Author Topic: rigol DS1054Z wrong rms measurement on long waveform  (Read 4227 times)

0 Members and 1 Guest are viewing this topic.

Offline ziopaper

  • Contributor
  • Posts: 8
rigol DS1054Z wrong rms measurement on long waveform
« on: May 08, 2015, 02:57:45 AM »
Hello,

I have a brand new beautiful DS1054Z.

If I measure a sinusoid from a funct generator (500mVrms, 60Hz) the measurement looks correct on the scope if I have a time scale of 10ms/div.

expanding the time scale to 500ms/div, the rms measured changes to 725mV (frequency measured also changes to 100Hz)

I tried at 12M points, 6M and 60k. Same result.  On the expanded time scale, if I stop the scope and enable the delayed sweep time base (pressing on the horizontal scale knob) and zoom in enough, the rms measurement will be correct again. This means that the scope has in memory the correct waveform, with enough details (no aliasing or other problems).

So how is it calculating the wrong rms on the long time scale? or is it downsampling to do the calculations? (also the period is wrong, so it is missing the zero crossings)

Also I noticed that with so many points, the rms is still calculated very quickly (too quickly?) while a tektronics DPO3000 would take several seconds to display rms of a long waveform at 5M points.

Is there anything I am missing?

Thanks a lot

Marco
« Last Edit: May 08, 2015, 08:48:42 AM by ziopaper »
 

Offline alsetalokin4017

  • Super Contributor
  • ***
  • Posts: 1563
  • Country: us
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #1 on: May 08, 2015, 02:14:09 PM »
For the measurements selected from the left menu items, I think the scope only uses what is displayed on the screen, or in the zoom window when you are zooming the horizontal timebase, not the whole sample set in memory. So yes, it's probably a sampling error/aliasing problem when you are showing a great many cycles in the screen or zoom window. I think... maybe.
The easiest person to fool is yourself. -- Richard Feynman
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 9203
  • Country: nz
    • Taupaki Technologies Ltd.
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #2 on: May 08, 2015, 02:36:53 PM »
Trigger level setting?
Avid Rabid Hobbyist & NZ Siglent Distributor
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #3 on: May 09, 2015, 05:29:41 AM »
trigger is set to edge, auto sweep, dc coupling.

I guess is doing the rms on the samples that are shown on the screen...I am not sure why. This is a rather simple function since all the samples are stored in the memory.

Any other comments are appreciated

One of the reasons I have this scope is looking at a switching waveform of a PFC, so I do need to read the rms of a switching waveform over a 60Hz period.

this is a bummer. Such a nice scope
 

Online wraper

  • Supporter
  • ****
  • Posts: 5694
  • Country: lv
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #4 on: May 09, 2015, 05:35:23 AM »
I guess is doing the rms on the samples that are shown on the screen...I am not sure why. This is a rather simple function since all the samples are stored in the memory.
Trying to calculate from the whole memory would be damn slow.
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #5 on: May 09, 2015, 05:39:46 AM »
calculating based on random samples gives the wrong answer. This is useless.
I would take slow vs useless anytime. Tektronix's one takes 5-10 seconds (not minutes), nothing wrong with it

Please don't tell me "just use the tektronix then". I believe this is an unacceptable bug, doesn't matter how cheap the scope is.

 I might contact Rigol about a firmware update on the issue
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 9203
  • Country: nz
    • Taupaki Technologies Ltd.
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #6 on: May 09, 2015, 06:09:44 AM »
trigger is set to edge, auto sweep, dc coupling.
As it should be, but "level" can make a difference on how the scope calculates a result.

Screen shot required.
Avid Rabid Hobbyist & NZ Siglent Distributor
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #7 on: May 09, 2015, 06:43:37 AM »
trigger level did not affect the rms measurement

see attached screenshot for a specific trigger level, at 3 different time base length.

60Hz sinusoid, 1V rms from function generator

thanks
 

Offline Sailor

  • Regular Contributor
  • *
  • Posts: 170
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #8 on: May 09, 2015, 10:01:01 AM »
From those screen shots I'd say that it's pretty obvious that the cause is what I suggested the other day (on that other forum), and which others are suggesting above i.e. the Rigol is only using the points that it has displayed on the screen for its computations. The total screen width is only 800 points. The waveform display width is only approx 600 - 640 (just eyeballing it). In your second screen shot I can count ~ 10 cycles per division, so in other words there are only 60 - 64 screen points for 10 cycles i.e. ~6points per cycle. It's inevitable that the computation will be wrong. Look at the first screen shot, I think there's approx 25 cycles in one division i.e. 2.5 screen points per cycle. That shot naturally shows the greatest error in the computation. The second has less error, and in the third shot, where you have ~ 60 screen points per cycle, delivers an almost exact value.

PS Closer scrutiny indicates ~12 cycles per division in the second shot, and ~30 cycles per division in the first shot - essentially right on the absolute Nyquist limit.
« Last Edit: May 09, 2015, 10:04:40 AM by Sailor »
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #9 on: May 09, 2015, 10:17:27 AM »
lol, small world.

yes, I believe this is what is happening. It doesn't make any sense to implement a function based on the displayed points (since the points are available in memory). Hopefully I can get a firmware update about this. If Rigol cares, fixing this should be very easy

I'll keep posting here in case of a positive outcome

Thanks Sailor (you confused me talking about integer number of cycles, as this was not the case)
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 1113
  • Country: de
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #10 on: May 09, 2015, 04:35:57 PM »
Hopefully I can get a firmware update about this. If Rigol cares, fixing this should be very easy

Yes, I'm sure they are already on standby for some firmware-on-demand for you...

Honestly, I think Rigol's current implementation has its justification. The one-button measurements provide quick, real-time quantification of what's on the screen. Because they are fast, they are useful for a quick check of the signal, as well as for interactive adjustment of signal levels, frequencies etc.. A measurement that takes 5 to 10 seconds to update would be annoying for the first use case, and useless for the second one.

I, too, have sometimes wished for an alternative, in-depth analysis mode that applies to the complete memory (or at least to the signal segment shown on the screen, but using the full sampling resolution from memory). But that is an alternative, maybe additional mode, in my view -- not "the right way to do measurements", with Rigol's current implementation being wrong. We are talking "feature wish" here, not "bug". If I had to choose just one of the two modes in my scope, I would go for the current, fast version any time.

And, let's face it: The DS1000Z series are budget, entry-level scopes. Rigol needs to cut some corners to keep the hardware cost down while still providing decent interactive responsiveness. Plus, they would be stupid to give you "everything" in the low-cost model, while still having more expensive instruments in their portfolio...
« Last Edit: May 09, 2015, 04:48:25 PM by ebastler »
 

Offline Armxnian

  • Regular Contributor
  • *
  • Posts: 214
  • Country: us
  • Computer Engineering Student
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #11 on: May 09, 2015, 05:15:02 PM »
As mentioned above, the values are calculated by what is displayed on the screen. This can be confirmed by many different tests. Input a 5v signal, and set your scope to 100mv/division, and use the Vmax function. It will display stars because it is unknown. If it displayed 5v, it would be accurate, but it wouldn't make sense, because you're not looking at 5 volts. You can try the frequency measurement as well. Input a sine wave and turn on the hardware counter and the software freq function. At a long enough time base, both will display the same value. But shorten the time base so the displayed wave is less than 1 period, and you will see the stars again. It is more logical (on an oscilloscope) for the value to reflect what is displayed.

Some settings also reflect more accurate values than others. Input a 1v signal, and observe it at relatively close settings, like 500mv/div, 1v/div, 2v/div. The software function for Vmax will display 1v for all those settings. But try going up to 10v or higher (x10 probe setting). Even though the signal is still visible on screen, the numerical value will not be 1v. Most likely a limitation of the hardware as to a firmware issue where the hardware isn't communicating correctly with the software.

Of course this discrepancy is easy to detect because you know what the actual value of the input signal is, but on the other hand you are looking for values that are incorrect. I've found that more vertical and horizontal settings give the accurate value compared to the wrong value. So if 5 different time bases give you 500rms, and 1 gives you 750rms, then the 5 are sure to be more accurate. You can most certainly find the correct value with quick tweaking.
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #12 on: May 13, 2015, 04:37:34 AM »

Yes, I'm sure they are already on standby for some firmware-on-demand for you...

Most probably not. Or maybe they are. The rep said he will pass my feedback to the eng team. Might be saying it just to make me stop bug him or maybe because my feedback is not bad.




And, let's face it: The DS1000Z series are budget, entry-level scopes. Rigol needs to cut some corners to keep the hardware cost down while still providing decent interactive responsiveness. Plus, they would be stupid to give you "everything" in the low-cost model, while still having more expensive instruments in their portfolio...

I am not sure that in their high level ones they do the measurements differently.

My opinion is that if people are checking a 60hz signal with 6M points, on a long time scale, they are using the scope wrong, and if in that case the measuremnt takes a long time, it is their fault. If you are checking a fast changing waveform (repetitive, not slowly changing like in a PFC), than you need the resolution, but you do not need the long time scale, so the total number of points you need can be kept low; also in this case the measurement will not take long if scope is set to a reasonable resolution/time scale for the application.

The only case in which I can think (my application, might be missing a lot of other cases) you might need a long time scale (for the low frequency) with high resolution (to capture the details of the high frequency) is something like a PFC. In that case, speed of calculation is secondary compared to having something that makes sense
 

Offline alsetalokin4017

  • Super Contributor
  • ***
  • Posts: 1563
  • Country: us
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #13 on: May 16, 2015, 03:56:57 PM »
Here's an example of why you might want the Left Menu measurements to use only the displayed screen data rather than the whole buffer. These are two shots of the same gated signal, one at 2 ms/div and the other at 500 ns/div. Note that the Hardware Frequency Counter at top right operates on the entire buffer. The cursor measurements give the correct frequencies by taking the inverse of one period of the displayed signal, either the gated blocks or the fine oscillations inside the blocks. The Hardware counter cannot report an accurate frequency of this higher frequency signal, but the Left Menu frequency measurement operating on the screen display data agrees with the cursor measurement.
The easiest person to fool is yourself. -- Richard Feynman
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #14 on: May 19, 2015, 02:40:04 AM »
I think we are talking about two different things.
I do not think it should use the entire memory buffer. I think it should use all the points of what is visualized, or at least offer the option of choosing.
The example you are showing, (correct me if I am wrong), you are zooming (changing the time scale) into a part of the acquired buffer. If the measurement (left menu, same as you are using) was based on all points of the displayed data, it will always give you the correct result. If it is too slow, in case not all points are needed for your application, you can change the number of points being acquired.
It do not think (and I don't know if there are any application in which it would be useful) that it should give a measurement of the entire buffer even if you are looking at a small part only

thanks
« Last Edit: May 19, 2015, 03:40:41 AM by ziopaper »
 

Offline ziopaper

  • Contributor
  • Posts: 8
Re: rigol DS1054Z wrong rms measurement on long waveform
« Reply #15 on: May 19, 2015, 02:41:39 AM »
btw, I checked with the Rigol rep, and he confirmed that also the higher end Rigol scope implement the measurements in the same way, so it is a design decision (without thinking of PFC apparently), not a matter of entry level oscilloscope
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf

 

http://opalkelly.com/