Author Topic: Help please, understanding Rigol1054z DC RMS measurement.....  (Read 7257 times)

0 Members and 1 Guest are viewing this topic.

Offline mtolesenTopic starter

  • Contributor
  • Posts: 13
  • Country: dk
Help please, understanding Rigol1054z DC RMS measurement.....
« on: December 12, 2016, 08:18:28 pm »
Hello all,

Properly a rookie question, but i am confused!, just got my Rigol1054Z scope, and i am experimenting with a mix of ac and dc signals.

When i use the scope to measure a dc voltage, on ea channel 1, i adjust sensitivity to 1volt/div  and then on the left side of the screen, i choose measurement (from the row of bottums on the left side of the screen) to show me RMS voltage fron channel 1 in the little displays on the lower side of the screen. If i connect my channel probe directly to my dc power supply, and adjust to, lets say, 3 volt, and double check with my digital voltmeter, i always have a little higher voltage on the scope, compared to the DVM. Then, i disconnect the scope probe completely from the scope, and change input on channel 1 from DC, to GND. This should internally connect input on channel 1, to GND right?. the strange thing is that RMS measurement, on the little display reading on the scope do  not drop to 0,00 volt as it should, it only drops to around 30 mV and freeze, sometimes even only around 60 mV, but not 0 volt. In other words,
If i just disconnect the propes, a small voltage og around 30-40 mV is shown in the Rms reading, and also if i reconnect the probe and zero it by connecting GND to the hot tip of the probe.
It is like i always will have a small DC voltage offset on my inputs even if they are fully grounded, is that really correct, can i somehow compensate for this? I have tried the scope calibration in the utility menu, but it did not help.

I have read somewhere to use Per.Vrms instead, but this measurement seems not to be able to measure DC.

It seems soooo complicated to measure a simple DC value correct on this scope, can that be right, why can i not "zero" my RMS reading?

br
Martin / denmark
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #1 on: December 12, 2016, 08:28:18 pm »
The DS1054Z doesn't have a true "GND" connection. Measurements taken in that mode still have noise.

Why would you want RMS of a DC signal? It's meaningless, the square root of the square of a number.  :-//


It seems soooo complicated to measure a simple DC value correct on this scope, can that be right, why can i not "zero" my RMS reading?

The specification for DC voltage accuracy on a DS1054Z is 5%. 60mV error on a 3V signal is well inside that.

Oscilloscopes are for looking at the shapes of things, not making very accurate voltage measurements. Use a multimeter for that.

 

Offline mtolesenTopic starter

  • Contributor
  • Posts: 13
  • Country: dk
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #2 on: December 12, 2016, 09:17:31 pm »
Hello Fungus,

Thank you for the answer, and i fully understand what you are saying.

My reason for my question is that i, as a part of a larger circuit, are experimenting with a RMS to DC converter chip (AD536) and i am making my test measurements by measuring the RMS value of the AC input signal to the AD536 on channel 1 og the Rigol, and then i am using channel 2 of the Rigol to look at the corresponding DC value, coming out of the AD536. For that i look of the RMS values of both channels, and it was in that regard, o noticed this dc offset. I do know i can use a good old DC Multimeter on the DC output of the AD536, but i just wondered why i had this stupid offset.

Thank you for the answer.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #3 on: December 13, 2016, 12:16:53 am »
Why would you want RMS of a DC signal? It's meaningless, the square root of the square of a number.

Not only is it not meaningless, it is also easy to calculate - and is a good test case to see if the instrument and your understanding are working in perfect harmony :)

Quote
Oscilloscopes are for looking at the shapes of things, not making very accurate voltage measurements. Use a multimeter for that.

Depends on the frequencies and crest factor, of course.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: mtolesen

Offline mtolesenTopic starter

  • Contributor
  • Posts: 13
  • Country: dk
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #4 on: December 13, 2016, 08:30:59 am »
Hello tggzzz,

Thank you for the supporting words, and actually you are completely right. For me to start with a complicated instrument (i have been away from designing electronic for around 20 years) like the Rigol 1054z has been a challange, so i need to start from scratch, and do my learning from practical expirements. So your sentence "see if the instrument and your understanding are working in perfect harmony" makes completely sense.
« Last Edit: December 13, 2016, 08:32:34 am by mtolesen »
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #5 on: December 13, 2016, 09:39:53 am »
One thing I forgot: DSOs have a limited number of bits of resolution in their ADC (8 in your case).

To get the best possible readings you need to make the waveforms as tall as possible on screen, that way you're using more of the ADC's range.

 
The following users thanked this post: maelh

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #6 on: December 13, 2016, 10:10:57 am »
Hello tggzzz,

Thank you for the supporting words, and actually you are completely right. For me to start with a complicated instrument (i have been away from designing electronic for around 20 years) like the Rigol 1054z has been a challange, so i need to start from scratch, and do my learning from practical expirements. So your sentence "see if the instrument and your understanding are working in perfect harmony" makes completely sense.

"Coming back" to electronics is fun and worthwhile. I'm simultaneously appalled and delighted at how few of the fundamentals have changed in the last 20/30 years. The main changes have been that the same old stuff (especially C, processors, debugging) is much cheaper and smaller, nanopower circuits, and ADC/DAC speed/resolution.

Don't be put off by SMD components; I prefer them to through hole. See the blog in my .sig for my experiences.

Oh, keep an eye out for David Hess' recent posts on the ds1054z. He mentions a mechanism by which its post-processing measurements can be deceptive. I don't have a 1054, so I can't comment, except to say that if they are correct then the 1054's measurements are less useful than people would like you to believe.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Online PA0PBZ

  • Super Contributor
  • ***
  • Posts: 5121
  • Country: nl
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #7 on: December 13, 2016, 10:11:27 am »
Simply said: With the sensivity set to 1V/div and having 8 divs on your screen full screen will be 8000 mV. Divide that by 256 (8 bits) and your resolution will be about 30 mV.
Keyboard error: Press F1 to continue.
 
The following users thanked this post: maelh

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #8 on: December 13, 2016, 10:16:49 am »
Simply said: With the sensivity set to 1V/div and having 8 divs on your screen full screen will be 8000 mV. Divide that by 256 (8 bits) and your resolution will be about 30 mV.

And here's the specifications for gain/offset of the signal after you twiddle the knobs:


 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #9 on: December 13, 2016, 10:25:48 am »
Oh, keep an eye out for David Hess' recent posts on the ds1054z. He mentions a mechanism by which its post-processing measurements can be deceptive. I don't have a 1054, so I can't comment, except to say that if they are correct then the 1054's measurements are less useful than people would like you to believe.

Stop scaremongering. His 'discovery' is a curiosity at best and isn't even applicable to most of the readings (eg. RMS - the reading being discussed here).

PS: Welcome to the forums, a place where certain people will endlessly post about how everything the DS0154Z is bad/wrong. But it isn't true, they usually just don't understand their instrument.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #10 on: December 13, 2016, 10:28:54 am »
Oh, keep an eye out for David Hess' recent posts on the ds1054z. He mentions a mechanism by which its post-processing measurements can be deceptive. I don't have a 1054, so I can't comment, except to say that if they are correct then the 1054's measurements are less useful than people would like you to believe.

Stop scaremongering. His 'discovery' is a curiosity at best and isn't even applicable to most of the readings (eg. RMS - the reading being discussed here).

PS: Welcome to the forums, a place where certain people will endlessly post about how everything the DS0154Z is bad/wrong. But it isn't true, they usually just don't understand their instrument.

I'm not scaremongering, as the caveats in my statement indicate.

There are also people that endlessly post about how everything about the ds1054z is good/correct.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16561
  • Country: 00
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #11 on: December 13, 2016, 11:28:43 am »
I'm not scaremongering, as the caveats in my statement indicate.

Did you provide a link to the discussion? Did you mention that it's limited to a particular type of measurement? Explain what the problem is or where/when/why it happens?

Noooo... you just said something vague and scary-sounding with no way for the readers to find out more information.

In reality the great new "discovery" being discussed over there is no worse than the fact that all vertical measurements are less accurate if you didn't maximise the waveform height on screen before taking them.

Every single voltage reading!! Can you believe that?? What a scandal!!!

Or...maybe we can just understand what an 8-bit ADC is and how it works.

Understand your instrument, that's the real trick.

 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #12 on: December 13, 2016, 11:57:08 am »
I'm not scaremongering, as the caveats in my statement indicate.

Did you provide a link to the discussion? Did you mention that it's limited to a particular type of measurement? Explain what the problem is or where/when/why it happens?

Noooo... you just said something vague and scary-sounding with no way for the readers to find out more information.

In reality the great new "discovery" being discussed over there is no worse than the fact that all vertical measurements are less accurate if you didn't maximise the waveform height on screen before taking them.

Every single voltage reading!! Can you believe that?? What a scandal!!!

Or...maybe we can just understand what an 8-bit ADC is and how it works.

Understand your instrument, that's the real trick.

Sigh.

The OP managed to read, understand, appreciate and quote my earlier comment in https://www.eevblog.com/forum/testgear/help-please-understanding-rigol1054z-dc-rms-measurement/msg1090290/#msg1090290

It is revealing that you chose to snip and therefore avoid addressing my second sentence in https://www.eevblog.com/forum/testgear/help-please-understanding-rigol1054z-dc-rms-measurement/msg1090341/#msg1090341 I doubt the OP will make that mistake.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 
The following users thanked this post: mtolesen

Offline MrWolf

  • Regular Contributor
  • *
  • !
  • Posts: 209
  • Country: ee
Re: Help please, understanding Rigol1054z DC RMS measurement.....
« Reply #13 on: December 13, 2016, 12:26:56 pm »
1) Do you use probe and channel on 10X or 1X setting? You need 1X. But be aware - probe switch has been reported defective time to time.

2) Use cursors, not auto measurements - cursors may be more precise on Z. Example:
Manually tune zero position of CH2 with vertical control to graticule zero. Set CH1 to same vertical range. Set cursors source to CH1. Move cursors according to levels on CH2. Now you have right reading in cursors box. Be aware of temeprature etc related drifts over time.

3) Use math channel and cursors. Since I failed to find a way just do display channel 1:1 on math channel (is there any?) I used "bogus operation" with results attached on screenshots. Basically same technique as in example 2.
Tried to calib for zero and then measure 3 millivolts.
Source is 16bit signal gen.

Agilent U1272A, delta-zeroed on same source:
0mVDC => -0.005mVDC
3mVDC => 2.979mVDC

To get rid of the noise:
[Acquire]
Mode=Average
Averages=1024
MemDepth=12M

Just ot get CH2 1:1 on Math channel using "bogus op"
[Math]
Operator=A+B
Operation=ON
SourceA=CH2
SourceB=fx
Offset=-%your_problem%
Scale=%CH2_scale%
=>[Options]
fxOperator=A-B
fxA=CH2
fxB=CH2

[Cursor]
Mode=Manual
Source=CH2

Move cursors according to Math channel line, not CH2.

4) Using same stuff as in before examples and taking some general ideas from
this appnote:
https://www.picotech.com/library/application-note/using-analog-offset-to-maximize-oscilloscope-resolution
You could float your whole DUT and get some precision DC voltage source. Tie them to same "virtual ground".
Measure voltage from + of precision source to + of DUT. Exact circuit may not be that trivial. Less the the diff between "virtual ground" and + of your DUT, smaller the vertical scale used - higher precision.
Just random idea: If you make 1:10 voltage divider and power it with your average PSU. Then voltage over 1 part will be 10x less than over PSU, but absolute adjustment precision will be 10x higher.

5) Use potentiometric voltage measurement
http://www.allaboutcircuits.com/textbook/direct-current/chpt-8/voltmeter-impact-measured-circuit/

6) Apply voltage from precision source to CH1, make math channel CH1+CH2. Tune DC error out of math channel using precision source.

And so on... So even with crude measurement instrument you can go high precision when taking advantage of basic physics...

PS. It was just an amateur brainstorm with cup of coffee, so EE pros may correct at any time...

« Last Edit: December 13, 2016, 01:51:27 pm by MrWolf »
 
The following users thanked this post: mtolesen


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf