Dave compares the FFT modes on 7 different oscilloscopes:
Rohde & Schwarz HMO1202 Series
Tektronix MDO3000
Keysight 3000X Touch
Lecroy WaveJet 354 Touch
GW Instek GDS-1104B
Rigol DS1054Z
Rigol DS2000
Which is the best?
Which one sucks the most?
Forum: http://www.eevblog.com/forum/blog/eevblog-845-oscilloscope-fft-comparison/'>http://www.eevblog.com/forum/blog/eevblog-845-oscilloscope-fft-comparison/
EEVblog Main Web Site: http://www.eevblog.com
The 2nd EEVblog Channel: http://www.youtube.com/EEVblog2
Support the EEVblog through Patreon!
http://www.patreon.com/eevblog
EEVblog Amazon Store (Dave gets a cut):
http://astore.amazon.com/eevblogstore-20
T-Shirts: http://teespring.com/stores/eevblog
Donations:
http://www.eevblog.com/donations/
Projects:
http://www.eevblog.com/projects/
Electronics Info Wiki:
http://www.eevblog.com/wiki/
💗 Likecoin – Coins for Likes: https://likecoin.pro/ @eevblog/dil9/hcq3

Hi, it's just an ordinary day in the Eevblog lab here. I've got five oscilloscopes, a spectrum analyzer, and a function generator. What are we going to do with these? Well have some fun. of course, let's go now.

I Recently took a look at this Rohde & Schwarz HMO 1202 series scope and I noted that it had a really good FFT mode and actually a quite large art number of FFT points as well 128 K points which is a little bit on the unusual. yeah, it's not. While it's on the high-end side for these scopes and somebody asked, well, can I compare it with some other scopes and see what it's like So that's what I've done here I've set it up and what we're comparing. what we're going to have a look at is the Rhone Schwartz compared to the Rye Gold Ds1000 Z which everyone is familiar with I'll put in the 2000 Z as well, but I believe there's actually no difference.

Oh sorry the 2000 Diaz mm believe there's no difference what they tell Adele Lacroix wave Jet touch Here we've got the GW in stick Gds 1104 B We've got the Keysight Msox3104t the good ol' Dry Gold Esa Eight One Five bottom the range spectrum analyzer but it gives us a baseline. So I've got my signal into our function Jin here and I've got it set up for an FM signal. So with the carrier frequency of one megahertz here with an FM frequency of 5 kilohertz and a frequency deviation of 500 hertz and if we have a look at that on our Rai goal, our DSA eight, One five, its bot on the range spectrum analyzer, but it does the job. So this is a real RS spectrum analyzer.

It actually gets in there and sweeps the frequency across. None. This digital Fft. rubbish.

We're getting exactly what we expect here. Here's our carrier frequency bang on one megahertz. We've got a span of 50 kilohertz s. That's 5 kilohertz per division.

so you can see our frequency modulation there and there and we're getting tiny little ones over here as well. There just start some harmonics of our five kilohertz there. and let's have a look at what the same signal we get on the various scopes. Now how the FFT mode or fast Fourier transform mode on a typical scope works is that it takes your regular input signal here.

And here's our signal like this and we can say if we go in here and that not used to all these different scopes that I've got anyway, we can see there's our waveform in there. It's a little bit crusty, isn't a bit of distortion anyway. doesn't matter, it's real easy and you can slightly see the little modulation on there. Maybe you can see it a little bit.

a little bit of shimmy in there, so that's our modulation. So what it's doing is when we turn on FFT mode, it'll actually compute the fast Fourier transform of this. and then we cannot just go in there and zoom in. But the FFT depends upon many different things that depends on the sample memory of the scope, how deep a memory you've got, it depends on the sample rate you're using, it depends on your current horizontal time base setting, and most importantly, it depends upon the number of FFT points and this road and Schwarz one is unique.
And not only that, it has a high number of FFT points. it has a hundred and twenty eight K points, but you can actually select it as well and that's really good and we'll have a look at the effect of that in a minute. but that effectively. And what that does the number of points is that basically the frequency resolution for each individual pixel in there.

It's got to calculate like a frequency bin. For want of a better term, it's got to calculate each individual point and there's a lot of processing involved in this. And to do actually a hundred and twenty 8k point FFT requires a lot of processing grunt. So typically it might be done in an FPGA or an ASIC.

When you do it on like a regular Joe Bloggs arm process or something, it's going to probably, you know, grind to a halt, but yet we can actually select that so you can see that when we drop down in, that it just gets coarser and coarser. And of course, it gets faster because it's able to calculate the fast Fourier transform faster. Now it does the FFT using a DFT or a discrete Fourier transform and depends and depending upon the manufacturer you get, they might implement that DFT in different ways. So ultimately, what the oscilloscope is doing with the FFT function is it's converting the time domain waveform here and doing a discrete Fourier transform on it and actually converting it into the frequency domain.

So now on the X-axis we've got frequency and amplitude on the Y just like we have before, except where a DB scale of course, and we can actually show you that. There we go. if we just put voltage like that and no good, you've got to have it on DB because we're talking about quite large relative magnitudes there. So of course, the FFT converts your oscilloscope into a rudimentary spectrum analyzer.

But depending on how well they've implemented that FFT how many FF TR points has it got like this one and how well it actually does everything determines how useful it is. And as you can see this road and Schwarz One excellence. Got a hundred and twenty eight K points F F T and it allows us to get excellent resolution. You can see our carrier in there, you can see our frequency modulation, and also if your scope has got the high resolution mode as well, you can whack that on and you'll notice that the noise floor will drop a bit.

It'll be a bit better. Has to recalculate that, thinking, thinking, thinking. But there you go. It's a little little bit better, a little bit better.

And also what the number of points does here is it actually effectively lowers your noise floor so there could be like details hidden down in there that we can't actually see. For example, so if I lower the number of points, we'll see that this side being component here might eventually vanish because the noise floor is effectively going to lift and I should be able to show that if I scroll like a right down to a small Center frequency just so that it's always going to be on the screen where it does the min. mm 48 points now and just watch where this noise level is and you'll see it slowly drop just each time. I increment the number of points there.
you notice that it's slowly slowly drop and you'll see all the way up there we go. So that is like a significant difference. And of course if you do average in all sorts of other you know stuff, you can really bring signals out of the noise. and here's a shot from an Infineon application.

Note what's showing the difference between like a small number of FFT points and at like millions of FFT points, you can see really how all the signals come right out of the noise. So if we have our maximum number of points here hundred and thirty 128 K and then we go into our choir menu. We're just in regular refresh mode at the moment, but if we actually go into average mode, for example, we can actually watch the noise floor drop. Here we go there we go.

It drops a little bit and it brings out a bit more signal-to-noise there. So there you go. It makes a difference. There we go.

you can see it, add to averages and tweak it up. Let's go for broke and then we can really go to town. We got our averaging on 512 every just thank you very much. But then we can also turn on our high resolution mode and wait for it.

Oh look at that! Beautiful! This is excellent. It's doing a really good job. And the other thing I like about it is when you turn FFT mode on right where we now time base. here.

When you turn FFT mode on, you don't have to go around with the menus or the select button. What it does is it now enables your time per division to work like this on your FFT Like this. Absolutely fantastic, So you can do that and then the position. The horizontal position control now works that and it just makes it incredibly incredibly simple to use and intuitive to use.

It's fantastic, of course, but if you want to go back to and adjust the time base, you've got to actually switch the FFT off. And then now we go back in and we were our time base again. like that. But yeah, it's just so intuitive.

Let me show you one that's not intuitive. and here we have the GW in stick GDS 1104 B Now this thing actually, it claims to be the Ducks guts in terms of FFT it claims one claims that can do a 1 Meg point FFT and well, I it's doing a lot because look, check it out I think it probably is doing it, so it must have some real grunty hardware in there to do a one main point. FFT It's absolutely incredible, but as you can see, we've got our carrier. We got our FM modulation and of course that component down there as well.

so it's just as good, if not better then the road and Schwarz one. But ah, just try and use this thing right? The horizontal time base still works on the time domain up here. Okay, so it's like that, right? So my horizontal time base is still there. Where's the let's go back to here? Okay, but now.
Well, this is fairly typical of scopes, but look how we have to dig around. Okay, I've got to use the the variable knob here. Okay, I've got to select which one I want, right? Even my horizontal per division or my center point. Okay, so now I can actually vary my centre point, but when that varies depending upon the setting of the horizontal per division.

So if you want to go all the way in well, all the way out sorry and then scroll across, you've got to dick around with that and then you can maybe move this across and it's just a lot lot of around. And trust me, when you're trying to set the damn thing up for the first time and it's jerky of course, because it's got to do or the FFT processing in the background and it's just it is really horrible to use. Absolutely awful, but it ultimately can do the business. And by the way, both scopes have been set to the hanging window.

Here you've got various different types rectangular, ham Hanan don't get confused between two and good old Blackman or Blackman Harris and so we've got that set to the same I've got to set to Hannon as I did on the roan Schwartz one because that actually will make a difference. You'll see it there we go: rectangular. Let's say I won't go into details of how all these various SAP modes work and things like that, but as. but it's just when you're comparing them, it's just important to use the same windowing technique.

And I've also got both scopes set to white one MIG point sample memory and they're both working at 50 Meg samples per second here: 1 Meg points as you can see same time based setting so it just allows us to get you know, do decent comparisons. and the third cab off the rank Here is the Keysight Msox3104t site because it it's purely automatic, but you know it's doing that. It's doing its deep memory, a business and as you can see, it's doing it perfectly well. The Keysight has a rated 64k Fft points, so not quite as good as around Schwartz.

Not nearly as good as the GW in Stick, but as you can see, does a pretty damn good job of it. But unlike the Rhoden Schwartz, this one is also requires you to dick around with the menus down here to actually set your your center and your span. But the good news is is that you can Actually because this is the new touchscreen scope, you can actually art typing exactly what you want, so you know. But it's not nearly as fiddly as the GW in stick, so it works.

You know it works fairly well and the Keysight also has much better velocity control here. And it doesn't slow down based on the Fft or anything like that so we can keep our spare at a low value and then we can actually go looking that way. We just jumped right up to Twenty Four Point Five gig, right? It has a hell of a hell of a velocity control, but it really is implemented beautifully so the what we had to dig around was so painful. you want to rip your hair out on the GW in stick.
we can easily come to one megahertz. It's almost as quick to use the knob as it is to type it in. Really? I'm not I'm trying to talk at the same time I'm not really concentrating. but yeah, you know you can really zoom.

You know, narrow in, straight on that. It is very nice. so yep, thumbs up to that. Now the Teledyne Lacroix Wave Jet Touch Three, five, four? Well I Just like the yaki site, this one also allows us to turn off the time domain waveform as well.

Excellent. but this thing has a rated eight K points FFT So but look I mean this is it is horrendously bad. It is awful. This is the absolute best.

I Can get it around trying to get all the settings right and everything else. This is the best I can get like I can't even I don't even know what's going on with the noise floor down here I cannot do anything. Look, it looks like dick and balls. that's what it looks like.

and there's only a small selection of FFT windows here. very confusingly. Vaughn hand that's actually another name for a Henin so it does have it. Better yet, like I don't I don't think I've ever seen it called Vaughn Han on any other scope.

Anyway, yeah that's all you get, but oh geez like it's just hopeless. but it kind of sorta there. but yeah, but um it does have a dedicated math control here and allows us to change the offset there. And the other good thing is is that the horizontal control does actually become the span for the thing.

so just like it did on the road and Schwartz so thumbs up there. but apart from that it is useless I mean this is I've been digging around this I Got one point memory on this thing and like this is the best I can get. it's just it's so rudimentary it almost doesn't work. But anyway I'll show you one that's even worse and sorry to all you Rygel Diaz 10:54 Zed Fanboys out there.

um yeah, these low-end scopes just do not cut it. This is the absolute best you can get with one of these low in Rygel's it's so how low is it? Well, it doesn't even tell you how many points it can do in FFT mode. it's it's not many clearly. So this is the most optimum art setting.

I Can get 20 microseconds per division and you know it's like that is worth tweaking this thing getting optimal settings for everything. That's the best it can do. it's just it's no good at all. I mean yeah, you can see a carrier but me, nothing else now.

I'll just show you how here how the horizontal art time base, which is up here 20 microseconds per division affects our ability to set a center frequency and our span as well. So right? I've got. Let's say center frequency right? That's as high as it high as it goes. Okay, absolute maximum.
our span is as high as it goes. There We go down to 25 kilohertz, up to 250 kilohertz. That's a 20 microseconds per division. And yes, you can see like it's now got the full range.

Okay, over there. so our our span is at 250 kilohertz per division. So 500 one Meg So there's our one made carrier which we want to measure so we can actually go in there and see it. Okay, but now I'll change my horizontal time base to 50 microseconds and you can see that it's dropped down.

Once again, this is the absolute maximum we can do here. We can only do a hundred kilo Hertz per division, can't do anything more. You'll notice that we're set to 1.2 Meg points here. it just cannot.

It's got a very low number of FFT points so it can't use all that sampled data and that's one of the keys with having a high number of FFT points. No point having hundred 50 million gigabytes of sample memory. if your FFT algorithm, it just can't use it. So I'm yeah.

so a signal is way outside the range here so we can obviously can't select that time by setting. Okay, and if we go were smart, we can only go to 800 kilohertz. Okay, so it got it. So the absolute best range that we can get where we're going to get the most resolution out of this thing is at 20 microseconds per division.

Because anything faster, let me go to 10. Okay, or you know, let's just go down to 1. for example, you know, looking right, we're at 5 megahertz per division. Okay, we're going to get no resolution in there at all.

So the absolute best: When you're mucking around trying to get the best FFT possible on your scope, you've got to do this. Find where you're I'm tweak. sorry, the horizontal I'm not showing on the screen here, but dump. Yeah, So that is the absolute best we can do.

So we want to go center, we want to go 1 megahertz and buh-buh-buh-buh right? And then we want to change our span like that. and bingo. That is the absolute best we can do on this thing. The absolute best.

So as you can see, it's best ain't good enough. So while these are in modern low-end digital scopes I have thoroughly impressive value for money. It's got more bills and whistles you can poke a stick out. It's got a ton of memory.

This thing has like 14 mega standard I think it has. well when you if you buy the option of oh I've got a Dickey I think I've got a Dickey T piece there I think I do Anyway, you get these things impressive amount of memory. It works great in the time domain. Everything else but the FFT on it is almost like a toy.

And we'll see if the Rye gold DSR 2200 is any different. We're at 50 microseconds per division, which wouldn't work before and sure enough, it doesn't work here ever. Look, we'll change it down to 20 microseconds per division. Bingo! We've now got our FFT and we can maybe zoom in on that and have a squiz.
But the good thing about the ROI goals is that yes, a Once you're in math mode, the horizontal actually does that. If you just press channel one bingo, you can go back and that changes your time domain and you press math again and bingo you're in. You can actually adjust that and then your time base can be used to zoom in. So Jersey Jersey Okay, but yeah, what? Wah-wah Thanks for playing.

That is no good at all. Whoo jumping around like a jackrabbit? There we go. Center 1 Meg Where it tend to be per division? Maybe we can change that, Can't we hang on? No, how do we? How do we adjust? Oh, that's right. we've got to go in here and we've got to go like that and then select now at 20 DB per division like we were on the other ones.

Goodness around around there we go. Yep, no good at all. Not actually sure why it's jumping around like a jackrabbit on that. 20 microseconds are like we're triggering smack in the middle.

There, nothing wrong there and then then if we go to 10 microseconds per division and our time domain, then it doesn't jump around anymore. And we got similar to what we saw on the Rygar 1000 said, but yeah, see, it's It's pretty much just a toy. You can see that something's there and that might be good enough for a lot of uses. But to analyze something like this FM modulated signal? you know, like use it more like a real spectrum analyzer.

It's no good at all. Now, if we go back to our Road and Schwarz here, it's a good example because it allows us to change the number of FFT points which is absolutely beautiful. and you'll notice that if we change. okay, we're at maximum 128 K points.

Change it to 64 K points and you'll notice that it's hard like that. Look, it is. Stopped here. Okay, and the center is at 833 Kilohertz.

That's why there's no signal anymore, because our 1 megahertz is out here like this. Okay, so using this particular time base of 5 milliseconds here, sorry for the finger in the way, I should use my poker. Five milliseconds per division and at 65 K points, we can't get that. so we're going to have to go basically turn.

This is where it becomes a bit annoying. Let's go down back to two milliseconds per division here and turn the FFT back on and you'll notice that we will hopefully get it there. We go now with our 65k points. We can actually go in there and see that, but all depends on the pond, the time base.

So yep, we're at the center now and bingo we can go in like that and see it. But of course if we change our if we don't have enough data there, let's go down to say an order of magnitude lower 200 microseconds per division, turn our FFT back on and look. we can't get see the resolution in there is no good. Okay, because we're calculating.

Even though we're still calculating a massive 65,000 FFT points in there ie. frequency bins, it still isn't enough. We cannot zoom in any further on that and you'll notice that it is now giving us a much greater frequency range. Else in our span is now a hundred.
Now this, Yeah, it's like 25 megahertz that span there. It's absolutely enormous. Good thing about this is that you want to see a bit more I Can just turn that menu off there. That's quite jazzy, but you can see how it's all a big trade-off in terms of you know, getting the right I Got to have the right memory depth set that's not going to work.

You've got to have the right time base as setting you've got to have. Well you know, generally if your scope can do you know allows you to change the number of points like this, then you set it to you know maximum unless you want really fast updating. So if I now set it to 8 K points FFT same as what we had on that crappy Teledyne Lacroix. Look, you can see that it's given us.

You know, like that is a similar sort of result at that particular time base. Let's see if we can tweak that. so if we turn off FFT let's give us a bit more. Let's go at the 1 millisecond can we get 1 milliseconds? Worth note because it only gives us our I'm sorry our span 500 kilohertz not good enough.

So the best time base we can do is 500 microseconds There we go 500 microseconds per division and now we can zoom in. but you know at least it is doing a better job than that tele dining. Lacroix You can actually see the separate signal components there. You know? it's fairly clean.

and of course as I said, you know if we go into their choir menu high or high resolution mode, maybe we can clean that up a little bit there. But we can certainly go in there and see these things. and with this road and Schwarz, the measurements on the road and Schwartz are really quite nice. I'll show the quick view thing but if you're going to cut which are and auto measure which doesn't work in FFT mode sadly.

but if you go like you turn the cursors on next peak, previous peak works an absolute treat. Look at that so hoping and start to appreciate the huge trade-offs in terms of sample rate, time based setting, number of FFT points, the higher the better you know it's worth painful if FFT functionality is something that you want in a scope because you know you could end up with a toy like those Rygel scopes and you know it's really no good. So there you have it. There's a little look at the FFT modes on our five different scopes here.

No, I didn't deliberately didn't use the Tektronix I M do 3000 scope because it's got a built-in Hardware spectrum analyzer so you know I guess I could get the FFT mode out and try that just for kicks shall we? Yeah, yeah, why not. Anyway, I'll do that after this bit dump and as you can see I Love this road and Schwarz works really really well. It's got a large number of FFT points and it's functional. its usability functionality.

the auto setup on the FFT gets you in the ballpark and it just really is quite a nice scope. Of course the art of the keysight which has which is advertised as having a specific or what this new touch model specific FFT functionality can do gated FFT it's extremely powerful which the road and Schwarz can't do any of that. so it's in terms of FFT the best is the Sytem Msox3104t. It does the job so you know, Hey, you got to give it a thumbs up for that.
Uh, and the Teledyne Lacroix Here is it. It is a joke. The Dick and Balls model yeah, don't like that at all, huh? Um, that's not good. Pair on self test failed.

What? What therapist is the power pleaser in qualified? Unbelievable. Um, well. I've got my signals connected in, but surely that shouldn't make a difference. That is greatly disturbing, huh? Now if we have a look at this on the Tektronix MDO 3000 scope here, this is not the FFT This is using the analog RF front-end because it's a mixed domain.

Oscilloscope does actually have an RF spectrum analyzer built-in and they'll be it. It's a digital sampling based system. so it does actually do an FFT approach, but it actually has you know specific hardware and software to actually yard do this. so it's not a that's why it's faster updating.

Even though this Sloka scope is generally slow as a wet week, it is much faster updating because it's doing the discrete Fourier transform of the or the FFT of the signal instead of actually doing the sweep based system. you know, generating a sweep and going across because the Rykor DSA eight one five to build up the image that we saw before actually takes like fifty seconds I Think to do an entire sweep because I had a low resolution bandwidth set in and everything else. but there you go. This gives an excellent result.

Check it out and its performance is better. As we've seen before, performance is better than the entry-level Roy Goal: DSA Eight One Five And as you actually saw before, we've got a modulation index set up here of 0.1 So the modulation index is the frequency. Devia is their frequency deviation divided by the FM frequency there. So it's up.

We've got 500 Hertz deviation with a 5 Kilohertz FM frequency. So I'll just interestingly show you what happens if we take that above 1. Okay, if we take the modulation index above, 1 will see our carrier actually start to drop and as we go up, say to 0.5 or something, we'll actually see more. Our site will see more tones down here getting smaller and smaller and smaller.

So it's 500 Hertz at the moment. Ok, this is 1 Kilohertz and do we see any note? Not really. let's take it up to 2 Kilohertz. There we go, but another can see.

There we go the increasing amplitude and then we start to see more of them. If we increase the modulation index, let's go up to 5. So I've got a modulation index of 1. so our frequency deviation is 5 kilohertz.
How? FM frequency is 5 Kilohertz and away? Look at that. There we go. Beautiful. Now if we take it above that.

So we've got a modular I've got now taken up to 10 kilohertz frequency deviation. so I've got a modulation index of 2. Bingo! Our carrier here has dropped and let's take it up to 20. So now you can see the carrier's actually gone back up and the side.

The first sideband here has dipped back down and then it goes up again. And this is all classic textbook art stuff for FM Frequency Analysis Theory So a Go look it up if you want. but it works. Excellent.

But if we actually try and do an FFT on our channel one signal here, look, it's got the same annoying thing as the GW in stick. You got a set. You know your horizontal controls actually still work on your time domain signal. So you've got to now dick around with these two multi-purpose controls.

I Hate having these 2 separate controls on the MD O 3000. It's just. oh, it's excruciating. Anyway, well did Actually in this case is a bit handier because this is a good use of the dual knobs.

I Guess you don't have to dick around with this so we can actually tweak that. you can see how slow this thing is. It's just whoa. We can easily overshoot with this.

so I Anyway, this thing does have a keypad, so actually we can type it in. Ah, Beautiful! Yes! I Love things with keypads are so wonderful, so wonderful and we want 100 kilohertz. There we go. we're in luck.

Flynn Okay now I Set this to what one MIG point memory I set it to Hannon window. Here we go and one megahertz Center with a hot well? how do I Cherie It said it's changed at 125 I'm sure we set it to a hundred anyway. Um, it looks like it needs to. it.

can't 100 kilohertz? There we go. Note: it doesn't like that anyway. so we need to work. uh-oh the sorry bar per division do Okay so let's go in there and 10 kilohertz, let's change to our carrier.

one megahertz. thank you very much and didn't like that. Did it. Ting kilo Hertz per division.

All right there we go. 12.5 Okay, now we're in life. Flynn Here we go. It's doing the business now.

But look, it's not updating. It's not updating like you'd expect to see the noise change that way. We got one where he got one. Um yeah, this is the MDO 3000.

In a nutshell, it is one of the slowest modern scopes I've ever used a What? I think it is the slowest. It is just horrendous every time you turn something on. It's got very limited processing power in this thing. so I don't know how many points? F F T This one actually does in math mode I wonder if it's in the manual? Let me go read it, but let me try and turn the channel off.

Can we still do the FFT I Don't know. But anyway, look how sharp this is. Absolutely incredible. Must have a massive number of frequency bins ie.

a massive number of FFT points that is calculating this is God Yes, it still works by the way with the waveform of excellent. So the Tektronix actually is has got to be the winner in then in this FFT shootout. By far, that's got to be equivalent to the million points in the GW in stick, no doubt. So no wonder it's as slow as a wet week is that? No, it doesn't tell us we've got one midpoint is our sample memory, but it doesn't tell us how many FFT points it's actually doing and there's yeah, there's no indication in there or you can tell us by help slow it is.
And yes, I Just read the manual and sure enough, this thing has up to two Meg points. FFT So absolutely brilliant. Um, yep. okay, sorry Tektronix No wonder it's a slow calculating two million Meg points.

But the GW instinct was faster doing a million. It was at more than twice as far. so this is still a very slow scope, right? Don't get me wrong, it's a dog. But hang on.

Why is it now like it's now tight? Are there we go? We got one. Yeah, it takes forever. It's still very slow, but it can give you the performance. not that you really need it because you're going to use the real Iris spectrum analyzer.

but anyway, this is indicative Dick, He should be indicative don't quote me, but should be indicative of the other non-m Do scopes in in text range because basically the M Do is one of their existing series with the RS spectrum analyzer hardware tacked on. and that's basically what it is. So yeah, this performance should be similar. So absolutely the M Do is the winner.

although I'm not sure if it can do gated. but I think we can? Can we do gate it on the real RF like we can on the keysight? I Don't know. Anyway, it it is a winner. So very very happy with that.

Hall can do advanced math and there we go that's zoomed in a bit. That's 2.5 kilohertz per division now, so check that out. Beautiful result. Absolutely gorgeous.

So there you go I Hope you enjoyed that. Well what's supposed to be a quick look? I Always say that, don't I and then I just walk alone and yeah, yeah, find extra things to do anyway. Look at the difference between FFT modes on various scopes. You can see how some of the entry-level ones are just toys.

pretty much. Yeah, you can detect that some carriers there, but that's about it. You can't see some basic RFM side bands and things like that. So um yeah, but I like that little road and Schwarz is very cute.

Look at it. But yet the winner Tektronix But the keysight one is really awesome as well. GW In stick a lot of points in at million points, but it's a bit annoying to use. but ultimately yes, it does do the business.

So there you go. It was not designed to be a tutorial on FFT tutorial or anything was just a comparison of all the different scopes. But I hope you liked it. If you did, please give it a big thumbs up and all that sort of jazz Catch you next time you you.
.

Avatar photo

By YTB

19 thoughts on “Eevblog #845 – oscilloscope fft comparison”
  1. Avataaar/Circle Created with python_avatars Eng. Eletricista Vanderlei Martinez says:

    voce nao pode doar um osciloscopio aqui pro Brasil ???

  2. Avataaar/Circle Created with python_avatars lasantha bandara says:

    Sir I would love to work with Oscillas Cop but I don't have the money to buy this device.If you have a used one, can you send me one for free

  3. Avataaar/Circle Created with python_avatars Mick D says:

    Johann Carl Friedrich Gauss – pure genius discovered FFT, but didn't have much application for it.
    FFT was rediscovered for cold war nuclear disarmament verification and useful for a million other things!
    Very in formative as usual, Dave – we all owe you big time – Merry Christmas!

  4. Avataaar/Circle Created with python_avatars وليد الحوراني ابو حيدر says:

    لماذا يستخدم هذا الجهاز
    لدي مثله ولم اعرف لما هو؟؟
    اجابة لو سمحتو

  5. Avataaar/Circle Created with python_avatars radiotec76 says:

    I wish I had just one of those flat screen scopes. I have a Tektronix 465 that doesn’t work and a 475 that needs alignment.

  6. Avataaar/Circle Created with python_avatars kot kot says:

    Hello friend, which will be better Rigol DS1202Z-E or GW Instek GDS-1202B?

