New Rigol DS1054Z oscilloscope

[i4004]

even rigol says "break the rules"!
 



btw. why would anyone be surprised to see this scope has less bandwidth when u use 4 or 3 channels? 


you have that data in scope datasheet on rigol web

Code: [Select]

Analog channel: 1 GSa/s (single-channel),
500 MSa/s (dual-channel), 250 MSa/s (three/four-channel)

and simillar goes for memory depth. more channels=less bandw. and memory depth.

you're not looking at a system that has 4 totally separate processors where each has 1gs samplign rate. you're looking at system with 4 inputs but just one part that is doing all that 1gs processing....so when you use all 4 inputs bandwidth falls to 1/4th.

beyond 25mhz on 4 ch sure u can use it, but you don't know what you'll be looking at ie which alias of what.... 


long time ago i had some nice illustrations of aliasing that would really make this very clear to pascal_sweden (and maybe others) but who's gonna find those now? 
but he could just imagine a grid of vertical lines(ie sampling) layed over the sinusoid signal(signal to be sampled), and then what will happen when few sinusoid periods span just 2 vertical lines and you never know at which signal amplitude you'll take that sample.
and how you can't really reconstruct that signal (you sampled) because you lack denseness of that vert. grid...

edit/added/found that aliasing illustration...


alan2k, if you saw this

you'll see dave uses 1/10th rule of thumb, ie for example 250ms/s equals 25mhz analog bandwidth...

pascal, you would probably use "insert image" tags and then put image url (that you just uploaded to forum) in it...ie

(i've put

Code: [Select]

https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/?action=dlattach;attach=112665

in img tags)

[Fungus]

Has anyone here actually proof that the DS1104Z can not do 100 MHz on all 4 channels at same time?

It *can* do 100Mhz on all 4 channels. The argument here is over how useful that number is to an electrical engineer.

Question: With 200 MSa/s, how can you trust that the signal you're looking at is really 100MHz?

(Answer: You can't, because aliasing becomes a massive problem as you approach the Nyquist limit).


For as long as I can remember the rule of thumb has been "usable bandwidth is one tenth of the sample rate". With a 1 GSa/s 'scope this rule limits you to 100MHz with one channel, 50MHz with two channels and 25MHz with three or four channels. These frequencies can be relied on. Above these frequencies? You need to start doubting.

For me the analog frontend should just limit the max. frequency to make sure that there is no conflict with the corresponding ADC specs

Well of course it should do it, but they don't because it's difficult (ie. expensive) and you don't seem willing to pay.

Me? I'm gonna relax and enjoy my 4-channel, top-build-quality, sub-100Mhz-with-four-channels-enabled, 300 Euro oscilloscope...  (thanks, Rigol!)

[Mark_O]

and that it can sample at it's highest rate (without fear of aliasing) that XX MHz with all N channels ON. The DS1104Z can not - period.

But with all 4 channels on, 250MSa/s, Nyquist is still 125 MHz, right?  Am I missing something?

Well, yes, and no.   

You're correct about Nyquist.  But the point you've missed (and you're not the only one) that Marmad has tried to make is that for that to work, you can't have significant spectral components above that Nyquist limit.  And on the Rigol, you do.  They roll off very slowly, above their -3dB reference point.  (And BTW, some scopes DO use higher-order filters on the front-ends, which roll off frequencies much more quickly than the Rigols do.  Rigol probably doesn't do that, because good, quality filters that don't negatively impact waveform integrity are expensive.)

We know that for sure, since swperk took the time to confirm that the hacked 1054z is down only 10dB at 400 393 MHz!  That means you're guaranteed to have aliases rolling over into your passband.  (Any time you have spectral content between 125-450 MHz.)

Now someone clever is sure to interject something like, "But why can't you still use it, if you know you don't have any frequencies above the Nyquist-limit, in the signal you're probing?"  And if you do know that, then sure, that's OK.  The problem is that you don't know that in general, and in principle, never can (unless you're using a different instrument to check, which kind of defeats the purpose).  And if what you're probing are digital signals with 'sharp edges', you'll have a lot of high frequency harmonics that won't be obvious unless you look at a spectral analysis.

