Rigol DS1000Z series (ds1054z, ds1074z, ds1104z and -s models) Bugs/Wish List

[Fungus]

So clearly it cannot be used anything like normal DSO. But then manual should reflect it. Customer is not to be expected reading forums for weeks when buying equipment.

It's honestly never occurred to me to try to accurately measure a rise time (or anything else) without being zoomed in on the area of interest.

I also don't try to measure 0.2V signals with my multimeter set to the 200V range, so... 

[Howardlong]

This discussion clears up a mystery for me; why does Rigol include a delay feature and misleadingly refer to it as delayed sweep when it is not. .

I belive this is a nomenclature problem. On Agilent DSOs, for example, until about six years ago, the zoom mode was referred to as delayed sweep mode, a throwback to the days of CROs. Rigol's heritage is from making low end scopes for Agilent, explaining the apparent anomaly.

[MrWolf]

It's honestly never occurred to me to try to accurately measure a rise time (or anything else) without being zoomed in on the area of interest.

Maybe you moved directly from CRO to "screen sampling" DSO. I barely used CRO and learned most stuff on normal DSO. On normal DSO you always look at the sampling rate and decide the actual precision of your auto-measurements based on that. For example at 1GS/s you would have high confidence that phases are spot on down to at least +-1ns (single shot), and transients to analog bw. This gives you massive edge over classical CRO - having "big picture" and precise measurements on a single screen w/o zooming in/out all the time.
This is a big trap for young players or when someone moves from normal DSO to "screen sampling" DSO.
So first step for fixing it for Z would be stop shooting at the moon and just disable auto-readings based on very simple formula I gave:
IF Signal.Transient <= (Timebase / 25) THEN Reading.disable()
...or at very least display <= in front of reading.
Of course real fix would be enabling memory-source measurements...

BTW - can it be assumed that info on this thread gets to Rigol or must report directly?

Edit: One more way to make it work. RUN mode - do as before, just display < or <= before readings or disable entirely. STOP mode - switch to signal analysis based on actual sample memory. No hit on realtime performance & valid data with single button press.

[Karel]

BTW - can it be assumed that info on this thread gets to Rigol or must report directly?

You can safely assume that they don't read anything on any threads on this forum...

Report it directly. If you don't receive any confirmation within a week, call them and ask for the status.
In my experience that is the only way to get your Rigol representative (Rigol North America or Rigol Europe) to really check and confirm a bug
and report it internally to the Rigol engineers in China. Also, do a follow-up check after 3 months. And after 6 months. And after...

[MrWolf]

Report it directly.

On reporter might be not enough. If anyone feels it is time for actual
sample memory based measurements for masses!
then act NOW.

https://www.rigolna.com/contact-us/

http://www.rigol.eu/contact-us/

EU exec is Wolfgang Bartels.
Look article:
http://www.elektroniknet.de/markt-technik/messen-testen/deutschland-ist-ein-wichtiger-potentieller-markt-fuer-rigol-87737.html
Google-translated:
https://translate.google.com/translate?sl=de&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.elektroniknet.de%2Fmarkt-technik%2Fmessen-testen%2Fdeutschland-ist-ein-wichtiger-potentieller-markt-fuer-rigol-87737.html
Quote:
"We place much emphasis on reliability, quality, service and sustainability."

Time to deliver.

[Karel]

I already have a dedicated line with the red telephone directly to Rigol...

[MT]

I got a bit annoyed by Mr Bartelemo when he said:

Customers - whether private, in research, development or training - who already have a device from us know that the quality is very good and the error rate is low.

[David Hess]

It's honestly never occurred to me to try to accurately measure a rise time (or anything else) without being zoomed in on the area of interest.

I also don't try to measure 0.2V signals with my multimeter set to the 200V range, so... 

There is merit to this but on the DSOs I have used including my ancient Tektronix 2440, accurate measurements are made even with voltage and time resolution significantly below optimum for the measurement.  The number of significant digits is also reduced to match the measurement accuracy and a warning is given if the sample rate or amplitude is too low.  If anybody is interested, the measurement algorithms used in older Tektronix DSOs are documented in detail.

Tektronix also went to considerable design effort and cost (1) to ensure that measurements and processing never truncated or degraded the sampled data.  For instance all of their old DSOs which use 8 to 10 bit digitizers, use 16 bit acquisition memories for processing and *then* a separate display memory.  So for instance averaging or high resolution modes result in a higher bit resolution waveform stored in the acquisition memory which the measurement routines then act on unlike the Rigiol DSOs.

(1) Back in the 1980s when they did this, memory was not cheap.  When you see those tiny record lengths (not so tiny when you have delayed sweep) on the first Tektronix DSOs, they all use 16 bit words so their 1k or 4k records are actually 2k or 8k bytes and as far as I know they have always done this.

This discussion clears up a mystery for me; why does Rigol include a delay feature and misleadingly refer to it as delayed sweep when it is not. .

I belive this is a nomenclature problem. On Agilent DSOs, for example, until about six years ago, the zoom mode was referred to as delayed sweep mode, a throwback to the days of CROs. Rigol's heritage is from making low end scopes for Agilent, explaining the apparent anomaly.

I am not so sanguine because Rigol has repeatedly done this going back to their early DSOs where they confused envelope mode and peak detection.  To my mind, Rigol premeditatively lies in their marketing and documentation.  And I notice that they never make a "mistake" in the opposite direction which Tektronix has done occasionally in the past.

On normal DSO you always look at the sampling rate and decide the actual precision of your auto-measurements based on that. For example at 1GS/s you would have high confidence that phases are spot on down to at least +-1ns (single shot), and transients to analog bw.

So first step for fixing it for Z would be stop shooting at the moon and just disable auto-readings based on very simple formula I gave:
IF Signal.Transient <= (Timebase / 25) THEN Reading.disable()
...or at very least display <= in front of reading.
Of course real fix would be enabling memory-source measurements...

But that would make the apparent and fictional performance of the instrument match reality!  Rigol cannot do that!

Tektronix actually did something like this with one of their early DSO designs.  If you have an analog trigger and compare it to the sampling frequency, then it is possible to detect when aliasing is likely.  The problem was while everybody's DSO suffered from the same aliasing problems, having an indicator on the DSO which alerts the user that aliasing is likely to be present makes the DSO look worse compared to the competition so this useful feature was dropped.

BTW - can it be assumed that info on this thread gets to Rigol or must report directly?

Rigol did not care years ago when I pointed out mistakes (lies) in their DS1000E specificaitons and documentation.  Why would they care now?

[Fungus]

On reporter might be not enough. If anyone feels it is time for actual
sample memory based measurements for masses!
then act NOW.

Isn't it better to understand that this is the way DSOs work.

That way you won't be confused when using strange DSOs (nb. It's not just Rigol that does this....)

Fixing it would require making all those low-end DSOs more expensive. You say you wouldn't mind if it all went slower in the name of accuracy but secretly you would. You'd be in here complaining about speed, and how your old 1990s Tek was faster.

[MrWolf]

Isn't it better to understand that this is the way DSOs work.
Fixing it would require making all those low-end DSOs more expensive. You say you wouldn't mind if it all went slower in the name of accuracy but secretly you would. You'd be in here complaining about speed, and how your old 1990s Tek was faster.

Fungus, I'm starting to get seriously worried... You are a fungus, not a mushroom. Who keeps you in the dark and feeds ...? Rise up, break from the chains. You can do it, fight for your freedom  You remind me of beautiful russian women who are ok with drunk husbands beating them. It's not ok! There is a better life out there  Don't suffer just because of having knobs!

Current Pico 2204A, costing just 109€ works just like my old 2205MSO and would wipe the floor with Z on low freq - because despite it has tiny 8kS memory, it actually uses all of it, not some 1.2kS interpolated down to a 300S(!!!) garbage!
What if we compare to actual counterpart 2408B, 100MHz, 4ch, 128MS. It's 128,000,000S vs 300S that scope can actually use for calculations...

But I know one other scope that you have to use just like Z - take measurements with cursors and scroll screen back and forth.
It's.... Velleman HPS140 

Now about fixing part. Did adding "memory" source option to FFT make Z more expensive? Did you receive 39600$ bill?
Or what is cost of "<" adding sign before measurement value?
Also there's this option:
RUN mode - do as before, just display < or <= before readings or disable entirely.
STOP mode - switch to signal analysis based on actual sample memory.
No hit on realtime performance & valid data with single button press.
And not only measurements, it would make decoding work!

So it's all about choices and not treating your customers like a mushrooms... Not so much about cost...
Yes there would be some programming involved but they already managed FFT so it's a broken in path...

And do not worry for me so much. I bought this box for very specific task at VHF freq.
According to Screen sampling rate = 1 / (Timebase / 25)  formula there is small window of timebases where Z can be used as more-less normal instrument: 5,10,20ns. So I'm very well covered with old Picos doing precision stuff below 25MHz and Z-box at VHF.

... but due to my pro line of work I feel that making society aware of s*it programming and it's implications is important so why not to have little cockfight now and then... 

[Fungus]

Isn't it better to understand that this is the way DSOs work.

Fungus, I'm starting to get seriously worried...

I know this is going to scare the absolute crap out of you, Wolfie, but I'm going to say it anyway:

Every voltage reading you've ever seen on your DS1054Z screen is probably wrong.

It's wrong because you didn't take the trouble to maximize the waveform on screen first. The DS1054Z only has an 8-bit ADC and it works in screen space. Think about all the implications of that! All those newbies out there who don't understand it.

Let's demand that Rigol switch to 16-bit ADCs right now to avoid any more confusion.

[MrWolf]

Let's demand that Rigol switch to 16-bit ADCs right now to avoid any more confusion.

Well yea since Rigol does not have DC offset like normal DSOs in this same price range:
https://www.picotech.com/library/application-note/using-analog-offset-to-maximize-oscilloscope-resolution
Not to mention much better basic DC accuracy in the first place for 2xxx series: ±3% of full scale ±200uV

But when you do not have proper tools it's worth having a brain. I actually got interested in the problem yesterday and if there has not been blonde force majeure forcing me off the lab corner...
Will report later in this thread:
https://www.eevblog.com/forum/testgear/help-please-understanding-rigol1054z-dc-rms-measurement/

[BravoV]

... but due to my pro line of work I feel that making society aware of s*it programming and it's implications is important so why not to have little cockfight now and then... 

Hmmm ... pro line of work ? ... programming ? ... just because your claim and self declaration ?

Just don't believe it until there is a proof, isn't that fair ?

[JPortici]

Fungus, I'm starting to get seriously worried

i see you're still new around here

[Fungus]

Let's demand that Rigol switch to 16-bit ADCs right now to avoid any more confusion.

Well yea since Rigol does not have DC offset like normal DSOs in this same price range:

How is the "price range" the same if you need to add a PC/laptop just to get a display and some horizontal/vertical/trigger controls?

PC: Remind us again the price of a 4-channel, 100Mhz Picoscope. I forgot.

[MrWolf]

Just don't believe it until there is a proof, isn't that fair ?

Is my formula Screen sampling rate = 1 / (Timebase / 25) not correct according to actual tests done?
Does it not characterize device and have practical value when estimating actual accuracy of readings?

Would this solution not work on specific hardware?
RUN mode - do as before, just display < or <= before readings or disable entirely.
STOP mode - switch to signal analysis based on actual sample memory.
No hit on realtime performance & valid data with single button press.
And not only measurements, it would make decoding work!

How is the "price range" the same if you need to add a PC/laptop just to get a display and some horizontal/vertical/trigger controls?

Well you really have to be a mushroom not to find some out of fashion PC essentially for free to tuck under the desk.

PC: Remind us again the price of a 4-channel, 100Mhz Picoscope. I forgot.

If we talk non-hacked, diff is not that big. If we talk hacked, low-end VHF and only analog work, maybe yea (that's by I bought Z-box). If we talk lower freq then Picos just wipe the floor with Z for lower price. It seems to me often people go for high freq just for "prestige" and do not actually need it. Especially amazing are audio people buying Z if there is Analog Discovery 2 with 14bit ADC for less.

[JPortici]

If we talk non-hacked, diff is not that big. If we talk hacked, low-end VHF and only analog work, maybe yea (that's by I bought Z-box). If we talk lower freq then Picos just wipe the floor with Z for lower price. It seems to me often people go for high freq just for "prestige" and do not actually need it. Especially amazing are audio people buying Z if there is Analog Discovery 2 with 14bit ADC for less.

only a fool would pay for the option you can hack for free

[MrWolf]

only a fool would pay for the option you can hack for free

Had any really free lunch lately? It reminds one occasion with my good friend. He's sort of "financial life hacker" and tries to get stuff for free always when possible. One time he bought large good looking set of kitchen utensils from presentable dude with BMW X5... in a dark poorly lit parking lot... Dirt cheap, much cheaper than in a store  Well you can guess what he discovered at home... All it was basically a moulage
So what can be learned from that - if someone gives you some hackable stuff (knowingly!) for cheap price - it's high probability that it's actual value of this stuff + margin for manufacturer + margin for ...

[ebastler]

PC: Remind us again the price of a 4-channel, 100Mhz Picoscope. I forgot.

If we talk non-hacked, diff is not that big. If we talk hacked, low-end VHF and only analog work, maybe yea (that's by I bought Z-box). If we talk lower freq then Picos just wipe the floor with Z for lower price.

https://www.picotech.com/oscilloscope/2000/picoscope-2000-overview
100 MHz, 4 channels - $1235
50 MHz, 4 channels - $659

Is that the pricing we are talking about? Then I don't get your "wipe the floor for lower price" comment, even looking at a stock DS1000Z in comparison. And if I compare a hacked DS1054Z to a $1235 Picoscope then - "maybe yea", as you put it, there might be just a tiny price advantage indeed...

[MrWolf]

Then I don't get your "wipe the floor for lower price" comment, even looking at a stock DS1000Z in comparison.

Yolo, I'm really glad I sort of play simulated poker here, it would be nightmare otherwise...

Did you failed to notice that 32768Hz signal produced correct auto-risetime-measurement (accounting analog bw, it's 25MHz scope) from 2ns to 2us timebase, and spot-on period from 5us to 5ms. On Pico that is. Did you register, it did show garbage only 1 time: never.

On Z it produced auto-risetime from 5ns to 50ns, and almost correct period only on 5us timebase. Correct period accounting for analog bw is never produced. For mere 32768Hz signal! So there is "bandgap" filled with garbage 100ns to 5ms risetime-wise and all settings besides 5us are garbage period-wise for specific signal auto-measurement. Only workaround is cursors and zoom, like Velleman HPS140. But o boy it has some sexy knobs eg "lots of hardware"  These knobs will be polished to the bone while "fools" who "cannot count their money" just sit and watch correct readings on their extremely boring instrument 

[David Hess]

Isn't it better to understand that this is the way DSOs work.

That way you won't be confused when using strange DSOs (nb. It's not just Rigol that does this....)

Only cheap toy DSOs do this.

Fixing it would require making all those low-end DSOs more expensive. You say you wouldn't mind if it all went slower in the name of accuracy but secretly you would. You'd be in here complaining about speed, and how your old 1990s Tek was faster.

It is not entirely clear how Rigol is producing their high update rate.  If the digitizer is producing a display sized histogram, then it really has a 1200 point record length or whatever the graticule width is most of the time and Rigol are cheap lying bastards.  If the graded indexed display is created by the processor (this seems more likely to me now), then Rigol are still cheap lying bastards.

Either issue is do to the programming and not the hardware; there is no conflict between accurate automatic measurements and high acquisition rate.

Also there's this option:
RUN mode - do as before, just display < or <= before readings or disable entirely.
STOP mode - switch to signal analysis based on actual sample memory.
No hit on realtime performance & valid data with single button press.
And not only measurements, it would make decoding work!

So it's all about choices and not treating your customers like a mushrooms... Not so much about cost...
Yes there would be some programming involved but they already managed FFT so it's a broken in path...

I do not think this is feasible for marketing reasons.  Having two different measurement modes would reveal that display base measurements are broken.

Well yea since Rigol does not have DC offset like normal DSOs in this same price range:
https://www.picotech.com/library/application-note/using-analog-offset-to-maximize-oscilloscope-resolution

Doesn't the Rigol position control shift the ground level?  Position controls usually operate as offset controls with the position signal added before the digitizer.  The difference is whether the signal is added before or after the attenuators.

[Fungus]

Isn't it better to understand that this is the way DSOs work.

That way you won't be confused when using strange DSOs (nb. It's not just Rigol that does this....)

Only cheap toy DSOs do this.

I mentioned several times that people are free to pay more money and get this problem "fixed". Going on an on about it is like complaining that Ford Fiestas can't do 200mph. Technically true, but not going to make headline news or scandalize anybody.

So really this discussion is a bit of a yawn.

As I said, it honestly never occurred to me to look at a rise time without zooming in on the wave. Isn't it 100% natural to zoom in and look at the rising edge when you measure it? Make sure there's nothing weird there? 

If Picoscope owners aren't doing that then it says a lot about their user interface.

[Fungus]

https://www.picotech.com/oscilloscope/2000/picoscope-2000-overview
100 MHz, 4 channels - $1235
50 MHz, 4 channels - $659

Is that the pricing we are talking about?

I suspect Wolfie's not being entirely honest, not even with himself.

Well you really have to be a mushroom not to find some out of fashion PC essentially for free to tuck under the desk.

Sure. I might even be able to find a matching CRT monitor to complete the setup (and get a real analog 'scope heat output to make it feel like the old days).

OTOH I might be able to put something useful in that space.

Had any really free lunch lately? It reminds one occasion with my good friend. He's sort of "financial life hacker" and tries to get stuff for free always when possible. One time he bought large good looking set of kitchen utensils from presentable dude with BMW X5... in a dark poorly lit parking lot... (snip stupid story)

The DS1054Z can be made into exactly the same device as the DS1104Z with serial decoders, advanced triggers, 24Mb memory, etc.

The DS1104Z with all those options costs $1100.

No, the DS1104Z isn't worth $1100. If I was spending $1100 I might get a GW-Instek or something like that.

The DS1104Z is definitely a bargain for $400 though.

[MrWolf]

Doesn't the Rigol position control shift the ground level?  Position controls usually operate as offset controls with the position signal added before the digitizer.  The difference is whether the signal is added before or after the attenuators.

Screenshot attached how it works on Rigol.

[nctnico]

Well you really have to be a mushroom not to find some out of fashion PC essentially for free to tuck under the desk.

Those things aren't free at all. I used to have 4 PCs running under my desk. Now only one and my electricity bill is considerably lower. Not to mention noise and replacing fans every few years.