  7. Avataaar/Circle Created with python_avatars Garry Chauvin says:

    Best of others

  8. Avataaar/Circle Created with python_avatars Christopher Gossett says:

    Looks like mission control on your bench

  9. Avataaar/Circle Created with python_avatars Paul O'Neill says:

    Dave creams himself over the 128kpt R&S FFT and takes a huge dump on the GW Instek 1Mpt FFT.
    I'm guessing Dave is looking to score some sweet sweeet R&S free test gear for his lab.

  10. Avataaar/Circle Created with python_avatars sửa bo vlog says:

    Goot . ..

  11. Avataaar/Circle Created with python_avatars Name Redacted says:

    GWINSTEK CLONED Tektronix scope menu. Along with the MORONIC FFT controls. My DPO4000 is AWFUL at FFT. You blow it up with a few adjustments, then you have to restart from scratch. AWFUL.

  12. Avataaar/Circle Created with python_avatars Patrick Cameron says:

    As folks have mentioned, the DS1054Z is capable of a lot better – but it's a trick since you can't directly set the sample rate and adjusting the center frequency can take a lot of knob twisting. On Dave's signal you are shooting for a sample rate of 5Msps to see a symmetric plot or 2 if you want a better plot that just shows the left side (nothing past 1MhH). Use auto to get a nice stable waveform and fine adjust to fill the screen vertically. Under acquisition set the depth to the first value larger than 16k (depending on whether you have 12 or 24M of memory). The FFT memory mode supports 16K samples so there is no benefit to more depth. I can't see a difference between normal/ave/hires but peak doesn't seem to work well. Under the math function turn on FFT and select Hanning window and memory FFT (rather than screen FFT) on the second settings page. The trick at this point is to walk the sample rate down to 5Msps one step at a time by adjusting the horizontal timebase (CCW). Set FFT Hz/Div to its largest available value and the center frequency to 1Mhz. If you don't keep the Hz/Div at max or you skip steps it can take a million turns to move the center frequency. Always adjust the center even if it shows 1Mhz because it LIES and sometimes doesn't update when you change the timebase. Once you've gotten the sample rate down to 5Msps and the center at 1Mhz you can reduce your Hz/Div to 5k. If you get to a point where the center won't go to 1Mhz check that you are still in memory mode with Hanning window. It sometimes returns to defaults for no apparent reason. Check these again at the end just to be sure because the plot will be terrible if Hanning is off. Adjust the vertical size/position of the plot and you are done.

  13. Avataaar/Circle Created with python_avatars Renaissance-Pete says:

    Dave, your comparisons of FFTs was very interesting. It would be great if you could compare the Tek 3104 with the GW Instek 4-channel MDO-2204EG (or the 2-channel MDO-2202EG) both of which have FFT of course, BUT, have a dedicated Spectrum Analyzer as well as a 2-channel AWG. In addition, both can do BODE plotting for FRA. And I don't know the street price on the Tek 3104 but I'll bet either Instek (MDO-2204EG or MDO-2202EG) will be less costly than the Tek. And, if you could demonstrat BODE plots with one of the Insteks that would be fantastic. I love your characterization of the RIGOLs "Dick & Balls" model… hilarious! I was laughing my balls off for at least 10,000 milliseconds, if not more!

  14. Avataaar/Circle Created with python_avatars rocifier says:

    Is the RS Pro 1100 series the same as the Gw Instek here?

  15. Avataaar/Circle Created with python_avatars وافي الله شفيع says:

    Where is siglent sds1202x 200MHZ osliscope?

  16. Avataaar/Circle Created with python_avatars ⵠⴷⴶⵁⴸⵏⵖ says:

    "dick and balls" he said

  17. Avataaar/Circle Created with python_avatars Rob Schaap says:

    comparing a 15K oscilloscope to < $500 ones?

  18. Avataaar/Circle Created with python_avatars Inspironator says:

    Thanks for disappointing me about my Rigol 1054Z's FFT capability, Uhhh, I think. :/ The cheap SDRs seem to produce a much better display resolution, so why these scopes are so wimpy is a big question.

  19. Avataaar/Circle Created with python_avatars Paul D. says:

    the Tektronix MDO3104 is $$16,000 new on Ebay

Leave a Reply

Your email address will not be published. Required fields are marked *