EEVblog #601 - Why Digital Oscilloscopes Appear Noisy

hi now there's a myth regarding oscilloscopes that simply will not go away and that is that digital scopes and be they ancient like this Tektronix tds 210 sort of a bit more modern like this Rygel our 1000 a series or something like this you know really kick ass high-end Tektronix 3000 series just released and the myth is that your old traditional analog oscilloscopes like this Tektronix triple to five or any analog oscilloscope is in quote marks lower noise than a digital scope and well that's not actually true and I want to explain it to you today so let's start off by take a look at this Tektronix triple to five nice analog oscilloscope 50 megahertz bandwidth fairly typical analog scope and look at that trace look at it it's just beautiful look how fine that is I've got no input connected to these scopes or any of these scopes as we'll see and they're all going to be by the way it's set to the same volts per division one volt per division in this case and one millisecond time base just so that we're consistent across all scopes but look at that you might think how beautiful is that there's no noise on that whatsoever these analog scopes are so massively low they practically don't have any noise at all they're brilliant and then you go to something like this ancient tds to twenty a hundred megahertz bandwidth analog scope one of the first real-time bandwidth scopes on the market and we'll take a look at the waveform it's you know it's a bit fuzzy look at the noise on there if you know anyone would say that is noisier than that taken down that analog Tektronix triple to five we saw before then we've got this six seven year old wry Goldie's 1052 we it's still sold it's almost obsolete but once again it's an even nicer 50 megahertz bandwidth the firmwares being hacked it is actually a hundred megahertz bandwidth front end and is a pretty you know a good example of a modern low-cost you know bottom of the range DSO and well look at the waveform once again all these time bases are identical the volts per division are all identical and look and we get in see seen those little occasional blips in there look at that us you know good four pixels high or something all that noise you would think well that one's actually slightly worse than the Tektronix tds 220 and then we'll have a look at this brand-spanking-new Tektronix 3000 series scope very expensive scope you know over ten thousand dollars worth just released it's a quality Tektronix brand you'd expect this to have you know be a really well designed scope and well once again same time base and same voltage setting no input look at the waveform ah it's all over the shop right that is the worst of these four scopes it looks like they got progressively worse or digital scopes have got progressively worse as time has gone on well is that fact or fiction so clearly I know what you're thinking Dave you're talking rubbish I can see it with my own eyes this one analog scope traditional analog scope is definitely the lowest noise scope this one is the next best the ancient digital then the slightly more modern digital is probably better again and this latest modern one is just absolutely hopeless I can see it with my own eyes well I'm here to tell you that your eyes aren't deceiving you yes this is better but you're not seeing the whole picture you're not thinking forth dimensionally now the thing you must remember with a traditional analog scope is obviously it has no storage capability and the brightness of the image on the screen is going to be determined by how long the trace spins in that position so if we had just had a little tiny runt pulse that went boom up there like that and it happened one in a million times you're never going to see it on an analog scope because in each sweep assuming that the trigger actually uh you know it was actually able to be displayed it only be displayed one in those million sweeps or whatever so you wouldn't see it it wouldn't be there long enough to produce a bright image on the screen like we're seeing with that trace there and yes I can make that trace fatter by turning up the brightness now that is like a blooming effect on the scope but it's not just that it's also displaying more information when you make it brighter and I've shown that in my previous video as I mentioned which I'll link in down below how your analog oscilloscope can be hiding the true signal so in this respect analog scopes aren't nearly as good as digital scopes for capturing the actual hidden data in there the hidden signal in there so somewhat confusing but a lot of people make that mistake thinking analog scopes display the signal better not necessarily so watch the other video and you'll find out why so the bottom line is any noise on the analog amplifier input down here or on the signal that you're feeding in any noise which is uncorrelated to your sweep speed or your trigger is just randomly appears it's not synchronized with the sweep then it's going to appear quite dim or non-existent because it's not going to show up all the time and that is why an analog scope will always show this beautiful clean signal like that effectively what it's doing is averaging your signal by way of brightness but let's take a look at the first example of our digital scope here ancient Tektronix tds 220 but as you should know a digital scope actually samples the signal then displays it so any noise or anything else in there is going to get captured in that data acquisition and displayed now in this case this signal looks relatively clean you can see little noise artifacts on there you can see it right but it's not that bad you would think that's not too bad at all especially compared to the Rygel one above but here's the first fact you need to know about digital scopes the amount of information you're seeing displayed on that screen there is going to be determined by the sample memory depth and width this ancient scope here the 200 series TDs it's only got 2.5 k of sample memory practically bugger-all and that is why we're getting a nice clean signal there so that's fact number one and I can demonstrate that on this Rygel here now you can see that it looks a little bit worse than the Tektronix one okay one hundred megahertz bandwidth same time base same voltage no import all those spikes but this ray Goldie's 1052 II has one Meg of sample memory so for any given time base in this case one millisecond per division then it can capture much more high frequency noise and actually display it on the screen and that's exactly what it's doing and I'll show you that if we go into the acquire menu here if I it's on long memory at the moment so it's got that one Meg points it's using that one one Meg points of memory but watch what happens to this signal if I drop it down to short memory I think it's only a couple of Keio it might be 10 K on this scope but we'll see a difference it's going to clean it up not by a huge amount but it will you'll be able to see it watch see it dropped up the oscilloscopes memory depth is a K or 16 K in normal mode and with 512 or 1 Meg in long memory but there you go it answered that for us so you can see that difference there it's going it's dropping at least a whole line of pixels there it's thinner by at least one pixel one least significant bit if it's displaying 256 for example that's because in log memory mode it is capturing more of that high frequency band that high frequency noise and it's soup and it's putting it on the screen so it's showing you more of a true representation of the signal than the analog scope is because this duh this that high-frequency data in long memory mode wouldn't be displayed on an analog scope because it wouldn't be it wouldn't be visible wouldn't be on the screen long enough to light up those phosphors and that's why an analog scope appears to have less noise but it actually doesn't and fact number two that you should know about digital scopes or any scope even analog ones the higher the bandwidth the greater the inherent noise of the amplifier and other front-end circuitry so in this case we can see this by turning our bandwidth limit off and on at the moment this Rykov scope is 100 megahertz bandwidth but I if I turn this bandwidth limit on it'll drop down to 20 megahertz bandwidth and we should see this noise drop a little bit more you won't see it a huge amount on this where we all see it on our high-end Tektronix in a second but I'll show you there we go there's slightly less you can see just on the bottom of the waveform there there's a little pixel chunks taken out so it's slightly less noise look at that hey fact number two now on our brand-new Tektronix mdo 3000 we'll be able to see both of these things much better than we could on the previous one now as before exactly the same time base exactly the same voltage input and the noise looks pretty horrible look at that but if I call up channel 1 here look the bandwidth is full and the bandwidth of this scope is 1 gigahertz got a massive bandwidth so the first thing we can do is change this 250 megahertz we'll see the noise maybe drop by a smidgen it might be hard but we'll give it a go here we go 250 yep slightly you can see that grow just a little bit there and 250 we drop it down to 20 same as we did on the rygel it's less noise again look at that so put that back to its for 1 gigahertz bandwidth we'll go to the acquire menu and we'll now muck around with the record length look it's 10k at the moment ok so not a huge amount ok now if we vary this let's drop it down to a thousand just like we had on that ancient Tektronix tds 200 series watch the noise on the waveform bingo look at that it's dropped significantly and once again if we drop the bandwidth down to 20 megahertz look at that our line is exactly almost exactly identical to what we were getting on the ancient TDS 220 because our those two rules the band width makes a difference and also the amount of the sample rate showing that high frequency content but we've dropped both of those down and bingo our noise has magically vanished look at that I'll turn it back now watch this so we're back to our one gig bandwidth so we're now on 10k that's got 100k look at that the line gets thicker one Meg line gets thicker again five mega you're probably yeah when seen might not see much difference there but 10 Meg that is as thick as it's going to get look at that we're on one volt per division with no input whatsoever and a 1 gigahertz bandwidth you would think this is the worst scope in history but it's not it's actually showing you real data so there you go there's nothing inherently wrong with digital scopes you've just got to understand those two reasons why they can show more noise in quote marks is not really noise it's actually real data that's there which is ordinarily being hidden on an analog scope because that analog scope just averages out over its screen so in that respect digital can actually be better because you can easily capture that high frequency data that's really there now if we go into the acquire menu again I can demonstrate that let's go into high res mode which puts on boxcar average in so it's averaging out some of that high frequency content boom look at that and then if we go into normal averaging mode we can do that as well but that is what happens when it a ver ajiz out that content and then we can of course combine our memory depth so we can do our boxcar average in our memory depth go right down to a thousand points are let's be reasonable let's go down to 10 K get a decent amount of memory look at how thin that line is now because we've turned on that boxcar averaging over time to filter out effectively that high-frequency content most exactly like the analog scope does except the analog scope does it using persistence of vision on your phosphor based screen and then of course you have other modes like your peak detect mode which can show which captures those Peaks and stores them better than your full Meredeth even with a foul a record length of a thousand here we can still get it to display all of that high frequency data because it captures it it's got that peak detect mode in the ADC and likewise envelope mode of course you can with infinite persistence you can capture that and it just fills it up fills up the screen like that there's something you can't get on an analog scope but that's a real information there over time you will never see on an analog now there are two types of digital scopes and this will make a difference one is like your eye goal 1000 series scope without what's called an intensity graded display or variable persistence display goes under all sorts of names like that or analog like display but something like this Tektronix 3000 series does have that and that's what this intensity button over here actually does if we hit that it's at a hundred percent at the moment that's why this waveform looks exactly like it will on a Rygel 1000 series it's all chunky you know and there's there's no sort of a variable intensity in that at all but if we drop that down you'll notice that the real signal going down going down going down look the real signal the real line in there is actually thinner than that and there's high frequency noise superimposed on top of that which you want which you'll see clearly if you have a source code like the Riga 1000 that doesn't have this intensity grader display you'll always see all of that high frequency noise there is no way for the oscilloscope to tell you the difference between ones that are that appear there all the time and just noise that just appears there periodically and that's one of the advantages of the analog oscilloscope of course because it shows you that intensity graded display just like these modern scales and modern digital scopes and that's what these modern digital scopes try to do they try to replicate that sort of things so if I show you that waveform with 100% of waveform intensity operates just like that cheap low-end Reigle one or any of those low-end scopes without this intensity grader display look if I turn it down then operates more like an analogue one and you can see I've got one mega sample memory now so it's showing lots of high-frequency detail in that waveform but you turn it down you turn it down you start to see that the true line actually gets thinner and thinner take a look at that because that high frequency information isn't displayed or captured nearly as often and that's why the that's why these digital scopes offer this intensity greater display because it tries to simulate the analog scope in that respect but in my opinion they're actually better digital scope is better than analog scope because you can actually pick up that information especially one with these intensity graded displays really fantastic look at that see it's almost all gone and there's the tiny amount the tiny waveform that's the one that's there all the time and the rest of that information is just more uncorrelated noise around that and you'll really see that here because I've added 30% noise to this signal so if we turn it right up to a hundred percent there we go look at that right there's a ton of noise that is deliberately added on that wave form because this scope allows me to do that you can see if I turn that intensity right down because that noise is effectively uncorrelated it goes away like that and you can see that the noise is uncorrelated because if I go into my choir mode and average down here BAM it disappears so that noise was actually superimposed on that signal now if I turn on the fast acquisition mode which this Tektronix 3000 has I can select different variable intensity display modes now the normal one we had before down here was showing up your low but if I set it to temperature mode here check this out we'll see what it does change the waveform intensity got it set to 100% so it looks just like it would on that you know cheap low-end or I go without any waveform intensity but we turn it down you will notice that the waveform changes color and outside they're the ones on the outside like the little hints of blue and green out in there they're their signal anomalies or noise or whatever it is on your signal that is occurring less frequently than that red line in the middle just like the color temperature gradient in light for example same sort of thing so it's that red part in the middle there which is your main signal which is showing up all the time that's your baseband signal with all the less frequent noise or it could be you know some other part of your signal cut it's real information there that's ordinarily hidden on an analog scope it shows up here as you know infrequent blue and green data this is really handy very powerful feature of digital scopes and especially this one with that dark color temperature graded display and also has an inverted mode here which actually does exactly as the name says it actually inverts the waveform so the less frequent stuff the glitchy stuff all that noise shows up brighter than your main waveform so we go up to 100% warm it's all there and you may waveform in the middle vanishes that's just a neat little different way to view your data on this MDO 3000 but that's the same noisy signal 30% noise added and can we see it on our analog scope well not really it's very difficult look it looks like a very clean sine wave but there's I've added all that noise to it you can't see it because it's uncorrelated to the sweep signal and it's only showing up and very briefly there but we should be able to capture that with the camera now I'll attempt to demonstrate that this analog scope can actually display that high frequency content because it's so short it doesn't light up the phosphor as much and your eyes can't see it well we can try and attempt to capture that by using a long exposure on our camera here so that's what I've got here I'll put what they used to use back in the old days is camera hoods you could buy that's what these ridges around here will fall on these analog scopes you buy these hoods you hook your film camera back then up to it and you can get long exposures well I'm going to do that I'm going to whack some t-shirt over the top like that and I'll turn out the lights here try and get as dark as possible and I'll see if I can capture the noise on a couple of typical signals first a flat line with no input and then a two and B the control so there shouldn't be much noise on that and then a 1 kilohertz analog signal that will have noise on a digital scope but you won't see it on the analog unless we do this and here's what I shot with the camera at different shutter speeds now the signal was barely visible that's 125th of a second and you'll notice that it does get a little bit thicker there and now it's sort of pretty much stopped so that is the real signal there a bit of blooming but basically there was a difference between the signal as originally that I could see with the eye and then what we recorded with the camera got a little bit thicker and here's that noisy signal as you can see barely visible at the low shutter speed and that's what it looked like to my eye but as you increase that shutter speed you start seeing all that dim phosphor you couldn't see with your eye and you can now see the noise superimposed on the waveform brilliant so there you go I hope that's cleared up the myth that digital scopes are noisier than analog scopes because they're not they're just better and they work differently and hence they're showing up all that high frequency stuff that your analog scope has pretty much been hiding from you all these decades by way of the the phosphor persistence on there and having that uncorrelated noise or signal just not being bright enough now back in the old days when we didn't have digital scopes you had to turn the brightness right up on this sucker to sort of you know see all that tire I think there's something eating in there is it oh I don't know sometimes you might have to get your camera and hook it up to actually saw it as I demonstrated there it wasn't a totally thorough demonstration for the camera but it did at least show the difference that when I've got at low intensity like that there is actually more information there that my eyes just aren't picking up because they aren't displayed as frequently but the exact same signal on a digital scope especially one that doesn't have variable persistence like this riga all this ancient Tektronix one it shows up and that's why the waveform looks thick or noisy but it's not yeah there might be subtle differences between analog phonons but it's not like this modern Tektronix 3000 series scope just released is going to have a noisier analog front-end than this ancient analog Tektronix scope no it's not the case and it's a little bit involved in terms of the 8-bit digital sampling of things like that but in the end what it comes down to is as I said the memory depth of the scope the more memory depth the more you gonna pick up all that high frequency content the greater your bandwidth the more noise you're going to see and that's inherent in analog scopes as well not just digital and of course your variable intensity displays if you got something like the older I go or the old Tektronix here that doesn't have variable intensity well you're just gonna capture everything and sometimes as I showed in the previous video that can be a good thing so there you go don't rag on digital scopes they're not that bad they can actually be better sorry for all you went a lot gray beads out there catch you next time you you

Channel: EEVblog

Views: 168,294

Keywords: analog oscilloscope, scope, oscilloscope, digital oscilloscope, digital, analog vs digital, waveform, capture, noise, tektronix, tektronix mdo3000, mdo3000, camera, long exposure, oscilloscope camera, Digital Storage Oscilloscope, sony nex 5t

Id: Znwp0pK8Tzk

Length: 24min 36sec (1476 seconds)

Published: Thu Apr 10 2014