My sorry attempt at a review - DS1104Z

[jadew]

I wish I knew how to add a youtube video, so here's the link:

The video is pretty much a fail, but it might at least offer a bit more info on the DS1000Z series. I did try to make it good, but that's all I could put together from a few hours of footage.

The quality sucks too... 

In case the video gets automagically embedded in here, here's the description:
"Sorry for my english and my crude editing skills.

The review didn't came out as I expected because I found it a bit hard to express myself in english with out thinking for a second about what I was about to say.

Anyway, I hope it at least gives prospective buyers an idea on what it can and can't do. I tried to cover the things that I found lacking in the other reviews that are currently out and it's made from the perspective of a DS1052E owner.

Thanks for watching"

[electronics man]

I recently got this scope I realy do love it ther are some problems but realy for the price it is amazing I meen4 channels for £445 is just incredible it is a very good scope rigol have set a new benchmark for spec. Even a very good agilant can't beet it memory depth it is realy amazing. It would be nice if the sample rate were higher though.

[jadew]

Yeah, I enjoy it as well and yes, it could have used a second ADC, 250 Msps is a bit low.

[marmad]

Thanks for your video! BTW, it's totally fine - and it does offer some new info on the DS1000Z. Not enough owners are publishing any in-depth info about the DSO.

A couple of points and questions:

1) The DS2000 allows measurements on MATH - both horizontal, vertical, and all - so if any of that is missing on the DS1000Z, I suspect it's a bug:





2) As mentioned in other threads on the forum, most of the new generation of cheaper DPOs (like the Agilent X-Series, Rigol UltraVision, etc) DON'T have Alt-Triggers. I theorized that perhaps the reason it's not included is because it would be difficult to support intensity-grading of two distinct time bases at a reasonable speed. In some situations, you can at least achieve a stable display of two non-time-correlated waveforms by using one of the dual source triggers of the Rigol - such as the Delay Trigger - which allows you to trigger after edges (or other conditions) from each channel with an adjustable </>/<>/= time between them. This image shows triggering on two uncorrelated waveforms:





3) Does the DS1000Z allow MATH functions using a REF waveform as a source? This is something missing from the DS2000 - and it would be nice if they would implement this function.


4) In another thread about the DS1000Z, there was some discussion of the fact that the DS1000Z appears to, at many voltage scale settings, to digitally scale (like a camera digital zoom) a smaller set (< 200) of the ADC sample values to fit the screen. This can result in a display that looks like it has horizontal stripes running through it (as in the following image). Unfortunately, Rigol fails to include this info in their datasheets (the DS2000 scales the 500uV setting). Any chance you might test this with a graded signal (such as AM signal) and see how many of the main scale settings (not the fine scale, just the coarse settings) this might be visible on?





Once again - thanks for your efforts!

[jadew]

Hey,

1) The DS2000 allows measurements on MATH - both horizontal, vertical, and all - so if any of that is missing on the DS1000Z, I suspect it's a bug:

In that case, I really hope it's a bug, but I'm worried it could be because the CPU can't handle it. I think I forgot to put that piece in the video, but I was complaining about how slow is everything (including saving a screenshot). It's possible that doing additional processing on top of the already slow MATH channel, might slow it down to a crawl.

I actually took this screenshot with the Source menu opened, but it didn't show up in the screenshot (I noticed other elements don't show up in screenshots either - like the channel labels).

3) Does the DS1000Z allow MATH functions using a REF waveform as a source? This is something missing from the DS2000 - and it would be nice if they would implement this function.

Nope, indeed it would have been very nice.

4) In another thread about the DS1000Z, there was some discussion of the fact that the DS1000Z appears to, at many voltage scale settings, to digitally scale (like a camera digital zoom) a smaller set (< 200) of the ADC sample values to fit the screen. This can result in a display that looks like it has horizontal stripes running through it (as in the following image). Unfortunately, Rigol fails to include this info in their datasheets (the DS2000 scales the 500uV setting). Any chance you might test this with a graded signal (such as AM signal) and see how many of the main scale settings (not the fine scale, just the coarse settings) this might be visible on?

I don't have a proper AM modulated signal, but I did take a look at some noise and indeed, the issue you're talking about is apparent on the lowest voltage scale (1mV/div).

I couldn't reproduce the exact thing you have in your screenshot (which looks like digital zoom), however I got something similar, but I don't think it's proof of digital zoom but rather an intensity grading issue. The reason for that is because there are visible spikes in my screenshots that are smaller in amplitude than the region with the same color grading, while the fact that most spikes are about the same amplitude could be due to the repeating nature of the noise:






Now... this one is the only one I could get that I'm positive it has digital zoom:



And a little zoom inside it:

[Dongulus]

Your English speaking is fine and you are perfectly understandable. Hearing yourself speaking always sounds worse than what someone else hears.

Thanks for the review. I've had my eye on this series of scopes and it helps to see one in action. I've had my concerns with the usability of the UI and it appears to be okay. It's good to know that it does have some quarks when there is a lot of menus, measurements, and multiple waveforms on screen.

[jadew]

@ Dongulus,

Thank you for the encouragement.

I'm not sure if this particular bit of information is apparent in the video, but I'm very happy with it and I think it's a great oscilloscope. It does fine with all 4 channels active, the problem seems to be when it's doing extra stuff.

If I were you, I'd hold off for a bit longer from buying one right now and see if the little issues it has will be fixed with a firmware update or a new hardware revision.

[marmad]

In that case, I really hope it's a bug, but I'm worried it could be because the CPU can't handle it. I think I forgot to put that piece in the video, but I was complaining about how slow is everything (including saving a screenshot). It's possible that doing additional processing on top of the already slow MATH channel, might slow it down to a crawl.

Well, even on the DS2000, getting the Measurements menu to switch to the MATH channel (change color) is a little erratic. After you've turned on the MATH function (BTW, they won't work for FFT) sometimes you have to first press one or two CH buttons and then press the MATH button to get the measurements menu to switch. So if you haven't tried this yet, try turning on MATH then alternate between pressing CH1, CH2, and MATH button. Even if the DS1000Z can't handle doing ALL measurements for MATH, you'd think it could handle a maximum of 5.

I don't have a proper AM modulated signal, but I did take a look at some noise and indeed, the issue you're talking about is apparent on the lowest voltage scale (1mV/div).

I couldn't reproduce the exact thing you have in your screenshot (which looks like digital zoom), however I got something similar, but I don't think it's proof of digital zoom but rather an intensity grading issue. The reason for that is because there are visible spikes in my screenshots that are smaller in amplitude than the region with the same color grading, while the fact that most spikes are about the same amplitude could be due to the repeating nature of the noise:

Thanks for testing that. From other tests it seems as if the DS1000Z is doing it (digital zoom) a smaller amount at other scales too - but if it's not noticeable in the displayed image (which always gets upscaled by at least a factor of 2x for the 400 vertical pixels anyway), at least it won't be visually distracting.

[jadew]

Well, even on the DS2000, getting the Measurements menu to switch to the MATH channel (change color) is a little erratic. After you've turned on the MATH function (BTW, they won't work for FFT) sometimes you have to first press one or two CH buttons and then press the MATH button to get the measurements menu to switch. So if you haven't tried this yet, try turning on MATH then alternate between pressing CH1, CH2, and MATH button. Even if the DS1000Z can't handle doing ALL measurements for MATH, you'd think it could handle a maximum of 5.

Tried it, didn't work.

Thanks for testing that. From other tests it seems as if the DS1000Z is doing it (digital zoom) a smaller amount at other scales too (< 100mv) - but if it's not too noticeable in the final image (which always gets upscaled by a factor of 2x for the 400 vertical pixels anyway) it at least won't be distracting.

I see, so the 2 pixels I was seeing in there are nothing out of the ordinary.

I took a better look at the screenshot you showed me and while the general feel is that of pixelation both vertically and horizontally, it still shows fine detail on the margins, which might mean that they're not doing any digital zoom and it's in fact just a case where their intensity grading algorithm fails.

[nctnico]

Thanks for testing that. From other tests it seems as if the DS1000Z is doing it (digital zoom) a smaller amount at other scales too (< 100mv) - but if it's not too noticeable in the final image (which always gets upscaled by a factor of 2x for the 400 vertical pixels anyway) at least it won't be visually distracting.

It would be nice to test again with a sine wave. If the signal is being scaled up then it should be visible. Perhaps it takes different frequencies to make the effect visible. Rigol could use oversampling on longer timebase settings.

[marmad]

I see, so the 2 pixels I was seeing in there are nothing out of the ordinary.

Well, the stripes are caused by scaling the sample values. When digital scaling is present, it's apparent (at the smaller scales) on the DS2000 when you turn off Vectors and look at the Dots. Interpolation plus intensity-grading tends to hide it more.

EDIT: Perhaps they've gotten more clever with how it's implemented on the DS1000Z series.

I took a better look at the screenshot you showed me and while the general feel is that of pixelation both vertically and horizontally, it still shows fine detail on the margins, which might mean that they're not doing any digital zoom and it's in fact just a case where their intensity grading algorithm fails.

On the DS2000, 250 ADC values are ALWAYS mapped to 10 divisions (the 8 you see +/- 1 above/below) - UNLESS: you are at the 500uV scale (which is 1mv x2) - or any fine vernier scale (e.g. 2.2V) - which is mathematically derived (as opposed to a fixed resistor on the PCB) - i.e. digital zoom. This chart posted by sync indicates that the DS1000Z is doing some mathematical scaling at other settings and based on other factors (such as number of channels on).

[electronics man]

Talking About the slow maths functions I was doing FFT on a sine wave that was increasing in frequency fairly slow I was just increasing the freq using the encoder on my function gen and the the FFT was realy slow to update I it took about 5 or 6 seconds to update each time. I would like to see if rigol will correct this with new firmware. I can't comment on other things as this is my first dso   

[Pasky]

Is the 1000z basically a 2000 series with a signal generator?

[marmad]

Is the 1000z basically a 2000 series with a signal generator?

No, the DS2000 has more bandwidth, more memory, more features, faster hardware, etc. - but they share a certain amount of the hard and soft tech - as do all of the UltraVision series of DPOs.

[Mark_O]

Is the 1000z basically a 2000 series with a signal generator?

No, not at all.  It's more of a scaled-down 2000, less expensive, with twice the channels. 

Either can have a signal-gen, as a purchase-time option (-S models).

[Mark_O]

No, the DS2000 has more bandwidth, more memory, more features, faster hardware, etc....

Hey!  With your DS2000 more, more, more mantra, you left out "more money" and "less channels".   

[electronics man]

Yeah and 24Mpts is still massive compared to other scopes whitch may only have a few meg

[marmad]

Hey!  With your DS2000 more, more, more mantra, you left out "more money" and "less channels".   

Well, unless you believe more of something is cheaper than less, that seems obvious.    As far as the efficacy of those 4 channels go, it really depends on what your work is. The effective reliably-interpolated maximum analog bandwidth of the DSO when using 4 channels is ~25MHz (well, ~31.5MHz if you're charitable), and it's clear from the OPs video that the UI is a bit on the slow side. Definitely handy for certain situations (like 4-channel protocol decoding) - but it's not a 4-channel 100MHz BW DSO.

[cybermaus]

I'm not sure if this particular bit of information is apparent in the video, but I'm very happy with it and I think it's a great oscilloscope. It does fine with all 4 channels active, the problem seems to be when it's doing extra stuff.

Glad to hear you say so, I had the impression the device was being bashed a little in this thread, and I am quite happy with it myself.

I come directly from a Analog CRO, so I may not know precisely what the norm is for DSO or what to expect. And I agree, maybe the UI could be faster. Especially when you do math or if it has a difficult trigger.

But I am quite happy with it, I am convinced it is a step up from most other DSO targeted to the hobbyist like DS1054 or Owon SDS or Hantek clones. (not sure about Owon TDS)

Sure a DS2000 would be better, and even better yet would be two DS2000 or a DS4000. But where does it end? If you compare it, compare it against its correct pricerange, and I still believe it then is the best choice. @Marmad: Don't go telling people a entry level Volkswagen is junk just because you own a top level Audi.

[marmad]

Sure a DS2000 would be better, and even better yet would be two DS2000 or a DS4000. But where does it end? If you compare it, compare it against its correct pricerange, and I still believe it then is the best choice. @Marmad: Don't go telling people a entry level Volkswagen is junk just because you own a top level Audi.

Well, I'm not trying to do that! Sorry if it came across that way; I think the DS1000Z is a great bargain - and I'm a fan of Rigol's entire UltraVision line. But with new products in general - and new Chinese products specifically (since most manufacturers from there seem to be a little reticent about admitting faults or weaknesses), it's important for prospective buyers to know as much as possible to make informed choices - and that includes the bad as well as the good. There are plenty of things wrong with the DS2000 too - just read awhile over in the other thread.

I was never making a value comparison of the DS1000Z to the DS2000 - I was only interested to know what FW changes/additions/omissions Rigol had made when they developed the newest addition to the UltraVision line. Each product fits a different segment of the market.

[Hydrawerk]

Definitely handy for certain situations - but it (DS1000-Z) is not a 4-channel 100MHz BW DSO.

Isn't there an equivalent sampling feature? With that you can have a 100MHz oscilloscope with only 250MSa/s. Like that old HP 546000 series. Well, they say nothing about it... http://www.rigol.com/prodserv/DS1000Z/

[nctnico]

Definitely handy for certain situations - but it (DS1000-Z) is not a 4-channel 100MHz BW DSO.

Isn't there an equivalent sampling feature? With that you can have a 100MHz oscilloscope with only 250MSa/s. Like that old HP 546000 series. Well, they say nothing about it... http://www.rigol.com/prodserv/DS1000Z/

250Ms/s is enough for 100MHz but it depends on the interpolation whether you get a pretty picture or not. The more pressing problem is aliasing and I'm pretty sure you'll see lots of aliasing on this scope with 4 channels on. An interesting tests would be to feed the scope 62.5MHz (and a little bit) and display 1, 2 and 4 channels.

[marmad]

250Ms/s is enough for 100MHz but it depends on the interpolation whether you get a pretty picture or not.

By "pretty picture" do you mean a reasonable facsimile of the waveform?    The sampling rate is almost certainly too low for reliable sin(x)/x - and 2.5 samples per period are not enough for an accurate linear reconstruction, pretty or not.

[jadew]

250Ms/s is enough for 100MHz but it depends on the interpolation whether you get a pretty picture or not. The more pressing problem is aliasing and I'm pretty sure you'll see lots of aliasing on this scope with 4 channels on. An interesting tests would be to feed the scope 62.5MHz (and a little bit) and display 1, 2 and 4 channels.

Any particular reason why you picked that frequency? I can't do 62.5, but I can do 64.

As for equivalent sampling - I haven't found that yet.

[cybermaus]

exactly 1/4 of 250 sample.

If you are *exactly* exactly, it would not show any artifact, but near the sampling frequency, so if you have 62.49 (or of the actual sampling is 250.01) some interesting things may show.