EEVblog #522 - Rigol DS1000Z Oscilloscope Quick Look

[EEVblog]

Dave had a very quick play with the new 4 channel Rigol DS1000Z series oscilloscope with dual channel ARB function generator at the Electronex 2013 exhibition.
And meet some EEVblog viewers.

[stormbr]

Thanks for sharing the new brands of Rigol on Electronex 2013. 

[grego]

I wish they had the MSO4000 on display!  Dang nab it!

[hikariuk]

Hmm.  Colour me interested...depending on what the UK pricing ends up being.

[crt]

1mV / div = Really ?

please see my post about 1074z [no one mention it]
https://www.eevblog.com/forum/testgear/rigol-ds1074z-inside-picture/15/

[EEVblog]

1mV / div = Really ?
please see my post about 1074z [no one mention it]

What post?
What are you saying about the 1mv/div range?

[dr.diesel]

I wish they had the MSO4000 on display!  Dang nab it!

Agreed, I'm tired of waiting.  Seriously thinking about a Hameg HMO3000.

[grego]

I wish they had the MSO4000 on display!  Dang nab it!

Agreed, I'm tired of waiting.  Seriously thinking about a Hameg HMO3000.

diesel hopefully ou can be patient a little longer.  In talking with tequipment they expect it to drop her in the US "middle of 4th quarter" - so we've got (if that's accurate) 2 more months to go.  Since I just harassed them last month about it I'll probably wait until next month before I harass them again.

[peter.mitchell]

please see my post about 1074z [no one mention it]

What post?

I think either, both are related:
https://www.eevblog.com/forum/testgear/rigol-ds1104z/
https://www.eevblog.com/forum/testgear/rigol-ds1074z-inside-picture

[crt]

What post?
What are you saying about the 1mv/div range?

Rigol hide 500uV/div feature, but unfortunately already find by EEVBLOG member 

[Hydrawerk]

64 levels of intensity is OK. It is the same as on my DSOX2002A

[vk6zgo]

Nice to see them using a analog TV  field rate video signal as a test.

As an unreconstructed analog OF, I'd like to think it's because of my consistent whinging over the past few years over the total mess lesser DSOs make of such a test.

In any case,it does a really good job of a single field,& even when Dave wound it out to show 3 or more,it was still usable!

Well done,Rigol!

[evanh]

Got mine today, a DS1074Z-S.  It looks the part.  Four channels and deep capture memory is a base model for me, I've skipped looking at anything less.

It feels sturdy.  Nice display.  Lots of settings ... not always staying as set though.  Menus are responsive.  The separate group of keys on the left for selecting measurements seemed extraneous at first but I think I might like them now.  Noise levels look good.  All in all an excellent purchase.


Here's some niggles that could be sorted with firmware updates:

 - The transition from full sweep redraws to the progressive drawing, or trending, that occurs at slow timebases.  It kicks in at too slow a setting, imho.  The refresh times feel unresponsive when waiting for a full sweep redraw.

 - There is a second concern, and more annoying, with the trending method - It only trends post trigger!  If the trigger is in centre of display then you end up waiting a half display of sweep time before any activity occurs.  This can be minutes!  This, combined with the time based (rather than display based) trigger position, kind of makes for a lot of fiddling to get immediate visualisation of acquisitions.


Updates:

27 Oct 2013
Trivial gripe:  CLEAR, AUTO and SINGLE keys really do seem to be useless.  Why would one want to clear all traces?  I've used auto setting before, it never gets what I want and it totally loses the settings I had set.  A separate button to set single trigger mode, say what?, there is a simple to use trigger mode button straight below that!

And on the small gripes list is the HELP pages.  There is some broken descriptions in there, eg:  Memory depth in the acquisition menu, it repeats a single channel description three times over and doesn't describe any multi-channel info.

Not so trivial:  In the acquisition menu, what does Sin(x)/x setting do?  I don't see any effect on traces.  The help makes a vague reference to interleaving.  This setting is one of those settings, I mentioned previously, that doesn't stay set.  Maybe it doesn't really need a user setting at all?

Actually, the Help system has been helpful.  I've definitely seen much worse.  Some kudos are deserved there.

When Zoom'd it would be nice to know the number of sample points covered by the zoom box.

... Just noticed another one to be wary of:  Peak acquisition mode places it's two min-max values in successive sample points in the trace.  This allows the memory depth to be shown as unchanging between Normal and Peak modes.  I had wondered why that parameter didn't change.  The problem with this configuration is when zooming in to look at the fine detail, instead of seeing a continuous min-max block fill, you see a triangle wave at half the sample rate toggling between the upper and lower values.  A sudo alias of sorts.  This is a behaviour that Peak mode is meant to avoid.

Once you are aware of this though, it could be used.  The question becomes:  Does the min and max alternately stored samples represent a progression in time?  If so, then there is extra info in them than just being a collective min-max of the whole interval.

... Just delivered the scope to a new temporary home (I might not be very detailed from now.) ... has a use for the waveform generator ... bugs galore!  :-O  First bug is either the outputs are not meant to be active on power up or the lamp in the Source button does not light to indicate an output is active on power up.  It syncs up when toggling either of the output on/off settings.

Second waveform generator bug is when changing pretty much any setting with modulation enabled causes it to set the depth to 100% and uses the depth setting to set the RMS level.  All without indication.  Turning modulation off then on corrects it.

Third waveform generator bug is when selecting waveform type Pulse the modulation is not immediately disabled as it seems to require.


Evan

[evanh]

I just got my old scope back (Made in 2001, and cost ten times the price) and had a look at the way it handles the slow timebase issue - It rolls the display (I couldn't actually remember that detail) rather than trending it.  That way it can both give a 100% live progress of the sweep and still cleanly line up any trigger that might occur.

Also, it kicks in at 50ms/div rather than the somewhat slower DS1000Z's 500ms/div (I think).

I feel the DS1000Z can do better.  It's display speed seems comparable, if not quicker, otherwise.


28 Oct 2013
We tested USB file storage today.  There's some oddball bugs with that too.  FAT16 and FAT32 are supported.  exFAT not supported.  Some controller chips causes the scope's USB to stop functioning, driver crash likely.  Happens immediately upon device insertion, not even the USB connected symbol appearing.  Had to power cycle to recover normal USB function.  Haven't got any controller names sorry.  Also tried an MP3 player, with the same result.

Card readers work but limited in what slot is accessed.  Again, depends on the controller chip.  One small SD-only reader we tried only supports one card inserted at a time, in this case both slots (SD and uSD) work.

Tried a SATA HDD slotted in an external USB cradle.  Worked a treat.  Although only the first partition showed up, and, obviously, had to be FAT32.  T'was using Windoze to perform the formatting - it wouldn't allow FAT32 volumes larger than 32 GB.  32 GB works fine with the scope.

There is various formatting arrangements, including with FAT32, that the scope refuses to deal with.  When this happens you see a quick one line message on the display that says something about filesytem init failed.  No volume appears.

30 Oct 2013
Broke the 32 GB Windoze limit by formatting with GParted.  USB file storage now tested to work with a 230 GB volume.  Which was measured to have 16 KB clusters.  And since 32 KB/cluster was doable under FAT16 one can pretty much assume that up to 512 GB is viable in FAT32.


Evan

Continued on page three at - https://www.eevblog.com/forum/blog/eevblog-522-rigol-ds1000z-oscilloscope-quick-look/msg325751/#msg325751

[rf-loop]

exFAT not supported. Is it bug? Really?

As long as this is true: "Microsoft has not released the official exFAT file system specification, and ..so on",   imho, there is not any reason use this in serious lab equipments.

[marmad]

- There is a second concern, and more annoying, with the trending method - It only trends post trigger!  If the trigger is in centre of display then you end up waiting a half display of sweep time before any activity occurs.  This can be minutes!  This, combined with the time based (rather than display based) trigger position, kind of makes for a lot of fiddling to get immediate visualisation of acquisitions.

Not sure where you're appropriating the word 'trending' from (2D charts?) but:
1) The DSO can't display pre-trigger data until it's acquired a trigger (it's a given of the way triggering systems work on a DSO). If you want to have less pre-trigger display (and less of a wait with slow horizontal scale settings), then just adjust the trigger position - it's the same on every DSO.
2) If the DS1000Z series is like my DS2000 series, then you can just switch from normal Y-T acquisition to Roll mode via the Horizontal -> Timebase menu setting, if that's what you prefer.

Trivial gripe:  CLEAR, AUTO and SINGLE keys really do seem to be useless.  Why would one want to clear all traces?  I've used auto setting before, it never gets what I want and it totally loses the settings I had set.  A separate button to set single trigger mode, say what?, there is a simple to use trigger mode button straight below that!

You might not realize it now, but CLEAR is very often handy - especially when having multiple REFerence traces, buses, etc. Also, the SINGLE key is NOT the same thing as the button to set single-trigger mode: once it's in single-trigger mode and a trigger occurs - the DSO goes to STOP. The SINGLE button takes it back to WAITing for the next single trigger - the mode button does NOT.

Not so trivial:  In the acquisition menu, what does Sin(x)/x setting do?  I don't see any effect on traces.  The help makes a vague reference to interleaving.  This setting is one of those settings, I mentioned previously, that doesn't stay set.  Maybe it doesn't really need a user setting at all?

I'm not trying to be unnecessarily mean here, but if you don't know what Sin(x)/x interpolation is on a DSO, you probably need to do a bit of reading on the subject. Such as this.

... Just noticed another one to be wary of:  Peak acquisition mode places it's two min-max values in successive sample points in the trace.  This allows the memory depth to be shown as unchanging between Normal and Peak modes.  I had wondered why that parameter didn't change.  The problem with this configuration is when zooming in to look at the fine detail, instead of seeing a continuous min-max block fill, you see a triangle wave at half the sample rate toggling between the upper and lower values.  A sudo alias of sorts.  This is a behaviour that Peak mode is meant to avoid.

This is how ALL DSOs handle Peak Acquisition - look at the chart from Tektronix:



No offense, but I think you need to do some basic reading on DSOs and their operation.

[evanh]

As long as this is true: "Microsoft has not released the official exFAT file system specification, and ..so on",   imho, there is not any reason use this in serious lab equipments.

That was more an observation really.  I presume this is not special to the DS1000Z but is common to all Rigol scopes and prolly most others too.  I'd be interested know.

[evanh]

- There is a second concern, and more annoying, with the trending method - It only trends post trigger!  If the trigger is in centre of display then you end up waiting a half display of sweep time before any activity occurs.  This can be minutes!  This, combined with the time based (rather than display based) trigger position, kind of makes for a lot of fiddling to get immediate visualisation of acquisitions.

Not sure where you're appropriating the word 'trending' from (2D graphics?) but:
1) The DSO can't display pre-trigger data until it's acquired a trigger (it's a given of the way triggering systems work on a DSO). If you want to have less pre-trigger display (and less of a wait with slow horizontal scale settings), then just adjust the trigger position - it's the same on every DSO.

My old scope achieves what I'm talking about.  I described it in my second post.

I've used the word "trending" here to label the progressive process of adding new samples to the displayed traces as the new data becomes available in a left to right filling of the display.  This is distinct from the normal full screen refresh that occurs after a capture period is complete.

2) If the DS1000Z series is like my DS2000 series, then you can just switch from normal Y-T acquisition to Roll mode via the Horizontal -> Timebase menu setting, if that's what you prefer.

Yah, fiddly and loses other settings.  Not the greatest solution.

Trivial gripe:  CLEAR, AUTO and SINGLE keys really do seem to be useless.  Why would one want to clear all traces?  I've used auto setting before, it never gets what I want and it totally loses the settings I had set.  A separate button to set single trigger mode, say what?, there is a simple to use trigger mode button straight below that!

You might not realize it now, but CLEAR is very often handy - especially when having multiple REFerence traces, buses, etc.

Like I said, a trivial gripe.  Not everyone will agree, obviously.

Also, the SINGLE key is NOT the same thing as the button to set single-trigger mode: once it's in single-trigger mode and a trigger occurs - the DSO goes to STOP. The SINGLE button takes it back to WAITing for the next single trigger - the mode button does NOT.

The STOP/START key performs that function.  The SINGLE key is completely pointless.

Not so trivial:  In the acquisition menu, what does Sin(x)/x setting do?  I don't see any effect on traces.  The help makes a vague reference to interleaving.  This setting is one of those settings, I mentioned previously, that doesn't stay set.  Maybe it doesn't really need a user setting at all?

I'm not trying to be unnecessarily mean here, but if you don't know what Sin(x)/x interpolation is on a DSO, you probably need to do a bit of reading on the subject. Such as this.

So, we are talking about filling in sample points in the stored trace that weren't actually sampled because the hardware interleaving means the sampler was on another channel?

Or maybe it's talking about trace rendering to display?  Shouldn't that be in the display setting menu rather than the acquisition setting menu?

... Just noticed another one to be wary of:  Peak acquisition mode places it's two min-max values in successive sample points in the trace.  This allows the memory depth to be shown as unchanging between Normal and Peak modes.  I had wondered why that parameter didn't change.  The problem with this configuration is when zooming in to look at the fine detail, instead of seeing a continuous min-max block fill, you see a triangle wave at half the sample rate toggling between the upper and lower values.  A sudo alias of sorts.  This is a behaviour that Peak mode is meant to avoid.

This is how ALL DSOs handle Peak Acquisition - look at the chart from Tektronix:
...
No offense, but I think you need to do some basic reading on DSOs and their operation.

Have a careful re-read of what I've said.  Maybe I need to be clearer.  Ask me what I meant if something is confusing.

[marmad]

My old scope achieves what I'm talking about.  I described it in my second post.

Yes, you described a Roll mode combined with triggering - how does that work exactly? The DSO is displaying pre-trigger data while rolling? And what happens when a trigger happens?

The STOP/START key performs that function.  The SINGLE key is completely pointless.

Strange, then Rigol has redefined those buttons on your DS1000Z because they act differently on the DS2000. The RUN/STOP button puts the DSO into AUTO mode - it does NOT reset SINGLE mode - so the SINGLE button on my DSO is far from pointless.

So, we are talking about filling in sample points in the stored trace that weren't actually sampled because the hardware interleaving means the sampler was on another channel?

Or maybe it's talking about trace rendering to display?  Shouldn't that be in the display setting menu rather than the acquisition setting menu?

It's about converting the sampled data to the most accurate rendering of the waveform possible. On my DS2000 it's under the Display Menu - but as is the case with many parameters - there could be an argument made for placing it in another menu category (maybe a space issue?).

Have a careful re-read of what I've said.  Maybe I need to be clearer.  Ask me what I meant if something is confusing.

No, you were perfectly clear. You wrote 'Just noticed another one to be wary of...' as if you thought Rigol was doing something unusual in it's Peak Detect implementation - or else you assumed that other posters didn't know how Peak Detect functioned. You also wrote: 'This is a behaviour that Peak mode is meant to avoid.'- which is incorrect.

[evanh]

My old scope achieves what I'm talking about.  I described it in my second post.

Yes, you described Roll mode - which the Rigol does perfectly well (as mentioned).

It's not a separate non-triggering mode.  It's integral and doesn't require any adjustments and doesn't impose extra limits or exceptions.

The STOP/START key performs that function.  The SINGLE key is completely pointless.

Strange, then Rigol has redefined those buttons on your DS1000Z because they act differently on the DS2000. The RUN/STOP button puts the DSO into AUTO mode - it does NOT reset SINGLE mode - so the SINGLE button on my DSO is far from pointless.

Ah, my inexperience with the Rigols is showing here.  Strange behaviour for the START to force auto triggering but that would be a good explanation.  I don't have it with me so can't confirm right now.

Hmm, I regularly stop a Normal trigger mode when inspecting a deep capture.  Then restart it again after completing my examination.  I presume pressing the START key will flip it to Auto triggering instead of just resuming the Normal triggering ...

No, you were perfectly clear. You wrote 'Just noticed another one to be wary of...' as if you thought Rigol was doing something unusual in it's Peak Detect implementation

Yes!

You also wrote: 'This is a behaviour that Peak mode is meant to avoid.'- which is incorrect.

We differ on opinion there.

[marmad]

It's not a separate non-triggering mode.  It's integral and doesn't require any adjustments and doesn't impose extra limits or exceptions.

Yes, I re-edited my comment above when I realized you were talking about something different than Roll mode. Now I assume you mean a Scan mode (from left to right - Roll mode always displays from right to left) and I'm curious what that DSO is actually displaying if there's been no trigger - and what happens to the display when a trigger occurs.

Ah, my inexperience with the Rigols is showing here.  Strange behaviour for the START to force auto triggering but that would be a good explanation.  I don't have it with me so can't confirm right now.

Hmm, I regularly stop a Normal trigger mode when inspecting a deep capture.  Then restart it again after completing my examination.  I presume pressing the START key will flip it to Auto triggering instead of just resuming the Normal triggering ...

On the DS2000, pressing RUN/STOP while stopped in Normal mode will just re-start Normal mode.

No, you were perfectly clear. You wrote 'Just noticed another one to be wary of...' as if you thought Rigol was doing something unusual in it's Peak Detect implementation

Yes!

...which, as I pointed out with the Tektronix diagram (which shows a triangle wave at high magnification of a Peak Detect acquisition) is not the case at all.

You also wrote: 'This is a behaviour that Peak mode is meant to avoid.'- which is incorrect.

We differ on opinion there.

Any acquisition mode that does post-processing on the samples for a specific effect is likely to show artifacts when magnifying the waveform - this is not behavior which Peak Detect could ever avoid.

[pinkysbrein]

It is bloody trivial to handle the corner case correctly in the display code ... it's just piss poor programming, as usual on oscilloscopes.

[evanh]

Yes, I re-edited my comment above when I realized you were talking about something different than Roll mode. Now I assume you mean a Scan mode (from left to right - Roll mode always displays from right to left) and I'm curious what that DSO is actually displaying if there's been no trigger - and what happens to the display when a trigger occurs.

My old scope has no Scan mode, only Refresh (whole screen update) and Roll (right to left).  But the Roll is integral to normal triggering and all other related settings.  It's purely just an automatic rendering method for slower timebases - 50ms/div and slower.

On the DS2000, pressing RUN/STOP while stopped in Normal mode will just re-start Normal mode.

So, it re-arms AUTO and NORMAL but not SINGLE mode?  That's better but still odd.

...which, as I pointed out with the Tektronix diagram (which shows a triangle wave at high magnification of a Peak Detect acquisition) is not the case at all.
...
Any acquisition mode that does post-processing on the samples for a specific effect is likely to show artifacts when magnifying the waveform - this is not behavior which Peak Detect could ever avoid.

Um, um, nooo...!  The Tek example is terrible.  I didn't look carefully at it first time around but I now see that method is done incorrectly.  The two stored samples, of one interval, are treated as separate intervals of time from a rendering point of view.  But they don't correspond to the equivalent shape through time of the original signal.

They are really nothing more that a min-max of one interval.  They should be treated as such when displayed.

[marmad]

Um, um, nooo...!  The Tek example is terrible.  I didn't look carefully at it first time around but I now see that method is done incorrectly.  The two stored samples, of one interval, are treated as separate intervals of time from a rendering point of view.  But they don't correspond to the equivalent shape through time of the original signal.

They are really nothing more that a min-max of one interval.  They should be treated as such when displayed.

So, according to you, Tektronix and Rigol are both doing Peak Detect the same way - yet 'incorrectly'? Sorry, but you're clearly the one who is incorrect - about how Peak Detect acquisitions are processed and displayed. The samples ARE separated by intervals of time - as all samples are.

If you truly think there is a "correct" method that corresponds to your belief of how it should work, please provide supporting documentation of ANY DSO manufacturer doing it that way.

[evanh]

I don't need any example to know a zig-zag rendering of min-max values is a wrong method.  But, given my strong reaction in the first place, I think it would be obvious I've come from an environment where it was done differently.