Author Topic: New Rigol DS1054Z oscilloscope  (Read 1482981 times)

0 Members and 3 Guests are viewing this topic.

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 2113
  • Country: hr
Re: New Rigol DS1054Z oscilloscope
« Reply #4375 on: November 11, 2018, 02:16:01 pm »
This is the moment where it would be useful to have jitter histogram. It is possible that jitter you're measuring spends very little time at the extremum so it gets into blind time for a Pulse width+ measurement routine.
It is obvious that measurement are not done at same dataset that is used to establish persistence on the screen...

I cannot reproduce it here with siggen. I don't have the module, sorry.

Regards,
Sinisa
 

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 1726
  • Country: ro
Re: New Rigol DS1054Z oscilloscope
« Reply #4376 on: November 11, 2018, 02:28:39 pm »
The modulation is actual jitter in a clock, ESP8266 module.

There is something very fishy there.  That is a huge jitter for a clock driving a radio at 2.4GHz.

What pin did you probe in the initial picture? (https://www.dropbox.com/s/jzfwh67ivyejl6t/DSC_0050.JPG)

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
Re: New Rigol DS1054Z oscilloscope
« Reply #4377 on: November 11, 2018, 02:36:23 pm »
RoGerorge, having both edges always present might be a  clue. I will
start looking for bloated pulse using trigger system to see if that may
be cause.

Thanks, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 7906
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: New Rigol DS1054Z oscilloscope
« Reply #4378 on: November 12, 2018, 02:30:48 am »
bitseeker -

I do not see any setting, other than choice to apply measurement to "screen"
or "cursors", to control sample set. I have it set to screen, and cursor measurement
looks correct. Stats does not.

Go to Measure -> Statistic -> Stat.Sel. -> Difference to display Cnt (instead of Min/Max), which is the number of samples collected for the currently displayed statistics. Then, you'll be able to confirm that it's sampling during the entire display persistence period and that it hasn't reset for some reason.

Quote
If the statistics were not examining all sweeps, even if this were true, over time it would,
all things being equal, get a sample for all possible values. Eg. the jitter would produce
multiple samples for same point in time, over long acquisition periods. I have done hours
to see see if the measurement would converge, does not.

Agreed. With a sufficient number of samples, the stats should be closer to the cursor measurements. Turning on the sample counts display will help confirm that you're getting lots of samples.

Also, you can go to Measure -> History -> DispHistory to turn on the measurement history followed by Measure -> History -> Disp Type -> Table to see the values that have been collected. That will enable you to confirm, for example, that the scope got at least one of the Min and Max values that you measure with cursors.

These are not a solution, but they may be helpful in determining if the data is being captured as expected.
I TEA.
 

Offline JHuston

  • Newbie
  • Posts: 1
Re: New Rigol DS1054Z oscilloscope
« Reply #4379 on: February 01, 2019, 12:13:09 am »
ANYONE GETTING AN ERROR ABOUT PRIVATE KEY:

YOU ARE ENTERING YOUR SERIAL NUMBER WITH A 8 INSTEAD OF AN 'S' AND THE SCRIPT IS NOT RECOGNIZING IT.  SERIAL NUMBERS ARE DS NOT D8
 

Online ebastler

  • Super Contributor
  • ***
  • Posts: 3025
  • Country: de
Re: New Rigol DS1054Z oscilloscope
« Reply #4380 on: February 01, 2019, 04:41:59 am »
ANYONE GETTING AN ERROR ABOUT PRIVATE KEY:

YOU ARE ENTERING YOUR SERIAL NUMBER WITH A 8 INSTEAD OF AN 'S' AND THE SCRIPT IS NOT RECOGNIZING IT.  SERIAL NUMBERS ARE DS NOT D8

And there are other ways to potentially mis-read the serial number, due to the small font on the screen. The serial number is also printed on a sticker on the scope's back -- easier to read than the small on-screen letters.
 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
pass/fail
« Reply #4381 on: April 01, 2019, 10:27:45 pm »
I was able to get pass/fail working to stop when a trace falls outside the mask, but can anyone tell me if it is possible to record only failed traces with DS1054Z?

It seems that if I have pass/fail active, then the record menu is dimmed out, and vice versa.  |O
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4382 on: April 02, 2019, 12:00:07 am »
I was able to get pass/fail working to stop when a trace falls outside the mask, but can anyone tell me if it is possible to record only failed traces with DS1054Z?

It seems that if I have pass/fail active, then the record menu is dimmed out, and vice versa.  |O

I don't think it's possible.

You can enable infinite persistence so all the failed waves will show overlaid on the screen.

 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
Re: New Rigol DS1054Z oscilloscope
« Reply #4383 on: April 02, 2019, 12:42:43 am »
 :wtf:

Wow, that's a really, really stupid limitation.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4384 on: April 02, 2019, 01:56:16 am »
:wtf:

Wow, that's a really, really stupid limitation.

Not at all. It needs the sample memory to continuously acquire signal and look for the next bad pulse.

It's not the same as filling the memory with pulses, stopping the data acquisition, figuring out the start point of each one for playback.

It's a hardware limitation, sure, but I wouldn't call it "stupid".

With a bit of scripting you could put it in single shot mode, wait for it to find a bad pulse, grab the wave, screenshot it, start it running again, repeat.

 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
Re: New Rigol DS1054Z oscilloscope
« Reply #4385 on: April 02, 2019, 05:42:22 am »
But the memory's fast enough to take the stats on failed sweeps in realtime - I would have thought that would mean it could just continuously write over the same sweep until there's a fail, then move to the next segment of memory - thus only storing failed sweeps.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4386 on: April 02, 2019, 07:09:58 am »
...then move to the next segment of memory.

The logic is obvious, yes, but things that seem simple/obvious aren't always easy/possible to implement when you get into the details.

eg. It still has to trigger on every waveform (it doesn't know if it's going to be bad or not until it has time to run the test it), maybe it has no ability to "rewind" when a wave passes. Maybe it doesn't know if a wave has passed/failed until several more waves have arrived and are in memory so a "rewind" would overwrite those newer waves and there might be bad ones in there.

« Last Edit: April 02, 2019, 07:13:52 am by Fungus »
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 2113
  • Country: hr
Re: New Rigol DS1054Z oscilloscope
« Reply #4387 on: April 02, 2019, 09:47:12 am »
I know R&S rtb2000 /rtm3000 can save failed captures to segements.
Very expensive Keysight dsox3000T can't, it will only save screenshot to USb disc, as will Rigol MSO5000 /ds7000,
So not very common thing to do although does seem logical..
 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
Re: New Rigol DS1054Z oscilloscope
« Reply #4388 on: April 03, 2019, 02:27:05 am »
...then move to the next segment of memory.

The logic is obvious, yes, but things that seem simple/obvious aren't always easy/possible to implement when you get into the details.

Agreed.

Quote
maybe it has no ability to "rewind" when a wave passes.


I don't see why RAM wouldn't be random access.

Quote
Maybe it doesn't know if a wave has passed/failed until several more waves have arrived and are in memory so a "rewind" would overwrite those newer waves and there might be bad ones in there.

I get it, there can definitely be complications, but it's still very surprising to me that it can't do it at all. Even if it could only do a limited number of traces, or only sample every now and then because of the extra processing time, it would still IMO be a very useful thing.

Failed traces are the meat and potatoes when you're trying to troubleshoot why you're getting failed traces!!
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16206
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: New Rigol DS1054Z oscilloscope
« Reply #4389 on: April 03, 2019, 02:35:35 am »
So a waveform Search feature that you can change the settings for and also transfer to the trigger might suit your needs better ?
Have a squiz at this:
https://www.eevblog.com/forum/testgear/siglent-sds1204x-e-released-for-domestic-markets-in-china/msg1370717/#msg1370717
Avid Rabid Hobbyist
 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
Re: New Rigol DS1054Z oscilloscope
« Reply #4390 on: April 03, 2019, 04:03:04 am »
So a waveform Search feature that you can change the settings for and also transfer to the trigger might suit your needs better ?
Have a squiz at this:
https://www.eevblog.com/forum/testgear/siglent-sds1204x-e-released-for-domestic-markets-in-china/msg1370717/#msg1370717

I suppose it would, but it would better if the mask thingy worked in record mode.

I think I could have captured the problem I was looking for yesterday with creative trigger settings, but once again the fail mask working with record mode would be better.

Oh well, can't have everything I suppose.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4391 on: April 03, 2019, 06:06:46 am »
Quote
maybe it has no ability to "rewind" when a wave passes.


I don't see why RAM wouldn't be random access.

It's not a PC. The sample memory in a DSO is just a big circular buffer being fed by the ADC. The main CPU has very limited access, most access is done through an FPGA whose main job is downsampling the data into a format suitable for display on screen.

What you're proposing would mean that pieces of sample memory need to copied around (you have to save only the interesting stuff, accumulating it in the memory). It might be impossible to do that in a DS1054Z (at least; impossible without stopping the ADC, in which case you'd be moaning that it misses waves when they all come along at once).

 :popcorn:
 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
Re: New Rigol DS1054Z oscilloscope
« Reply #4392 on: April 03, 2019, 08:17:28 am »
It's not a PC. The sample memory in a DSO is just a big circular buffer being fed by the ADC. The main CPU has very limited access, most access is done through an FPGA whose main job is downsampling the data into a format suitable for display on screen.

What you're proposing would mean that pieces of sample memory need to copied around (you have to save only the interesting stuff, accumulating it in the memory). It might be impossible to do that in a DS1054Z (at least; impossible without stopping the ADC, in which case you'd be moaning that it misses waves when they all come along at once).

 :popcorn:

It must be able to have some level of "stopping" the recording to be able to do segmented memory, which only records when a trigger occurs.

I agree it would be one extra layer of processing to then have to do the masking, but clearly it can already manage to do that in realtime in order to keep the stats on failures.

Surely it isn't that hard to move the pointer in a circular buffer back to the start of the last sweep, even for an FPGA. There's no copying around, just moving the pointer back if it's a "pass" sweep that you don't want to keep. The sweeps will all be the same size too which makes things even simpler.

It's all moot points because it's one of the few features that this oscilloscope is missing. I'm still very happy with it overall, but I did get a big surprise that this one wasn't there.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4393 on: April 03, 2019, 08:39:34 am »
I'm still very happy with it overall, but I did get a big surprise that this one wasn't there.

If they disabled the "record" menu then they're aware there's a conflict between the two things. I'm guessing there was no solution.

 

Online ebastler

  • Super Contributor
  • ***
  • Posts: 3025
  • Country: de
Re: New Rigol DS1054Z oscilloscope
« Reply #4394 on: April 03, 2019, 08:40:06 am »
Maybe the classification of "good" vs. "bad" traces lags behind the recording? That would make it somewhat messy to go back and remove the good traces from the buffer, to retain the outliers only, since recording has already moved on and stored the next few sweeps.

Obviously feasible "in principle" -- you essentially want two memory buffers: a short-term one for interim storage until a trace has been classified, and a long-term memory to retain outlier traces. But that may not be supported by the Rigol hardware.

Anyway, this is just speculation. I'm afraid we simply have to live with the fact that the DS1000Z series does not offer this feature; and I wouldn't expect it in an upcoming firmware update (if we get another one...)
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4395 on: April 03, 2019, 08:53:19 am »
Obviously feasible "in principle" -- you essentially want two memory buffers: a short-term one for interim storage until a trace has been classified, and a long-term memory to retain outlier traces. But that may not be supported by the Rigol hardware.

There might be no way to copy sample data around in the sample memory, eg. Only the ADC has write access.

We know the sample data has to be downsampled to "screen" resolution before it can be compared with the pass/fail mask. This introduces a time delay between sampling/classification.

If there's a time delay between sampling/classification and no way to move sample data around then you can't do it. Period.

Anyway, this is just speculation.

Yep.
 

Offline magicsmokelost

  • Contributor
  • Posts: 7
  • Country: au
Re: New Rigol DS1054Z oscilloscope
« Reply #4396 on: April 04, 2019, 10:47:11 pm »
There might be no way to copy sample data around in the sample memory, eg. Only the ADC has write access.

We know the sample data has to be downsampled to "screen" resolution before it can be compared with the pass/fail mask. This introduces a time delay between sampling/classification.

If there's a time delay between sampling/classification and no way to move sample data around then you can't do it. Period.

Valid point, but you could eliminate the need for moving sampling data around by just jumping back to right after the failed sample you just found. You'd lose the interim sweeps but at least you'd get the first fail and be able to continue sampling.

This would mean we could just leave the thing going and come back later and see all (most of?) the failed sweeps that were recorded. That would be much better than the current solution which is to disable the recording feature entirely.
« Last Edit: April 04, 2019, 10:48:44 pm by magicsmokelost »
 

Offline metrologist

  • Super Contributor
  • ***
  • Posts: 1799
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4397 on: April 04, 2019, 11:35:06 pm »
I am adding it to my want list. Yes I do.

Is this an application for a SCPI automated test solution?
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10120
  • Country: 00
Re: New Rigol DS1054Z oscilloscope
« Reply #4398 on: April 05, 2019, 07:31:55 am »
Is this an application for a SCPI automated test solution?

I reckon so.
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 2113
  • Country: hr
Re: New Rigol DS1054Z oscilloscope
« Reply #4399 on: April 05, 2019, 09:26:15 am »
Mask testing in ds100'Z is done in software.What you want cannot be done. Period.
If 15000USD Keasight 3000T doesn't have that option why you would think it was trivial or necessary on cheapest one.
 Only R&S new 2000/3000&4000 series have it.
It could be added to new Siglents and new 5000&7000 series Rigols because they have architecture that is powerful enough  but it would need a change in a way acquisition engine works. Nothing trivial or easy about that

On 1000Z  it would be more useful to make decode to work from recorded data (and possible too). On 1000Z, they could add option to save a screen capture on fail, like Keysight and MSO5000/7000 do.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf