Author Topic: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems  (Read 505444 times)

0 Members and 1 Guest are viewing this topic.

Offline jnissen

  • Regular Contributor
  • *
  • Posts: 63
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #75 on: November 14, 2014, 07:36:32 pm »
The 100 kHz period of the jitter showing up at a delay of 5us and not at a delay of 10us could be produced by frequency modulation of the sample clock like they do with spread spectrum modulation to meet emissions requirements.  Such modulation should not be used on a sample clock unless the effect is compensated for and some frequency counters and oscilloscopes do this deliberately to avoid synchronous sampling.  Maybe Rigol used a clock module which includes that function or the PLL which generates the sampling clock is being modulated somehow.


I've got a suspicion that is exactly what were dealing with here.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #76 on: November 14, 2014, 07:43:27 pm »
The triggering is digitally done after digitization so there is nothing for an external signal to affect.

If that's the case it could explain the big difference in behavior between AC & DC coupling.

I think they are unrelated.  The jitter at 5us and every 10us after that looks like a problem with the sampling clock being frequency modulated at 100 kHz producing a peak to peak jitter of about 1ns on Dave's DS1054Z and about 5ns on EcProject's MSO1104Z.  The AC trigger coupling problem results in 8ns of peak to peak jitter on Dave's DS1054Z at the trigger point.

Quote
As regards differences from unit to unit, that could be small differences in frequency between scope and source, i.e. it's not a unit to unit difference, it's a source to source difference.

The jitter caused by the AC trigger coupling selection is at the trigger point so the jitter (and frequency) from the signal source is irrelevant as long as the average DC level does not change significantly and even if that was causing the problem, then the trigger point would not be offset as shown at 9:21 in Dave's video.  What is shown at that point in his video resembles what happens when a high frequency reject filter is used with analog triggering but I am sure there could be a more esoteric cause with the digital trigger.

The jitter is also not what I would expect from aliasing although if you look closely at the Tektronix example at 12:21 in the video, you can see the results of what might be aliasing immediately around the trigger point.  The Tektronix sample rate is high though so more likely it is just an artifact caused by noise interacting with the trigger.
 

Offline Nonorthogonal

  • Contributor
  • Posts: 11
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #77 on: November 14, 2014, 07:49:15 pm »
 

Offline alexanderhiam

  • Contributor
  • Posts: 25
  • Country: us
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #78 on: November 14, 2014, 07:52:38 pm »
I got my DS1104Z as soon is they were selling in the US, and it's SW version 00.02.00.SP1. The system info doesn't show a hardware version. It has the AC coupling jitter, but doesn't have the horizontal offset jitter at all. In place of that though it has a wonderful feature where it likes to reset itself at random intervals, like while I'm in the middle of looking through a large capture... it's a fun one.
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #79 on: November 14, 2014, 07:57:06 pm »
What's the liklihood of there being a software fix/cludge for this in the future (which might rescue earlier machines), or should we wait for Rigol to fix the hardware and tehn buy?
If you don't measure, you don't get.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #80 on: November 14, 2014, 07:57:59 pm »
Would be interesting to compare results with internal vs. external function gen - something like leakage of an internal clock to the analogue trigger circuitry would probably have a different effect depending on whether or not it was synced to the scope's clock.

The triggering is digitally done after digitization so there is nothing for an external signal to affect.

David, what makes you conclude these scopes have digital triggering?  Do you mean like the R&S RTO scopes?
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #81 on: November 14, 2014, 07:59:41 pm »
Major problem discovered 2 years after one of the most popular scope (DS2XXX) was released.
Apparently, nobody uses AC coupling trigger mode these days  ;)

AC coupled triggering is useful when DC offset is used or if the vertical positioning is done before the trigger pickoff so it acts as a DC offset control; vertical positioning and offset are often combined but they do not have to be.  If DC triggering is used in these cases, then altering the vertical position or offset also alters the trigger level which is usually not the desired behavior.

For this reason oscilloscopes which have a combined position and offset control often always use AC coupling when internal triggering is used.

Oscilloscopes which support peak to peak triggering can avoid this issue even when using DC coupled triggering.
« Last Edit: November 14, 2014, 08:07:47 pm by David Hess »
 

Offline mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13693
  • Country: gb
    • Mike's Electric Stuff
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #82 on: November 14, 2014, 08:02:06 pm »
Could the dual-slope trigger thing be just a synchronisation/number of cycles thing - does adjusting holdoff make a difference?
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline EE-digger

  • Frequent Contributor
  • **
  • Posts: 333
  • Country: us
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #83 on: November 14, 2014, 08:11:05 pm »
As far as the jitter at delay goes, it would seem like the the architecture of the acquisition and memory systems might come into question.

On my TDS784D, I can plug in any delay value and see no jitter whatsoever.  It became scary with really long delays so I turned on FM to be sure, so I could see no "jitter" at zero delay and increasing "jitter" as the delay was increased.

On the other hand, the problem could simply be incorrect indexing when mapping acquisition memory to display memory, and adding the intensity grading process on top of that.

If it is simple, how fast is Rigol at turning around solutions?  If they are as fast as they are with revising security measures, we're all set :) :)
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #84 on: November 14, 2014, 08:15:36 pm »
Would be interesting to compare results with internal vs. external function gen - something like leakage of an internal clock to the analogue trigger circuitry would probably have a different effect depending on whether or not it was synced to the scope's clock.

The triggering is digitally done after digitization so there is nothing for an external signal to affect.

David, what makes you conclude these scopes have digital triggering?  Do you mean like the R&S RTO scopes?

There are a couple of reasons:

  • The teardowns and reverse engineered schematics show no trigger pickoff from the vertical signal path and no auxiliary trigger circuits.
  • The multitude of triggers supported are easily implemented digitally after digitization in an FPGA or ASIC.
  • These oscilloscopes do not support traditional equivalent time sampling.
  • Increasing digital integration makes digital triggering cheaper to implement.  The reason to use digital triggering has changed from being able to support additional trigger types and high acquisition rates to just being cheaper to implement.

Update: I forgot to include the trigger output jitter and the excessive delay in the trigger output.  Both indicate that a digital trigger is used.
« Last Edit: November 14, 2014, 11:13:11 pm by David Hess »
 

Offline Larry Benko

  • Newbie
  • Posts: 1
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #85 on: November 14, 2014, 08:45:05 pm »
2 month old Rigol MSO2202A scope with Firmware 00.03.00.SP1 and Hardware 2.2.  I can NOT rerpeat either the 5 us. or AC jitter Dave saw.  Waveforms are all stable.  Current firmware is 00.03.01 which I have not loaded.
 

Offline xs4hkr

  • Newbie
  • Posts: 5
  • Country: nl
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #86 on: November 14, 2014, 09:05:22 pm »
The 5 us delay problem doesn't occur on a 1 week old MSO2072A (factory calibration date: 29-07-2014, firmware version: 00.03.00.SP1, hw version: 2.2). The AC coupled trigger jitter problem doesn't occur when using probe compensation signal output but does occur when using an equal signal (square, 1 kHz, 3 Vpp) generated  with a function generator.
« Last Edit: November 15, 2014, 12:01:41 am by xs4hkr »
 

Offline Chipguy

  • Supporter
  • ****
  • Posts: 320
  • Country: de
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #87 on: November 14, 2014, 09:10:48 pm »
I can NOT confirm any of that on my MSO1104Z-S. Maybe I did it wrong?
My short video is here: http://youtu.be/wYKInWNtzeY
Where is that smoke coming from?
 

Offline xs4hkr

  • Newbie
  • Posts: 5
  • Country: nl
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #88 on: November 14, 2014, 09:15:37 pm »
Chipguy: you should change the coupling (to ac) in the trigger settings.
 

Offline Kevman

  • Regular Contributor
  • *
  • Posts: 146
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #89 on: November 14, 2014, 09:18:41 pm »
I've been thinking about getting one of these to complement my CRO, if only "because..."

I'm having trouble grasping how big of a deal this is, though. Can someone give a use case for AC coupled trigger?

I didn't think my 2245A had AC trigger coupling till I just looked at it (it has AC, DC and 3 different 'noise rejects') I always just figured it followed the coupling setting for the channel...
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #90 on: November 14, 2014, 09:19:56 pm »
Did you look at the trigger-out?

No, barely had time to shoot and upload this video last night.
Going walkabout  today.
 

Offline Chipguy

  • Supporter
  • ****
  • Posts: 320
  • Country: de
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #91 on: November 14, 2014, 09:29:08 pm »
Chipguy: you should change the coupling (to ac) in the trigger settings.

Ah yeah right.
I just did that, same result. No jitter at all.
Not even worth to shoot another video.
Where is that smoke coming from?
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37661
  • Country: au
    • EEVblog
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #92 on: November 14, 2014, 09:35:21 pm »
The AC coupled trigger jitter problem doesn't occur when using probe compensation signal output but does occur when using an equal signal (square, 1 kHz, 3 Vpp) generated  with a function generator.

That is interesting!
 

Offline Chipguy

  • Supporter
  • ****
  • Posts: 320
  • Country: de
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #93 on: November 14, 2014, 09:47:34 pm »
It was a bit of work to find the right setting, but now at 20 MHz I can confirm the AC trigger jitter.
It is pretty big, awful.
Sinewaves at 1 MHz do not show any jitter, but square waves do.
It's not depening much on the amplitude. Almost the same at 100mVpp and 1Vpp.
Where is that smoke coming from?
 

Offline egonotto

  • Frequent Contributor
  • **
  • Posts: 648
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #94 on: November 14, 2014, 10:01:51 pm »
Hi,


the AC-trigger problem is a delay between 45ns to 55ns. You can say a delay of 50ns with jitter 10ns p-p.

Therefore you need the right input to see this.

The builtin 1KHz has to large risetime.

Best Regards
egonotto
 
 

Offline pciebiera

  • Newbie
  • Posts: 4
  • Country: us
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #95 on: November 14, 2014, 10:03:37 pm »
Another odd thing I just noticed.
When you hit run/stop it just freezes the frame buffer, but then then if you move around the waveform it updates.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #96 on: November 14, 2014, 10:05:52 pm »
The AC coupled trigger jitter problem doesn't occur when using probe compensation signal output but does occur when using an equal signal (square, 1 kHz, 3 Vpp) generated  with a function generator.

Oddly enough the old Tektronix 2465 series analog oscilloscopes do something like this in reverse with their calibration output signal.  When viewed on the oscilloscope, it will be jitter free but it is synchronized with the sweep so when viewed on a different oscilloscope or instrument, it will show significant jitter.  From their manual:

Due to internal circuitry constraints, the calibrator signal is not synchronized during trace holdoff.  This does not affect the accuracy of the calibrator signal that is present during a trace display.  However, if the calibrator signal is used to calibrate other instruments, the sweep must be shut off.  If it is not, the signal will appear to jitter and will give false (low) frequency counts.  The sweep is easily shut off by setting the trigger mode switch to single sequence.
 

Offline mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13693
  • Country: gb
    • Mike's Electric Stuff
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #97 on: November 14, 2014, 10:13:29 pm »
The AC coupled trigger jitter problem doesn't occur when using probe compensation signal output but does occur when using an equal signal (square, 1 kHz, 3 Vpp) generated  with a function generator.

That is interesting!
again, suggestive of the issue relating to slight differences in frequency when using an external source that's independent of the scope's own clock
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline (In)Sanity

  • Regular Contributor
  • *
  • Posts: 229
  • Country: us
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #98 on: November 14, 2014, 10:23:35 pm »
Has anyone tested this with a good old fashion analog  ~20 Mhz sig gen ? 

Jeff
 

Offline rs20

  • Super Contributor
  • ***
  • Posts: 2317
  • Country: au
Re: EEVblog #683 - Rigol DS1000Z & DS2000 Oscilloscope Jitter Problems
« Reply #99 on: November 14, 2014, 10:27:12 pm »
While I was trying the trigger issue on my DS2072, i noticed another bug.

With DC-Coupling there is a trigger-level marker, with AC-Coupling the trigge-level marker dissapears.
So it's more difficult to set the trigger correct.

This is correct behaviour, it makes absolutely no sense to draw a line on the screen when the trigger coupling is set to AC; the filtered waveform used for triggering isn't shown on the screen, so what would you be cross-referencing the line to? It's actually very good that they hide the line, shows that they've thought about it. (If only they tested it :) )
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf