New Rigol DS1054Z oscilloscope

[rosbuitre]

Hi
I was doing some measurements and gave me errors in RMS, I began to try and show the results, it is a sine 1Khz (Siglent SDG805) entering in 4 channels and a DMM Keysight 34461A, there are important differences between channels (first image 9%) and pressing AUTO again gets worse (second image 21%) without touching anything
This with the latest firmware (Rigol I sent him last week) self-calibration and reset to factory
Which may be happening?

Regards

[hayatepilot]

Try the same measurements with the vertical to full scale. The accuracy should be better than 3% full scale.
In your example you have a full scale of 4V. 3% of that is 120mV. 

A oscilloscope is not a Multimeter. 

Greetings

[JohnPen]

I have repeated the same set up on my scope and find channel 1 is always very close to the nominal 0.357 my particular meter shows.  However the other 3 channels are between 20 and 30mv higher reading.  This is with 6 cycles displayed.  If you change the TB to display 1 cycle the reading is still very close on CH1 (343mv) but CH 2,3 and 4 become very inaccurate ~ 476mv to 485mv.  However change the display to around 20 displayed cycles and all channels read between 369 mv and 385 mv.  It looks rather as if the best way for accurate readings is to have plenty of cycles displayed!  To be fair in my case CH1 is always very close to accurate 22 displayed cycles shows 367-8 mv, 2 displayed cycles shows 358-9 mv and 1 displayed cycle shows as 342-3 mv.  The motto would appear to be use CH1 for best measurement accuracy in my case or have more cycles displayed.

[ankerwolf]

Hello,
look at the ADC-resolution and the integration-time and filtering (<300kHz) of your DMM and compare this with the 8-Bit-resolution and (not) full-scale display (<8Bit) and 2..20 cycles (~20ms @ 1kHz) of the oszilloscope.
Then you will get the answer.
LG Wolfgang

[JohnPen]

Following on from my previous message using Channels 2 or 3 or 4 individually and using full screen I have a range of 362 mv to 342 mv for 22 plus displayed cycles to 1 displayed cycle.  This is an acceptable accuracy for a general purpose scope.  Other equipment would be used for higher accuracy.  So do not try and measure too many Channels simultaneously perhaps this is the best way forward.  Although my Channel 1 does seem to be accurate, as mentioned previously, with multiple channels displayed.

[rosbuitre]

Thanks for contributions 

Regards

[Fungus]

A oscilloscope is not a Multimeter. 

This.

An oscilloscope is basically looking at the position of the dots on screen for minimum/maximum values. A couple of pixels up/down can be significant.

The dots will always a bit fuzzy because radio waves can easily be picked up due to the high bandwidth of the input circuitry.

[Fungus]

Following on from my previous message using Channels 2 or 3 or 4 individually and using full screen I have a range of 362 mv to 342 mv for 22 plus displayed cycles to 1 displayed cycle.  This is an acceptable accuracy for a general purpose scope.  Other equipment would be used for higher accuracy.  So do not try and measure too many Channels simultaneously perhaps this is the best way forward.  Although my Channel 1 does seem to be accurate, as mentioned previously, with multiple channels displayed.

Does the 'error' change if you re-run the self calibration?

[JohnPen]

I ran the Self Cal before doing those level checks.  I will try another Self Cal later and see if there are any major differences.  Incidentally I did restrict BW to 20 M on all channels to avoid possible higher frequency interference.

[Dago]

Has anyone run in to an issue where the scope does not update the screen at the proper rate?

I don't know what I managed to do or what caused it but I managed to get my DS1074 to have a very low update rate constantly. No matter what the time base was the screen was updated maybe once a second. I went through all the options and made sure I had no decoders etc. on but couldn't solve it. It was also persisting over reboots.

I just actually managed to solve it, pressing the auto button somehow restored the refresh rate. Would still like to know what caused it and has anyone else ran in to this issue.

[JohnPen]

Further checks on Vrms on the Channels before and after a new Self Cal run.

Setup  1volt sine wave at 1khz.  All 4 channels AC coupled 200mv range using x10 probes to common source.
Vertical measurement Vrms set for each channel.  Bandwidth restricted to 20M.

Display size 6 divisions channels superimposed.

Displayed Cycles on Screen           CH1 mv      CH2 mv    Ch 3 mv     CH 4  mv

~50                                                368          365            366            364
~24                                                367          371            370            369
~11                                                368          371            378            377
~5                                                  369          392            392            391
~2                                                  360          427            425            424
~1                                                  344          477            474            473

Selecting Auto  Channels displayed as 4 separate traces.

Readings the same.   

Set Cal re-run.

Re-test all new readings were within 1/2 mv of the original table above.

There was some negative drift of 2-3 mv with time but whether it was the scope or my FY3200S I have no way of telling.

It does however suggest that you need around 20 cycles on screen to display reasonable accuracy for channels 2-4.
However CH1 seems to provide consistent accuracy down to 2 displayed cycles so measure with Ch1 if you want the highest accuracy for a particular signal.

Final comment these reading are all relative I do not have an absolute guaranteed 1 volt sine wave source.  My nominal meter states 357 mv with an accuracy of 1.25%.

[rosbuitre]

Similar result who JohnPen, without recalibration.

[Fungus]

Displayed Cycles on Screen           CH1 mv      CH2 mv    Ch 3 mv     CH 4  mv

~50                                                368          365            366            364
~24                                                367          371            370            369
~11                                                368          371            378            377
~5                                                  369          392            392            391
~2                                                  360          427            425            424
~1                                                  344          477            474            473

Selecting Auto  Channels displayed as 4 separate traces.

Yep. More pixels on screen gives a better chance to see min/max values.

With only one wave visible the noise is significant.


The really interesting thing here is that channel 1 is more accurate. I'm not sure why... (luck?)

[rosbuitre]

Hi
As before and 1 click in Vertical scale CH1

Regards

 

[Orange]

Hi
As before and 1 click in Vertical scale CH1

Regards

Channel 1 is clipping on the top screen ??, move it down a bit and see what happens to the RMS value.

For correct AC RMS measurements always use AC coupling, else you can be deceived by small DC offsets on your signal.

[rosbuitre]

Hi
As before and 1 click in Vertical scale CH1

Regards

Channel 1 is clipping on the top screen ??, move it down a bit and see what happens to the RMS value.

For correct AC RMS measurements always use AC coupling, else you can be deceived by small DC offsets on your signal.

Hi, move it down channel 1 and not change the value (not 'clipping').
All channels setting with AC coupling.

[Samogon]

That easily could be proprietary.

What do you mean? You can distribute proprietary GPL'ed software? Never heard of it.
Or did you mean something else?

It could be QNX for example. Non GPL unix like OS.

[alsetalokin4017]

The scope computes the measurements only from the data that is displayed on the screen. So if you have partial cycles at the ends of your traces this will affect the measurement values. This is why (to a point) more cycles on the display gives better measurement accuracy.

However there really isn't any excuse for the "Pluses" counter to tell you that you have 5 "Pluses" when you actually have six showing.

[JohnPen]

Please note from the DS1054Z 'freeze' messages my checks regarding the pulse counting accuracy.  See below.

I can duplicate *4017 results for the 'Pluse' counting.  However I have found that it only occurs when Acquire 'Average' is set.  All other Acquire modes show the count correctly.  There appear to be 2 ways of achieving an accurate count either only use 2 displayed channels alternatively if needing all 4 channels DO NOT use Acquire mode set on 'Average'.  I have tried up to 38 pulses on the Display rather sad!

John

[timofonic]

That easily could be proprietary.

What do you mean? You can distribute proprietary GPL'ed software? Never heard of it.
Or did you mean something else?

It could be QNX for example. Non GPL unix like OS.

I don't think they would license a RTOS for this, maybe I'm wrong. They would use an Open Source one.

They can use the so called "TIVOization" approach, despite they can provide GPL sources. Did anyone looked into the firmware to see what's there?

[Terry Lingle]

I am having trouble downloading the update file on a dial up connection  it gets to about 30%  then fails.
I would like to know if others have this issue and if the file can be acquired from a different site .
Most downloads resume when i click them after a failure but the Rigol site.

[Fungus]

I am having trouble downloading the update file on a dial up connection  it gets to about 30%  then fails.
I would like to know if others have this issue and if the file can be acquired from a different site .
Most downloads resume when i click them after a failure but the Rigol site.

http://www.freedownloadmanager.org/

[nrxnrx]

Hi!

Please consider not installing SP2 (the latest firmware). It has made my (and others') scope sluggish. Unless the release notes ( http://beyondmeasure.rigoltech.com/acton/ct/1579/p-0019/Bct/-/-/ct7_0/1?sid= ) show something you urgently need, I'd just install SP1 and wait.

Try the URLs below. I'll keep them up for at least a week.
It's a semi-vanilla apache server, so everything should work as expected.
To everyone else: that server is limited to ~30Mbits/s upload, but use as you please.

https://hamsterpower.org/f/rigolfw/4_03SP1.zip
https://hamsterpower.org/f/rigolfw/4_03SP2.zip
or
https://hamsterpower.org/f/rigolfw/4_03SP1.tar.xz
https://hamsterpower.org/f/rigolfw/4_03SP2.tar.xz


1,936,095 bytes - 4_03SP1.zip
1,927,701 bytes - 4_03SP2.zip
1,879,484 bytes - 4_03SP1.tar.xz
1,870,804 bytes - 4_03SP2.tar.xz

--
Mihai

[c4757p]

That's... an interesting glitch my 1kZ picked up in the middle of otherwise reasonable operation.

What the hell?

[Terry Lingle]

Thanks nrxnx  I got both files  and will follow suggestion  to not install SP2
Awsome board   
thanks all