New Rigol DS1054Z oscilloscope

[CatalinaWOW]

While adding to the obvious, using ANY instrument at the limit of its capabilities should be done with timidity/caution/low confidence.  When you walk on the edge of a cliff there is much to go wrong.

So far that hasn't been much of an issue for me.  I have had much use of the scope well within the boundaries of it's capabilities.  A few cases where the need was so far outside its capabilities that there was no danger of confusion.  And a very few cases where it took some skull sweat to get useful information out of a scope that was obviously one foot over the edge of the cliff.

[McBryce]

I think Scopes are going the way computers did over time. Back in the days of DOS/Windows 3.1, the user had to have a much better understanding of the computer and had to do a lot of things manually. Now computers are used by a huge amount of people where the majority have no idea of what's actually going on in the background.
I notice exactly the same trend with scopes. Back in University, the newest cutting egde scope we had was the Tektronix 2465, which had just been released around that time. There was no Auto buttons, no decoding, no auto-measurements and as for colours... Green is fine for all 4 channels and the limited readouts the scope offered. But we were taught every detail of the scope and how to take measurements. The closest thing to "Auto" was the "Find Trace" button. So you either really deeply understood the instrument or it was pretty much useless.
Today people expect the scope to almost do everything for them and do it perfectly, and I don't think newer enthusiasts are really spending the time required to fully understand how it all works and how to extract the information they want to measure. I really like all  the fancy new options that the Rigol (and other) scopes offer (and yes I really own one), but I still don't trust anything that I haven't configured myself.
I have a feeling that many of the "naysayers" are of the newer generation, who put way too much trust in the scopes abilities and have expectations, that many of the older users don't. A scope is a tool, the capability to take accurate measurements should come from the user.

McBryce.

[coppice]

I think Scopes are going the way computers did over time. Back in the days of DOS/Windows 3.1, the user had to have a much better understanding of the computer and had to do a lot of things manually. Now computers are used by a huge amount of people where the majority have no idea of what's actually going on in the background.
I notice exactly the same trend with scopes. Back in University, the newest cutting egde scope we had was the Tektronix 2465, which had just been released around that time. There was no Auto buttons, no decoding, no auto-measurements and as for colours... Green is fine for all 4 channels and the limited readouts the scope offered. But we were taught every detail of the scope and how to take measurements. The closest thing to "Auto" was the "Find Trace" button. So you either really deeply understood the instrument or it was pretty much useless.
Today people expect the scope to almost do everything for them and do it perfectly, and I don't think newer enthusiasts are really spending the time required to fully understand how it all works and how to extract the information they want to measure. I really like all  the fancy new options that the Rigol (and other) scopes offer (and yes I really own one), but I still don't trust anything that I haven't configured myself.
I have a feeling that many of the "naysayers" are of the newer generation, who put way too much trust in the scopes abilities and have expectations, that many of the older users don't. A scope is a tool, the capability to take accurate measurements should come from the user.

Automation is usually an excellent way to reduce the need for skills and ability. Unfortunately not demanding skills tends to make them wither, and then people can't understand why the automation isn't doing what they expect.

[MrW0lf]

I have a feeling that many of the "naysayers" are of the newer generation, who put way too much trust in the scopes abilities and have expectations, that many of the older users don't. A scope is a tool, the capability to take accurate measurements should come from the user.

Counting photons instead of lambs before sleep? User wo superpowers cannot compensate for handicapped design. My Z bite the bullet because it took away even theoretical possibility to take accurate measurements by skilled user (due to scrapping raw data and messed up statistics, all in the name of nice looking CGI). Now with new scope I can do timing stuff down to sample jitter (3ps RMS).  Needs bit "hacking" but at least possible. I expect this from any scope - delivering to full hardware capability. Even little AD2 tries hard to deliver max it can.

Exact reason why sold my Subaru. Whats the point of symmetrical AWD if cannot turn off stability control? Only car where I ever felt need for some electronic assistance was F1.

[Assafl]

I am not sure I understand the rationalizing behind the need of a scope to deliver the full capability of its hardware. Why?

I admit they need to deliver to specs, but the whole idea of lock keys and having multiple models is that they are selling different capabilities and specs at additional price points.

[MrW0lf]

I am not sure I understand the rationalizing behind the need of a scope to deliver the full capability of its hardware. Why?
I admit they need to deliver to specs, but the whole idea of lock keys and having multiple models is that they are selling different capabilities and specs at additional price points.

Nobody questions hacking DS1000Z for features or bandwidth. I like to push scopes for accuracy. Seems that people with CRO background hardly realize quite insane accuracy that can be achieved based on processing of true unaltered sample points - by means of mathematical / statistical gain. In vertical can do far more accurate than 1% DC even with 8bit ADC (using sampling noise). In horizontal can do picosecond timings with entry level and femtoseconds with better models. Its just like long memory single shot, completely beyond what similar analog bw CRO could do. Maybe someday it will sink in. Or maybe not because less excuse to buy frequency counters and other narrow functionality stuff...

[NorthGuy]

In vertical can do far more accurate than 1% DC even with 8bit ADC (using sampling noise).

According to the specs, the accuracy is about 4% of full scale - 5 bits. I don't think there's any reason for them to understate it, but who knows, I'm not a marketer.

The resolution is 8 bits, so (8 - 5) = 3 bits are already redundant. With oversampling, you can increase resolution further, but this will not improve accuracy.

[ebastler]

In vertical can do far more accurate than 1% DC even with 8bit ADC (using sampling noise).

According to the specs, the accuracy is about 4% of full scale - 5 bits. I don't think there's any reason for them to understate it, but who knows, I'm not a marketer.

The resolution is 8 bits, so (8 - 5) = 3 bits are already redundant. With oversampling, you can increase resolution further, but this will not improve accuracy.

I assume MrWolf was referring to the PicoScope, the only true digital data acquisition system, which probably has better specs. 

Agree that only precision (resolution) can be improved by oversampling, not accuracy -- so MrWolf's terminology was not precise (or accurate? ). I would not say that the additional three bits are "redundant", though. While they do not help with accuracy, they do improve the resolution.

[MrW0lf]

The resolution is 8 bits, so (8 - 5) = 3 bits are already redundant. With oversampling, you can increase resolution further, but this will not improve accuracy.

With oversampling you can improve accuracy also, taking advantage of sampling noise. 0.1% of range no problem with scope DC spec +-3% of full scale, if have access to unaltered data.

[MrW0lf]

I assume MrWolf was referring to the PicoScope, the only true digital data acquisition system, which probably has better specs. 

Why only, also have AD2 - very accurate thing. Suppose same with Teks, GWI etc. Its normal with any scope operating with raw data and proper floating point math. I would gladly have Tek MDO4000C, just ask for address where send to 

While they do not help with accuracy, they do improve the resolution.

Im not native english, help me out what you mean. So after performing calibration and getting values comparable to good DMM its somehow still not accurate, only high res? Even when repeatable etc... Bummer, maybe secretly DMMs are also only high res and show right values just by coincidence?



Heres little earlier test, where AD2 also present. As you can see with proper techniques scopes deliver comparable to simpler DMMs in vertical.

[ebastler]

The resolution is 8 bits, so (8 - 5) = 3 bits are already redundant. With oversampling, you can increase resolution further, but this will not improve accuracy.

With oversampling you can improve accuracy also, taking advantage of sampling noise. 0.1% of range no problem with scope DC spec +-3% of full scale, if have access to unaltered data.

Huh? In my understanding, accuracy is always limited by systematic measurement errors, e.g. correct amplification factors etc.. Oversampling can only improve precision (statistical deviation). See Wikipedia's definition under "accuracy and precision" for example:



EDIT: Saw your further post, where you also mention calibration (in addition to oversampling). That will help, indeed: Use oversampling to improve your precision, then use calibration to improve your accuracy. Then hope that the scope does not drift out of calibration... 

[MrW0lf]

Then hope that the scope does not drift out of calibration... 

Thats good point. Initially was afraid that it will drift away, but actually all you need is proper warmup at least 0.5h. Before publishing tests did re-check on next day with same reloaded cal after warmup. There is still drift when warm but pretty small. So yes, pushing the limits - but performance is repeatable - can be used in practice.
Edit: Note my tests are all done in scope interface, real time, not offline with MatLab etc.

[NorthGuy]

Of course you can improve the accuracy by calibration, but the specifications usually list the accuracy of a calibrated instrument.

There are many factors which affect accuracy. Accuracy determined over a range of conditions will be worse than the accuracy at 25C. Different units of the same scope will have different characteristics and the specs specify the worst case.

If you have an accurate reference, you certainly can perform your own characterizations, which wouldn't be any worse than the characterization by the manufacturer. But I doubt it is better than 1% for DS1054Z.

[bitseeker]

I see you've found rf-loop's work and for those that have not seen it already:
https://www.eevblog.com/forum/testgear/rigol-ds1074z-weird-signal-level-problem/

It just follows on from other work this long sever member did many years ago, I think in his first post on the forum.
https://www.eevblog.com/forum/chat/rigol-ds1000e-series-possible-errorfail-in-sin(x)x-interpolation/

Thanks for those links, tautech. I hadn't seen those threads before (they were prior to my joining).

[stuartmp]

Hi All,

This is my first post to this forum and thread.

I am interested in upgrading my current oscilloscope and I have been looking at the Rigol DS1054Z

I watched Dave's video EEVblog #703 - Rigol DS1054Z Oscilloscope Review Summary


However I'm confused about something you said at 7:25 on this clip.

You say that there is no Software or hardware filter.

I am confused because I read in the manual on page 135 - chapter ch6-21 on Filters
http://www.batronix.com/pdf/Rigol/UserGuide/MSO_DS1000Z_Plus_UserGuide_EN.pdf

It says the MSO1000Z/DS1000Z provides 4 types of filters (Low Pass Filter, High Pass Filter,
Band Pass Filter and Band Stop Filter). The specified frequencies can be filtered by setting the bandwidth.

Can someone please try and clear this up for me as I am used to using a scope with these types of filters.


Thanks

Stuart

[David Hess]

If you switch it off, I assume the scope reconstructs the signal by simple "connect the dots" interpolation of the samples. That, of course, will reduce the apparent amplitude since the peaks are not properly reconstructed.

With one exception, without sin(x)/x interpolation, the peak and RMS values do not change with sampling rate whether aliasing is present or not.  The exception is when synchronous sampling produces a DC or very low frequency result which exceeds the record length but that is not relevant here.

I have little idea whether Rigol's automatic measurements work correctly in this case though.  I suspect not and I know the automatic RMS measurement does not under certain conditions.

[technogeeky]

Hi All,

This is my first post to this forum and thread.

I am interested in upgrading my current oscilloscope and I have been looking at the Rigol DS1054Z

I watched Dave's video EEVblog #703 - Rigol DS1054Z Oscilloscope Review Summary


However I'm confused about something you said at 7:25 on this clip.

You say that there is no Software or hardware filter.

I am confused because I read in the manual on page 135 - chapter ch6-21 on Filters
http://www.batronix.com/pdf/Rigol/UserGuide/MSO_DS1000Z_Plus_UserGuide_EN.pdf

It says the MSO1000Z/DS1000Z provides 4 types of filters (Low Pass Filter, High Pass Filter,
Band Pass Filter and Band Stop Filter). The specified frequencies can be filtered by setting the bandwidth.

Can someone please try and clear this up for me as I am used to using a scope with these types of filters.


Thanks

Stuart

It has the mentioned filters. That video is very, very old. Perhaps they weren't included at the time.

[metrologist]

I see you've found rf-loop's work and for those that have not seen it already:
https://www.eevblog.com/forum/testgear/rigol-ds1074z-weird-signal-level-problem/

I am not seeing this on my scope, but I am feeding 10 MHz pulse. Is this still repeated on current firmware? I may need a faster signal.

[rf-loop]

I see you've found rf-loop's work and for those that have not seen it already:
https://www.eevblog.com/forum/testgear/rigol-ds1074z-weird-signal-level-problem/

I am not seeing this on my scope, but I am feeding 10 MHz pulse. Is this still repeated on current firmware? I may need a faster signal.

Of course can not see. First you need knowledge and after then also tools. Now there is both missing.
How I know both are missing. Just when you tell "but I am feeding 10MHz pulse".  Please explain what was real theory and idea behind this.

[alsetalokin4017]

Hi All,

This is my first post to this forum and thread.

I am interested in upgrading my current oscilloscope and I have been looking at the Rigol DS1054Z

I watched Dave's video EEVblog #703 - Rigol DS1054Z Oscilloscope Review Summary

(video link snipped)

However I'm confused about something you said at 7:25 on this clip.

You say that there is no Software or hardware filter.

