Author Topic: Anomalies and/or Bugs in/between the Rigol MSO/DS1000Z, MSO/DS2000A, MSO/DS4000  (Read 30444 times)

0 Members and 1 Guest are viewing this topic.

Offline i4004

  • Regular Contributor
  • *
  • Posts: 71
  • Country: hr
taken from here
https://www.eevblog.com/forum/testgear/rigol-ds1000z-series-font-size/msg559800/#msg559800


aaa..there's the marmad i know, getting all feisty  over nothing.
again.  :P



Well, no - sorry, I have not learned anything new from your posts - which continue to be filled with loads of incorrect speculation and bad information.  :)

damn, i see simillar posts from teenagers when they first appear on the web!  ;)
ego-trips.
adults just correct the wrong info.
you learned your camera is crap. i bet recent smartphones have better quality.




Quote
There is no "virtual 12bit stuff". It seems you still don't understand how the DSOs are working. I suggest you go back and re-read the posts in the other thread.

nope, i think you didn't understand the terminology i used, in a simillar fashion that you calculate bits of vertical resolution to end up with a decimal point numbers. 9.8bits i think you calculated somewhere.  :palm:
if the adc is 8bit (and it is on both 1k and 2k scopes) than any way to achieve higher res will be virtual because it's not real, it's achieved by tricks.
(as rf loop explains shorter and simpler than your pdf paper, not that pdf is bad)
it's just that 2k is devoting more cpu to doing it, ie z is not really doing it, like we saw from the screenshots.
i understand exactly how they get virtual "12bits" from real 8bits. i mentioned it in the thread, if it's wrong, correct it.




Quote
No, you're wrong again on multiple points. Either your eyes are bad, you don't use a color-corrected monitor, or you don't understand what you're looking at.

my god man, my god! come here and we'll test our eyes, we'll see their color performance, we'll see of color (on monitor) affects this (at all) and we'll also talk about pixelization by lower resoltuion, which seems not really quite your field of expertise.....
if you can't see a drastic resolution decrease in z portion of the screen i posted above, then you'0
re quite unfamiliar with the way images are scaled.....
see this, both images originated from same lores source, one is jaggy, one is less jaggy but more blurred...the only difference is resizer.....





now, it is not really important for us to go into details of how and why (although we already did, and can do some more) but the fact is z has consideratly less vert. res. with same number of bits of adc. and same display resolution!


Quote
There is no lower resolution - they are exactly the same -

meh, make a 320x240 paint image of diagonal thin line(or sine form), and then make the same thing (same line, slanted in a same angle) on 1920x1080, and you'll probably get what i'm talking about.
probably!

Quote
and the only thing that the DS1000Z has going for it is slightly more contrast (due to a brighter backlight). But it's "washed-out" and less "sharp" (this is NOT the same thing as contrast) because of it's worse sub-pixel anti-aliasing. This is apparent - even to those with poor vision :) - when you examine the images close-up (using pa3bca's photos) :

you're going on too far in a wrong direction. you talk about small nuances that less will notice(i would, and better than you, i wear no glasses and can read z display just fine, font is not small for me) and skip the "elephant in the room" ie the jagginess of the z sine forms.

Quote
First, here is the rendering of "black" on both DSO screens - normal and inverted to "white". If it's not obvious which one is closer to true black, then you need a new computer monitor (or you need to calibrate the one you've got):

what a nitpicking parade!
it's probably true, it's probably correct, but i just can't bring myself to the fact that you see the 1000(z?) trees and yet not see the forest!  :-//


Quote
I repeat: there is no higher res - they are both in Normal acquisition mode on displays of equivalent resolution.

this tells me 2k has much better resizer, the thing that converts less adc levels to more pixels the display has.
when both displays are 480 pixels high.

again, could you, please, make a png screenshots from both scopes on that AM signal?
i' m really interested in this, ie why it's happening.
« Last Edit: November 30, 2014, 08:52:43 am by i4004 »
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
damn, i see simillar posts from teenagers when they first appear on the web!  ;)
ego-trips.

No, this is what adults say to children who continually interject nonsense that is incorrect, off-topic, rude, or irrelevant.   :)

Quote
nope, i think you didn't understand the terminology i used, in a simillar fashion that you calculate bits of vertical resolution to end up with a decimal point numbers. 9.8bits i think you calculated somewhere.  :palm:

If you bothered to actually read or comprehend anything that was posted here, you would know I was quoting those numbers from a Tektronix document. This just proves, once again, that you have no interest in understanding anything that doesn't fit within your own narrow concerns or opinions.

Quote
if the adc is 8bit (and it is on both 1k and 2k scopes) than any way to achieve higher res will be virtual because it's not real, it's achieved by tricks.

It's math - plain and simple. Is all math just a trick? Is the fact that we turn real-world analog signals into digital number-streams via electrons and math a trick? I suppose all those things would seem like tricks to, for example, a monkey.

Quote
...and we'll also talk about pixelization by lower resoltuion, which seems not really quite your field of expertise.....

 ;D  Much more than yours, I'm quite sure.
 
Quote
you're going on too far in a wrong direction. you talk about small nuances that less will notice(i would, and better than you, i wear no glasses and can read z display just fine, font is not small for me) and skip the "elephant in the room" ie the jagginess of the z sine forms.

I was responding to and writing about exactly the things that people in the other thread were interested in and asking about: the legibility of the fonts and the overall display quality (the thread is called "Ridiculously small Rigol DS1000z series font size"). You, otoh, were driving off-topic.

Quote
this tells me 2k has much better resizer, the thing that converts less adc levels to more pixels the display has.
when both displays are 480 pixels high.

again, could you, please, make a png screenshots from both scopes on that AM signal?
i' m really interested in this, ie why it's happening.

This has all been discussed in-depth in another thread; I suggest you read it there. Oh, wait - you're the guy that doesn't know how to use Google to do forum seaches, right?  :palm:  Ok, I suggest you FIRST learn how to do forum searches, then find where we were discussing the mapping of the ADC of the DS1000Z, then read it there.  :)
« Last Edit: November 30, 2014, 01:47:48 pm by marmad »
 

Offline i4004

  • Regular Contributor
  • *
  • Posts: 71
  • Country: hr
Quote
If you bothered to actually read or comprehend anything that was posted here, you would know I was quoting those numbers from a Tektronix document.

i know that.

but do tell: how many levels is that bit number, and in what circumstances?
bleh
i like hess' response better:
Quote
I am not sure what you found but I suspect the numbers, 6.8 bits and 9.7 bits, came from the ENOB (effective number of bits) specifications that Tektronix often includes or used to include.  They represent the dynamic range of the digitizer and have nothing to do with mapping ADC levels to the display.

You can find the same type of ENOB specifications in the datasheets for the ADCs Rigol is using but they represent ideal values; the real numbers are lower do to outside factors like clocking and noise.

Quote
It's math - plain and simple. Is all math just a trick?
marmad, i'm pretty sure you would pick 'true' 10bit or 12bit adc (without extra processing) versus 8 bit with virtual 12bits, am i not correct?
so all this writing of yours flops right there.

math can be used in all sorts of ways. i would prefer 1:1 mapping with more bits in adc, and so would you.

Quote
Much more than yours, I'm quite sure.

for you, these 2 segments

are the same?
and you talk about my vision and reasoning?  :clap:

Quote
You, otoh, were driving off-topic.

kinda, sorta, but because i saw great difference in the image (of both scopes) you posted there, and that image shows fonts in some way?  :-DD

but i agree with you, some offtopic is good!
you said it quite well in other thread!
https://www.eevblog.com/forum/blog/eevblog-683-rigol-ds1000z-ds2000-oscilloscope-jitter-problems/msg556083/#msg556083

  ;D

Quote
Oh, wait - you're the guy that doesn't know how to use Google to do forum seaches, right?
that's quite mean, as i already explained
https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg530392/?topicseen#msg530392
Quote
".just like i wrote in added edit "it's over, google found us just now, now everybody knows!" which was written prior to your post.
so yeah, now you can, but then? no."
you also have this
https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg521251/#msg521251


so don't use that garbage anymore, it's doesn't become truth if you repeat it.
it wasn't easy to find then. now?
much easier....there's a yt video on first page of results, just like i wished for!  ;)

Quote
Ok, I suggest you FIRST learn how to do forum searches, then find where we were discussing the mapping of the ADC of the DS1000Z, then read it there.

what, this stuff between me and you
https://www.eevblog.com/forum/blog/eevblog-683-rigol-ds1000z-ds2000-oscilloscope-jitter-problems/msg555796/#msg555796
?

what about if i tell you how would i do it, and then you tell me how would you do it?

ok, presuming we can't have 10 or 12bit adcs (and we obviosuly can't) i would resize those 256 voltage levels to desired display resolution by using something like lanczos resize (in hardware in real time) or alike.
now you. how would you do it?
256 levels to 400 pixels. would you use 'double pixels' and why on earth would you do that?
hehe....
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
Following up on Teneyes' post on the Trigger Output jitter change (normally ~8ns on both DS1000Z and DS2000), I did a little more experimentation.

In the first set of images from the DS2000 running FW 03.01.00.04, switching the trigger coupling from DC -to- AC -to- LFR/HFR has the effect of changing the jitter from 8ns -to- 1.2ns -to- 4 alternating bands of 1.2ns preceded by gaps of 800ps.

In the second set of images from the MSO1000Z using FW BETA 04.01.03.00, switching the trigger coupling from DC -to- AC/LFR has the (non)effect that Teneyes noticed - i.e. the output jitter doesn't get smaller (although the delay out changes - notice the trigger delay settings used to center the waveform). But the interesting thing that I noticed was that using the HFR trigger coupling produced a rock-solid Trigger Output (absolutely no jitter - last image) - although the input frequency of the triggering signal probably can't be too fast since the HFR is a ~75k low-pass filter (I tested it working to  500kHz).
« Last Edit: December 05, 2014, 07:00:05 am by marmad »
 

Offline XFDDesign

  • Frequent Contributor
  • **
  • Posts: 442
  • Country: us

It's math - plain and simple. Is all math just a trick? Is the fact that we turn real-world analog signals into digital number-streams via electrons and math a trick? I suppose all those things would seem like tricks to, for example, a monkey.


If you follow his claim (as it wasn't an argument; it wasn't backed up by anything), then all existence is "virtual" as it's merely "tricks" of physics and chemistry. The transistor's function is "virtual" by nature of it's use of "tricks" in applied physics and the use of math. The resulting voltage tolerance of two capacitors in series is a "virtual" capacitor, by mere function of the use of physics (and the application of math). Etc.

In an attempt to return the subject at hand, does anyone have a list of bugs with the DS4000 series scopes? I'm particularly interested as my TDS5034B died, and as a result am no longer buying Tektronix anything (same series model now costs 2x, and is make just the same in China. If the decision is "a $5k scope that is made in China and lasts 3 years" versus "a $15k scope that is made in China and lasts 3 years" well the Tek amount of money will last me 9 years, which is almost as long as the TDS5034B lasted me. But, the concern is the bug list. People on the forums have said that the 4k series has a lot of bugs, yet all my googling has only unearthed a SPI Decode bug and that is it.  :-\
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
In an attempt to return the subject at hand, does anyone have a list of bugs with the DS4000 series scopes? I'm particularly interested as my TDS5034B died, and as a result am no longer buying Tektronix anything (same series model now costs 2x, and is make just the same in China. If the decision is "a $5k scope that is made in China and lasts 3 years" versus "a $15k scope that is made in China and lasts 3 years" well the Tek amount of money will last me 9 years, which is almost as long as the TDS5034B lasted me. But, the concern is the bug list. People on the forums have said that the 4k series has a lot of bugs, yet all my googling has only unearthed a SPI Decode bug and that is it.  :-\
Have you seen the one created by Altemir? It's fairly comprehensive, but I notice that he doesn't have the latest FW listed (02.02.01.01) and that he hasn't been on the forum for 2 months - so it hasn't been updated (e.g. I know the latest version allows decoding of recorded frames). But you can probably ask over in that thread for more details about other bugs that might have been fixed in the latest FW.
« Last Edit: December 02, 2014, 03:55:44 pm by marmad »
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
But the interesting thing that I noticed was that using the HFR trigger coupling produced a rock-solid Trigger Output (absolutely no jitter - last image) - although the input frequency of the triggering signal probably can't be too fast since the HFR is a ~75k low-pass filter (I tested it working to  500kHz).
Yes, Very interesting.
I have a suspicion that the trigger out in the HFR mode uses a fix hardware timer ,
possible made for Mixed mode.
I will ask Motocoder to test hs DS2000 with his leak beta version on the HFR mode.to see how that looks.

Screenshot attached.
 

Offline XFDDesign

  • Frequent Contributor
  • **
  • Posts: 442
  • Country: us
Have you seen the one created by Altemir? It's fairly comprehensive, but I notice that he doesn't have the latest FW listed (02.02.01.01) and that he hasn't been on the forum for 2 months - so it hasn't been updated (e.g. I know the latest version allows decoding of recorded frames). But you can probably ask over in that thread for more details about other bugs that might have been fixed in the latest FW.

I hadn't, and my SearchFu is apparently weak. Thank you very much for the pointer!
 

Offline Mark_O

  • Frequent Contributor
  • **
  • Posts: 939
  • Country: us
Following up on Teneyes' post on the Trigger Output jitter change (normally ~8ns on both DS1000Z and DS2000), I did a little more experimentation.

In the first set of images from the DS2000 running FW 03.01.00.04, switching the trigger coupling from DC -to- AC -to- LFR/HFR has the effect of changing the jitter from 8ns -to- 600ps -to- 4 alternating bands of 600ps preceded by gaps of 400ps.

I'll buy the 8ns, but your '600ps' looks more like 1.2ns to me.  I think you may have forgotten one factor of 2x when you multiplied?  (0.6 div, yes, but 2ns per.)

Same for the 600ps+400ps.  That winds up 1.2 + 0.8ns, and the 4 bands combined span the same 8ns as the original, just segmented.
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
I'll buy the 8ns, but your '600ps' looks more like 1.2ns to me.  I think you may have forgotten one factor of 2x when you multiplied?  (0.6 div, yes, but 2ns per.)

Same for the 600ps+400ps.  That winds up 1.2 + 0.8ns, and the 4 bands combined span the same 8ns as the original, just segmented.

Yep - good catch - somehow I used 1ns/div when counting tick marks. Fixed.
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
Another strange anomaly in the DS1000Z - first spotted by rf-loop in this thread.

As pointed out by rf-loop (and confirmed by me) the Sin(x)/x ON/OFF switch on the DS1000Z does absolutely nothing to change the interpolation - if the DSO is using Sin(x)/x, it stays Sin(x)/x - and if it's using Linear, it stays Linear - regardless of the switch setting.

But I think Rigol has completely screwed-up the labeling of the function of this switch. I think, instead of interpolation, it's doing some kind of BW-limiting (maybe 50MHz?) on all of the input channels simultaneously - perhaps to try to prevent aliasing when 3/4 channels are active with the reduced 250MSa/s rate. But it only seems to be active between 100 - 5ns/div.

Image #1 shows 100MHz sine to CH 1 & 4: "Sin(x)/x ON"
Image #2 shows 100MHz sine to CH 1 & 4: 20MHz BW limit on CH1 - "Sin(x)/x ON"
Image #3 shows 100MHz sine to CH 1 & 4: 20MHz BW limit on CH1 - "Sin(x)/x OFF"
« Last Edit: December 08, 2014, 02:34:29 am by marmad »
 

Offline leppie

  • Frequent Contributor
  • **
  • Posts: 269
  • Country: za
perhaps to try to prevent aliasing
So many useless buttons/features on this scope :(

Is there a summary post for all known DS1000Z issues yet?
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
The Rigols were the first oscilloscopes I considered when looking for a new DSO and I expected better.  Making measurements (and protocol decoding) directly from the display is something I would expect in a toy no matter who makes it.

It's SOP below a certain price point, so I guess it's just a market full of toys. Too bad for you; you'll likely never be able to buy a low-cost DSO.

I bought several low cost DSOs at that time including a Tektronix 2440 but they are all old enough to vote and drink which is not an avenue I would suggest to a neophyte.  None of them were newer than 21 years old which is a measure of how little I thought of Rigol's instruments at that time.

Quote
Personally, I wouldn't count on an oscilloscope for making precise measurements (a habit I learned from trying to count graticule tick marks on analog scopes) - just as I wouldn't count on an MSO for complex logic analyzing - feels rather like using a screwdriver to hammer a nail.

Using the display record to make automatic measurement is a hack; its like making measurements from a screenshot instead of the actual data and just adds another layer of processing and aliasing.  I would like to know why they did it this way.  I would suspect it was for performance reasons except that their slow DSOs work this way also.

One of the annoying results is that measurements made of a saved waveform change depending on how it is displayed.  As discussed in another EEVblog thread, there is more than a loss of precision if the automatic measurements are jumping by 50% on a saved waveform.  That is just broken.
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
Using the display record to make automatic measurement is a hack; its like making measurements from a screenshot instead of the actual data and just adds another layer of processing and aliasing.  I would like to know why they did it this way.

As mentioned multiple times already, Agilent does it - and likely did it before any of the Chinese manufacturers did it - so I would guess that's where the idea comes from. So you continually bringing it up solely in terms of Rigol is both silly and a moot point.
« Last Edit: December 13, 2014, 08:36:24 pm by marmad »
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Using the display record to make automatic measurement is a hack; its like making measurements from a screenshot instead of the actual data and just adds another layer of processing and aliasing.  I would like to know why they did it this way.

As mentioned multiple times already, Agilent does it - and likely did it before any of the Chinese manufacturers did it - so I would guess that's where the idea comes from. So you continually bringing it up solely in terms of Rigol is both silly and a moot point.

Agilent doing it does not make Rigol's implementation any better.

Which models of Agilent DSO operate this way?  Didn't they outsource design and production of their low end DSOs to China?
 

Offline Teneyes

  • Frequent Contributor
  • **
  • Posts: 498
  • Country: ca
As mentioned multiple times already, Agilent does it - and likely did it before any of the Chinese manufacturers did it -
Old connection of Aligent to Rigol is noted in this article back in 2009
And I am sure there was some exchange of technology (spying,agreements)
« Last Edit: December 13, 2014, 09:07:03 pm by Teneyes »
IiIiIiIiIi  --  curiosity killed the cat but, satisfaction brought it back
 

Offline marmadTopic starter

  • Super Contributor
  • ***
  • Posts: 2979
  • Country: aq
    • DaysAlive
Agilent doing it does not make Rigol's implementation any better.

I wrote nothing about "better". What I did write, repeatedly already, was that you bringing it up in multiple threads as if it's some strange anomaly of Rigol scopes is just nonsense. Agilent, Siglent, Hantek, and who knows how many others, do it. We've already been discussing this issue on this forum for years before you joined.
« Last Edit: December 13, 2014, 09:53:34 pm by marmad »
 

Offline i4004

  • Regular Contributor
  • *
  • Posts: 71
  • Country: hr
any of the listed bugs/anomalies fixed in recent firmware(s)?

is rigol reading this, or is it just fixing the "peaks", ie bugs that are most obvious?

as someone else mentioned in another thread, such threads would be more usefull if the findings and fixes were updated in the first post.

 

Offline rowifi

  • Contributor
  • Posts: 37
Not been following this thread, so this may already be known.
Our DS2202A shows a strange behavior when triggering.
It does not trigger with the vertical gain set low, even though the signal is visible on the screen

Using edge trigger and using Normal mode, use the built in scope probe test pin. The square wave is 3V peak to peak. Set Vertical gain to 1V/div and set trigger to say midway ( 1.5v). The Trigger works and the waveform is shown.
.
Reducing the Vertical gain in steps, the waveform will reduce and the trigger level indicator also reduces accordingly, but when the vertical gain is 10V/div  or lower  the trigger fails and a WAIT message is shown.

The scope does record the waveform for one sweep when the vertical gain is changed, and it therefore 'appears' to be triggering, but the WAIT message is displayed and the scope is not updating the waveform.
 
Forcing a trigger or using auto mode shows that the signal is still present and measurable by eye.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf