rigol ds1054z inconsistent measurement from channel 1 only

[crazyhog]

Hi all,

I just brought my first oscilloscope, the Rigol DS1054z.  I notice channel 1 is giving low Vrms readings.  To test, I hooked up all 4 channels to the same AC source.

ch 1: 9.9 vrms
ch 2: 11.2 vrms
ch 3: 11.2 vrms
ch 4: 11.3 vrms

The difference between channel 1 and other channels seems excessive.  I've verified the probes by swapping them; the probes are fine.  I've also ran the self calibration multiple times.

I also noticed when I increase the horizontal scale from 5ms to 20ms, channel 2,3 and 4 drops to 10.3 vrms.  Why does changing horizontal time scale changes vrms result?

Am I right the difference is excessive?  Should I be concerned that my scope is defective.  I am new to working with oscilloscopes and EE.  Your help is sincerely appreciated.

[alsetalokin4017]

The DS1054z has known "issues" relating to the RMS measurements. These seem to be related to the latest firmware version.

So... yes, and no. Yes, your scope is behaving "normally" for what it is. No, one would indeed expect the RMS readings to be both accurate and consistent across channels, so it's a bug, probably in the software.

See my screenshots below.

Changing the horizontal scale puts more cycles on the screen, and the scope computes the left-menu measurements from the displayed screen data. So, within limits, the more complete cycles you have displayed the more accurate the measurements will be.

You may also want to experiment with the "Per.VRMS" or Period RMS measurements, which will just use a single full cycle to compute the RMS. See the third shot below.

In short.... there are problems and issues with the RMS measurements on the DS1054z running the latest firmware, and perhaps going back to earlier firmwares as well.

[crazyhog]

Thank you very much!  Rigol really should fix this bug.  I was seriously contemplating returning the oscilloscope.  For new guys like me, we don't have the experience to differentiate between serious defects vs minor software bugs.

Thanks again!

[JFJ]

... the scope computes the left-menu measurements from the displayed screen data...

In which case, wouldn't making better use of the scopes vertical dynamic range, by reducing the V/div settings, improve the accuracy of the readings?

For the setup pictured in the first post, each channel could be changed from 20 V/div to 5 V/div. That would greatly improve the vertical resolution (though, it may make the screen appear to be more cluttered).

[wraper]

Thank you very much!  Rigol really should fix this bug.  I was seriously contemplating returning the oscilloscope.  For new guys like me, we don't have the experience to differentiate between serious defects vs minor software bugs.

Thanks again!

They released a new firmware about 9 months since it was introduced by previous one and no fix. Bug exists for almost a year and no possibility to downgrade. Actually CH1 is measuring correctly, all others are not. If you switch off CH2, then CH3 will measure "correctly" too. Basically they screwed up something so data from the previous channel somehow "leaks" into the next one.

[Fungus]

For nice regular waveforms like yours there's another measurement called "Per.Vrms" which gives the correct result. "Per.Vrms" looks for a single, complete wave on screen and calculates the RMS of that (ie. from one zero-crossing point to another zero-crossing point on screen).

The "RMS" reading is a bit misleading because it looks at the visible data on screen and calculates an RMS value from that. This gives an approximately correct result when there's a decent amount of waves visible on screen or for things like random noise, but ... it's very easy to have (eg.) 1.25 waveforms on screen. In this case the "RMS" value will still be incorrect, even after Rigol fixes the bug(!)

Lesson for the day: You need to know what complex instruments like modern DSOs are actually doing, don't just blindly accept numbers you see on screen based on what you think they should to be doing.

[Fungus]

... the scope computes the left-menu measurements from the displayed screen data...

In which case, wouldn't making better use of the scopes vertical dynamic range, by reducing the V/div settings, improve the accuracy of the readings?

In general: Yes.

Always make the wave as tall as possible on screen when decimal places in the numbers are important.

DSOs aren't really voltage measurement devices. They're approximately correct, sure, but they're mostly for looking at the timing and shapes of waves. Voltage values on screen are only about 5-10% accurate unless you spent a lot of money (and have a calibration certificate to prove it).