I am confused because I read in the manual on page 135 - chapter ch6-21 on Filters
http://www.batronix.com/pdf/Rigol/UserGuide/MSO_DS1000Z_Plus_UserGuide_EN.pdf

It says the MSO1000Z/DS1000Z provides 4 types of filters (Low Pass Filter, High Pass Filter,
Band Pass Filter and Band Stop Filter). The specified frequencies can be filtered by setting the bandwidth.

Can someone please try and clear this up for me as I am used to using a scope with these types of filters.


Thanks

Stuart

Yes, the recent scope firmware does include these filters. However, if you are used to using a "real" scope with these filters you may be less than impressed with Rigol's implementation of them. For example they can only be implemented on the Math channel and I don't think they can be combined with other Math operations.  In my own brief and casual testing they work "OK" but it would be better if someone like MrW0lf could evaluate them thoroughly. But I don't think he has his z-box any more...

[metrologist]

I am not seeing this on my scope, but I am feeding 10 MHz pulse. Is this still repeated on current firmware? I may need a faster signal.

Of course can not see. First you need knowledge and after then also tools. Now there is both missing.
How I know both are missing. Just when you tell "but I am feeding 10MHz pulse".  Please explain what was real theory and idea behind this.

Knowledge is not an absolute thing one has in unity, like a bit. Same with tools, of which I have a variety, but not a 100 MHz signal generator. I may find a few oscillators though. I had not intended to formulate any theory at this point. I was merely trying to duplicate your instrument settings with a signal I had available, and did not observe the same results, for which I postulated my own guess but have no other reason other than it is a different frequency than your test.

You seem offensive in your reply, so just be assured that I am not out to challenge or discredit you. I am just interested in learning about my scope and try to maintain an unbiased and objective perspective about that.

[ebastler]

I see you've found rf-loop's work and for those that have not seen it already:
https://www.eevblog.com/forum/testgear/rigol-ds1074z-weird-signal-level-problem/

I am not seeing this on my scope, but I am feeding 10 MHz pulse. Is this still repeated on current firmware? I may need a faster signal.

The strange sin(x)/x implementation remains unchanged in the current firmware. Just confirmed this again using a 100 MHz input signal: Everything looks the same as in rf-loop's earlier tests. The amplitude in dot mode changes when sin(x)/x is enabled; line display mode (in single-shot mode with sin(x)/x OFF) produces a smooth interpolation instead of connect-the-dots.

You will indeed need a higher frequency input signal to demonstrate this. These effects only become significant when there are just a few samples per signal period. With four channels enabled (250 MHz sampling frequency), use at least 50 MHz to make the effect clearly visible.

[metrologist]

I do note that interpolation is not disabled and is most apparent in the overshoot/ringing artifacts, and these increase with interpolation on. As I was looking again I recalled a 100 MHz source I have an also duplicate the signal level drop at 100ns sweep.

[rf-loop]

This is example about how it need work. (This is not made using Rigol)
In this test example used  samplerate  and  scope input sinewave frequency as: fNYQUIST/1.82     (equivalent as with 1GSa/s and 275MHz )



This image is just for imagine continuously running and Sinc off. (Of course not Rigol)
Rigol do (roughly) this kind of Sinc if we think there is dot middle of each line. If these imagined points are used for Sinc (draw via sample points rule) we can see very close what Rigol "looks like" do. (who knows how it really do but it looks like it in my some old tests)




Same running but now dots. (NO Sinc)
(note: last singlle acq dots bit more bright)




Then single shots. In all next three images sampled data is same for show that it really draw via real sample points.




 ADC sample dots.






Lines Sinc (Sin(x)/x) off.  (draw exactly via ADC sample dots.)






Lines Sinc on. (draw exactly via ADC sample dots.)
(pity it do not show these real sample dots here)



This is how oscilloscopes need do Sinc. (and all serious oscilloscopes do, look Keyshit, look Tektronix, look Siglent, look R&S, look Goodwill.

Note that Sinc is good for sine or some other "smooth" waveforms. It is NOT best in all cases for fast square-rectangle types of signals.  Later when I have time I will show some these things but not in this thread for avoid too deep off topic.

[Fungus]

You say that there is no Software or hardware filter.

Can someone please try and clear this up for me as I am used to using a scope with these types of filters.

It sounds like you have a very specific requirement in mind.

I'd try to get hands-on experience with a 'scope before spending any money on it. Manuals and sales brochures never give the full picture (this applies to all 'scopes, not just the DS1054Z).

At the very least hunt down a youtube video of the feature being used.