You want to be using an instrument that you know you can rely on for the results it's providing.  Otherwise, you can wind up wasting a lot of time, or come to completely erroneous conclusions.  And to do that with the 1000z Rigols, you have to know where its limits are.  And work within them.  That's all.

[Mark_O]

Has anyone here actually proof that the DS1104Z can not do 100 MHz on all 4 channels at same time?
Or are the speculations just that it can not, because it is supposed to have no fancy anti-aliasing filter or whatever it might be called?

There's no need, since it's a fact, not 'speculation'.

For me the analog frontend should just limit the max. frequency to make sure that there is no conflict with the corresponding ADC specs (sample rate) in the path beyond,

And there you go (again), with what you think things should be.   

and my understanding is that a 100 MHz scope has proper filtering by default to take care of this requirement in combination with > 200 MS/s ADC.

Your understanding is incorrect.  I.e., you're wrong.   Sorry.    Which is why people are posting to clarify... so you don't lead others astray, in your ignorance.

100 MHz scope == proper filter in the analog frontend to limit max. frequency to 100 MHz, period.

Please notify Rigol immediately!     They need to do an immediate recall!

Isn't the anti-aliasing filter just a band pass filter?

No, it's not a BPF, it's an LPF.  It rolls off only the high end.

How complex can that be?

Well, it can be fairly complex, as it turns out.  You can't filter amplitude without also affecting the phase, of components below the frequency of the amplitude you're trying to influence.  So there are tradeoffs in any filter design.  You pick the right one for the type of job it's intended for, and pricing considerations.

It's pretty obvious you've never taken a course on filtering theory, but that's not stopping you from pontificating about it.   

[Creep]

Mark_O: can't one just turn off the other channels to check how the signal looks in one channel mode if you get a feeling something might be off? Then if you have confirmed it, you can safely turn on the rest of the channels knowing that what you see is the accual signal. Of course, this assmums that one is triggering off the channel in question.
Would there be any accual problems going about it this way?
P.S. I know that that would be a hassel and I'm not saying it's convenient. Nor am I saying I understand this stuff all that well, it's just an idea I got while reading the previous comments.

[David Hess]

Mark_O: can't one just turn off the other channels to check how the signal looks in one channel mode if you get a feeling something might be off? Then if you have confirmed it, you can safely turn on the rest of the channels knowing that what you see is the accual signal. Of course, this assmums that one is triggering off the channel in question.
Would there be any accual problems going about it this way?
P.S. I know that that would be a hassel and I'm not saying it's convenient. Nor am I saying I understand this stuff all that well, it's just an idea I got while reading the previous comments.

I like using an old combination analog and digital storage oscilloscope just for this reason.  If I suspect a problem do to aliasing or anything else, I can push one button and see the analog waveform at full bandwidth confirming my sanity.

To be fair however, I have hardly ever found a discrepancy even on an old DSO and peak detection works about as well.

[pascal_sweden]

I was actually thinking about the same. If you need to measure 4 channels in parallel for a detailed problem analysis, just check each measure point separately first on the first channel (other channels not used) in FFT mode, to check if there are frequencies above the Nyquist frequency. After having verified this for your 4 test points, it is safe to hook up the 4 channels to your test points, forget about exotic anti-aliasing filters and enjoy your 300 USD investment =)

Here is my understanding of the Rigol scope and its limitations: The Rigol DS1104Z has limitations in the filtering of the analog front end implying that you only have 25 MHz bandwidth in reality when using ALL 4 channels. When you treat the Rigol DS1104Z scope as a 25 MHz scope, and only use it for debugging designs up to 25 MHz, you can safely rely that the high frequencies that could exist in your design, are properly suppressed with the low-order low-pass filter in the analog front end of the Rigol scope.

But even when you are only debugging 25 MHz designs with this scope, you will still have the risk of higher frequencies right? I mean, even if you use it as a 25 MHz scope, there could be higher frequencies that give false readings? Or do I miss understand here?

Actually, regarding high frequencies and noise. If your design itself is within frequency limitations, any other frequencies in the signal path must come from interference and noise. But if the amplitude is very low, do they still impact a lot? Can someone shed a light on actual frequencies of interference signals and noise, their respective amplitude, and their impact? Can noise and interference beat the Nyquist frequency in terms of frequency and amplitude? =)

What about the other Rigol series, and their reliable bandwidth?
For the Rigol DS1104Z: ALL 4 channels, 25 MHz is reliable bandwidth.
For the Rigol DS2302A: ALL 2 channels, what is the reliable bandwidth?

[Fungus]

What about the other Rigol series, and their reliable bandwidth?
For the Rigol DS1104Z: ALL 4 channels, 25 MHz is reliable bandwidth.
For the Rigol DS2302A: ALL 2 channels, what is the reliable bandwidth?

The names have the format DSXYZWS

The formula is:  X*100/W

[marmad]

But even when you are only debugging 25 MHz designs with this scope, you will still have the risk of higher frequencies right? I mean, even if you use it as a 25 MHz scope, there could be higher frequencies that give false readings? Or do I miss understand here?

No, you are correct - but in general, you shouldn't be attempting to measure signals that contain such high frequncy components (just as, for example, you wouldn't want to be measuring the same signals with only 1 channel on, but a sampling rate of 250MSa/s due to your time base/memory depth settings).
 
Another option would be to enable the 20MHz bandwidth limiter of each channel when using 3 or 4 channels. The DS1000Z appears to be using AD5207s (EDIT: and/or perhaps they're using the HC4053s) to alter the RC networks of the low pass filter - and I'm assuming that they apply the 20MHz bandwidth limiting there.

For the Rigol DS2302A: ALL 2 channels, what is the reliable bandwidth?

200MHz (2GSa/s) is perfectly fine on the DS2000A. 300MHz starts to push the filtering boundaries a bit - and as many of us have mentioned in other threads, the DS2302A starts to run into the same possible problem (although much less severely) when you run it with both channels on (1GSa/s - 500MHz Nyquist).

[Fungus]

But even when you are only debugging 25 MHz designs with this scope, you will still have the risk of higher frequencies right? I mean, even if you use it as a 25 MHz scope, there could be higher frequencies that give false readings? Or do I miss understand here?

No, you are correct - but in general, you shouldn't be attempting to measure signals that contain such high frequncy components (just as, for example, you wouldn't want to be measuring the same signals with only 1 channel on, but a sampling rate of 250MSa/s due to your time base/memory depth settings).

With four channels on the DS1054Z might be a better scope than the DS1104Z because it has a much harder cutoff filter (above 50MHz).

[David Hess]

Another option would be to enable the 20MHz bandwidth limiter of each channel when using 3 or 4 channels. The DS1000Z appears to be using AD5207s  for gain adjustment of the amplifiers in the front end - and I'm assuming that they apply the bandwidth limiting there.

Are you suggesting they used a para-phase amplifier (Gilbert multiplier with the outputs crossed) or something similar for gain adjustment controlled by the DC output from the AD5207?  That is a pretty old technique which I would not expect in a modern DSO.

The current trend is to do this type of calibration in software.  Even old DSOs do this.

[W7NGA]

sorry .. I just had to chuckle. I own several Tektronix 2465B's that will do 4-channel and 400 Mhz acquisitions in their sleep!

I paid $200 for each one and they look as new. yes, there are applications where a DS1054Z would better serve, but I don't have to worry about aliasing or finding a decent 50-ohm termination!

dan W7NGA

[Bert Camper]

W7NGA, can you do a single shot 400MHz on 4 channels with your Tektronix 2465B? The answer is no.
This is comparing apples and pears.

--Bert

[marmad]

sorry .. I just had to chuckle. I own several Tektronix 2465B's that will do 4-channel and 400 Mhz acquisitions in their sleep!

Sure... but sometimes that's not the most important concern. Aside from the plethora of things that a DSO can do that an analog scope can't do, the sheer size of the screens on the 2465Bs would have me pulling my hair out at some point (and I know, because I own an analog Tek too).

[pascal_sweden]

With four channels on the DS1054Z might be a better scope than the DS1104Z because it has a much harder cutoff filter (above 50MHz).

Is this really true? My understanding is that DS1054Z and DS1104Z are completely identical hardware wise. So the analog filtering should be exactly the same. The only difference is a SW configuration, but how can that SW configuration impact the behavior of the analog front end?

[W7NGA]

geez .. where is the tongue-in-cheek emoticon when I need it!

I enjoy my Agilent MSO7014B you see in the background when it suits the task at hand.

let's be real .. if you need single-shot 400 Mhz capture you won't shouldn't be reaching for the DS1054Z 

as far as the screen size, it's amazing how much great engineering was accomplished with the eye-strain and mental anguish caused by staring at a small CRT.

I seem to remember having no problems at all staring at my Tek 465 all day and night.

rehashing the analog versus digital oscilloscope comparison is ... sophomoric at best. good engineers use good tools that fit the application. 

then again, would you trust someone with such a messy bench? 

dan W7NGA

[marmad]

as far as the screen size, it's amazing how much great engineering was accomplished with the eye-strain and mental anguish caused by staring at a small CRT.

You use the tools you've got. Although I'm fairly sure if you had to give up either your Tek 2465s or the Agilent 7000, I know what would go. 

rehashing the analog versus digital oscilloscope comparison is ... sophomoric at best.

Perhaps... but maybe so is coming on this thread to post another "look at what my bargain analog scope can do" comment. 

What is actually more interesting - seeing as how this new, $400, 4-channel DSO has ~80% of the waveform display area of your Agilent 7000, ~25-50% of the intensity levels of your 7000, ~30% of the waveform update rate of your 7000, and ~3x more memory than your 7000 - is how much that Agilent 7000 cost?

[carpelux]

I noticed a strange ting today.

Received my new DS1054Z a couple of days ago, and today was the first day I had a chance to test it out.

As my aim was to free it from its chains i thought I should test it's performance before and after applying the upgrade. The result surprised me!

I tested two things, a sweep with a 1V pp sine wave until it showed 0.7V (-3db), and the rise time using a Jim Williams Pulse Generator from free_electron.

I used a approx 1m RG58, a Rigol DG4162 (low z) and at the scope end termination with a tee and a 50 ohm resistor. The pulse generator was connected directly to the tee and the terminating resistor.

Before upgrade to DS1104:
Sine sweep: 110 MHz
Rise time: 2.5 nS

After upgrade:
Sine Sweep: 120 MHz
Rise Time: 2.4 nS

To me it seems that at least my sample of the DS1054Z was not bandwidth limited to 50 MHZ out of the box.

[W7NGA]

I don't recall ever besmirching the Rigol scopes. I think the DS1054Z and DS2000 series are amazing.
I was reacting to what $400 can buy if you need high-bandwidth and multiple channels, with relaxed triggering requirements.

as I recall, I bought the Agilent 7000 scope with a few days engineering consulting income. it was a bargain at the time and afforded me accomplishing my objectives.
it has never locked up, crashed, smoked, or caused me irrepressible anguish .. other than paying for it. 

perhaps I simply felt that after 23 pages, this thread could use an injection of analog humor.

dan W7NGA

[marmad]

I was reacting to what $400 can buy if you need high-bandwidth and multiple channels, and relaxed triggering requirements.

Understood. But you have to imagine that for many of us that don't live in the US, the options for finding and buying reasonably-priced, decent condition (or even fixer-uppers) used-test gear can be much more difficult.

perhaps I simply felt that after 23 pages, this thread could use an injection of analog humor.

Always welcome - although not always 100% translatable in text form. 

[David Hess]

as far as the screen size, it's amazing how much great engineering was accomplished with the eye-strain and mental anguish caused by staring at a small CRT.

Those small CRTs were the "retina" displays of the era and one of the reasons Tektronix continued to use monochrome CRTs with external LCD color shutters in their high end TDS DSOs.  EEVBlog Dave mentioned Agilent doing something similar with high resolution monochrome CRTs in their older DSOs for the same reason.

Their DSOs which used the same 5" diagonal CRT from their 100 MHz and faster oscilloscopes were often driven with 10 bits of horizontal and vertical resolution for 1024 x 1024 which comes out to 240 dpi horizontal and 315 dpi vertical.  The 10 bits of vertical resolution was not even always wasted since the 7854 uses a 10 bit digitizer and most of the other DSOs have averaging modes.  More commonly some of the resolution was dropped but a respectable 1024 x 512 or 512 x 512 was still common.

I seem to remember having no problems at all staring at my Tek 465 all day and night.

I feel the same way now when using my 7904 (500 MHz 24 kV 5") despite its small CRT size compared to my 7603 (100 MHz 15 kV 6.25") or any of my other 100 MHz oscilloscopes with 5" CRTs even though I acquired the 7904 accidentally.

[David Hess]

W7NGA, can you do a single shot 400MHz on 4 channels with your Tektronix 2465B? The answer is no.
This is comparing apples and pears.

Another reason it is a poor comparison, which is pointed out occasionally on the TekScopes@yahoogroups.com email list, is that the 2465B and similar oscilloscopes were high end instruments of their time and still would be considered so if they were still manufactured.  That you can find used ones in good condition now for hundreds of dollars just reflects low demand and great uncertainty.

A better comparison to a 2465B would be a modern DSO costing $10,000 like a DPO4034B and I would certainly desire the later if I could get it for the same price as a used but working 2465B.  Of course you can calibrate (with difficulty) and repair a 2465B.  The DPO4034B is more like an Apple product; when it breaks, just throw it away and buy a new one.

[Fungus]

As my aim was to free it from its chains i thought I should test it's performance before and after applying the upgrade. The result surprised me!

It's easy to switch back/forth between the two. Install the Ultra Sigma utility and do ":SYSTem:OPTion:UNINSTall" to go back to DS1054Z, ":SYSTem:OPTion:INSTall XXXXXXXXXyourkeywithnohyphensXXXXXXXXX" to install a key.

To me it seems that at least my sample of the DS1054Z was not bandwidth limited to 50 MHZ out of the box.

Doesn't sound right to me.

[Deckert]

...that implies that it's BW frequency response is at least XX at -3db - and that it can sample at it's highest rate (without fear of aliasing) that XX MHz with all N channels ON. The DS1104Z can not - period. Since it does not adhere to these normal expectations, this would be no different than breaking one of the other implied specifications: e.g. that a 300MHz DSO has a 450MHz BW (even though that BW is at -9db).

Does anybody know if the DS1104Z supports ETS (Equivalent Time Sampling)? Yes, I know it's then only valid for repeating waveforms, but then again, a lot of the early DSOs did just that - eg. operated in ETS all the time in order to fulfill the analog front-end bandwidth requirement, while single-shot bandwidth was typically much less.

For example, I have the HP 54602B 150MHz DSO. It can display a 150MHz sine wave on all four channels at the same time, provided they're repeating waves, like that of a sine wave generator. The scope itself only as a 20MSa/sec ADC, so realistically one-shot captures are limited to no more than 2MHz on one channel.

I'd be happy with ETS for the 4-channel, 100MHz usage scenario.

--deckert

[marmad]

Does anybody know if the DS1104Z supports ETS (Equivalent Time Sampling)?

No, none of the Rigol UltraVision DSOs do.