Because you've got such a large input impedance, you can see that change that I just played around with there on the trend chart there. and we can probably do that again.

Let me get the leads and I'll actually twist them. Okay, so what I've gone and done now is actually, uh, twisted the leads like that and you can see that that has significantly reduced the pickup there. But of course, it all has to do with the number of power line cycles. So when you're playing around with your multimeter, especially these bench ones um, that can do a really fast integration times and stuff like that, you need to know about your number of powerline cycles and how not only how it can influence uh, the display resolution um, but also can influence your noise pick up there it is just magically vanished and if you want the most accurate readings like you're going to put it on like a hundred power line uh cycles.

And to give you an example of this, I'm actually feeding in five volts Dc superimposed with a one volt peak-to-peak 50 hertz, uh sine wave and you can see that it's bang on five volts because we've got the number of power line cycles equal to one. And if we go to point two, you know there we go. It's jumping around like a jackrabbit. 0.02 0.06 There you go, it's jumping around like crazy.

But if we go to the number of powerline cycles at least equal to one, it magically vanishes and we can see that perfectly on the Uh trend chart here, you'll notice it's precisely uh, plus minus 0.5 of a volt there. One volt peak-to-peak That's exactly what we're uh seeing And if we actually extracted that data and looked in, Hopefully we can see a sine wave, but if we change our number of power line cycles instantly go up to one. Bingo, it's stopped. We're actually getting a flat line there now and we can go back to point two power line cycles and you can see at point two, it's getting a bit.

We probably won't if we actually looked at the data and zoomed in. We probably wouldn't see a perfect sine wave there, but the lower that we go, the more solid you see that's going to get right. So what I've done is, I've pulled this data from the multimeter, put it into a spreadsheet here, and we can graph it, and I'm changing the modes in the power line cycles and you can see like four distinct modes here. Now, this flat one.

over here. This is five volts. We're of course feeding in five volts plus minus half a volt, 50 Hertz signal on there and we're completely flat line in here. And you may have guessed this is one power line cycle.

So that's 20 milliseconds cycle time or aperture time as it's called, or sampling time. They're all basically the same thing. It's just different. uh, terminology.

some manufacturers might use a different term, but basically an aperture time there of 20 milliseconds. So that allows us to get at 50 Hertz signal is to be different for 60. But at 50 Hertz we get one complete main cycle so it averages out. And that's why we get a flat line.

and it's not averaging out mathematically later as we just saw. it's actually doing it in the integration or sampling time of the analog to digital converter. The if it takes this much time to sample it, in that time, the 50 Hertz noise has gone exactly one complete cycle and it's just averaged itself out and we get five volts. Magic.

But at this point here I then switch to 0.2 power line cycles or 4 and now which is 4 milliseconds aperture or our sample time. And as you can see we start to see I, you know, up to like we start to see the peak there, that one volt peak to peak signal there. But you might have noticed this. It's kind of like modulated.

You might have seen this before in your oscilloscope. This looks like classic Alison. This is all to do with your Nyquist stuff right where you need at least twice the sample, right? otherwise you get aliasing. So we're clearly getting sampling artifacts here of our 50 Hertz signal.

And it it's not good enough because we're only sampling at with an aperture time of 4 milliseconds or 0.2 Powerline cycles. Now at this point here I switched over to 0.06 powerline cycles or 1.2 milliseconds aperture time and as you can see, we really start to get a pretty decent signal. It's still not absolutely perfect because our sample rate's not very high, and at this point, over here I switched to 0.02 powerline cycles or 0.4 milliseconds aperture time. As you can see, we get pretty much a perfect sine wave there, and you can see how it's effectively changed what looks like changing frequencies there at at each point because we're taking more samples each time we set.

Each time we change that number of power, line cycles is changing our sample rate effectively. So there you go. We get sampling artifacts just like you would on an oscilloscope or a data logger or anything. A bench multimeter is no different.

it's just a sampling system. That's it. It's not rockin' science, just how these things work. And of course it all has to do with the input impedance.

That 10 meg ohms is actually quite high and it picks it all up. and if you put go whack a 1k resistor in parallel with it, it's going to knock it on the head. And when if you go and measure uh, like a low impedance voltage source like a battery because it's got like milliohms output, well a source impedance and you measure that. That's why we can get just the probes there, we can get like a volt of noise.

Yet when we measure our battery like that, we will get 1.30254 and we'll only get the noise will only be a couple of least significant digits like that. So that's all to do with the impedance of your measurement source. In this case, the impedance of our measurement source is 10 meg and it's just pink. and we've got these big antenna uh, leads on here picking up the 50 Hertz, which is like plus minus a volt.

So there you go. Very interesting stuff. Number of powerline cycles that has to do with the integration time, which is different to any sort of smoothing or averaging mode which the meter might do after that because you're doing that after the measurement and not before. So it's all to do with the measurement time of the analog to digital converter.

And of course, it's going to slow down your measurement the more number of power line cycles you have. but you get increased accuracy and rejection of 50 60 Hertz noise, so I hope you found that interesting. If you did, please give it a big thumbs up. As always, discuss it down below and check out my alternative platforms like Odyssey.

I think I'm close to 60 000 subscribers on Odyssey now. We don't win a chicken dinner. Catch you next time you.