Why difference between scope measured voltage and DMM voltage

[prwiley]

I recently purchased a RIGOL DS1054Z and have been having fun using it to look at a little CanaKit amplifier that I built awhile ago (https://www.canakit.com/20w-bridged-audio-amplifier-kit-ck193-uk193.html).

I am looking at the voltage across the output (speaker connection) with the scope and my EEVBlog BM235 at various volume levels. One example: With a 60Hz signal a 1.0Vrms from my Rigol DG1022 the scope measures 6.22Vpp and 6.06Vrms. Screen shot of scope is here: https://photos.app.goo.gl/i63DGD5UYSyQN9sz9

My BM235 DMM, EEVBlog edition, is set on V auto is measuring 5.339V.

Seems like a big difference between the two to me, but what do I know, I'm coming back to all this after many years.

Could someone suggest what the source of the difference might be? I not that at any setting of the pot, the DMM is generally between 70 and 75 percent of the scope value.

[Benta]

Try AC coupling your 'scope input.
It's just a switch.

[exe]

yep, I think it's coupling "issue". I see sine amplitude is 7.68-3.18=4.5V. The AC RMS value should be 4.5/sqrt(2) ~= 3.2V. Not sure why it says RMS is 6.06V, I calculate it to be Vac_rms + Vdc = 3.18 + 3.2 = 6.38V .

Anyway, suggestions: try to make waveform taking the whole screen and display more periods if possible. This makes reading more stable. But always use sanity check, often automatically calculated numbers are "incorrect" due to many reasons (e.g., coupling, too low amplitude, noise, incorrect probe attenuation, not enough bandwidth, loading effects, etc).

UP: formula I used is incorrect, see https://electronics.stackexchange.com/questions/240242/using-a-voltmeter-to-measure-rms-value-of-a-sine-signal-with-a-dc-offset

[rstofer]

Unless you have terminated the cable at the scope with a 50 Ohm terminator, make sure that your signal generator is set for Hi-Z output.  If you have terminated the cable in 50 Ohms, make sure the signal generator is set for 50 Ohm output.  This causes a x2 error.

Next, make sure that your scope probe setting (x1-x10) agrees with the scope input setting.  Press the CH1 button briefly and there is a button for Probe 1X versus 10X on the menu.

As discussed above, make sure you use AC coupling.  It is also available on the menu with the Probe setting.

[prwiley]

Thanks everyone,

55Hz, 1Vpp. scope sees 4.48Vpp, DMM 3.544V

Probe agrees with probe attenuation software setting: check
More wave on screen: check
Signal gen output "High-Z": check
Tried AC coupling: check, but:

https://photos.app.goo.gl/awCr2RLSar5pxijB9 
https://photos.app.goo.gl/Hwap1VX3tEU45TDK7

[Benta]

So many issues here, you don't really know what you're measuring.

Your signal is a DC voltage with an AC voltage overlayed.

So what are you measuring?

Your 'scope displays the AC signal correctly when AC coupled: 1.62 Vrms, 4.48 Vpp

Your BM235 measures a DC voltage plus "something", in this case probably 3.44 V plus sine spikes on top, which might result in 3.544 V. It's anyone's guess. I suspect the BM235 is correct, but this would need a Fourier analysis and I don't have time for that.

You're using the instruments incorrectly.

[ogden]

But always use sanity check, often automatically calculated numbers are "incorrect" due to many reasons (e.g., coupling, too low amplitude, noise, incorrect probe attenuation, not enough bandwidth, loading effects, etc).

Right. Let's not forget 8bit ADC of the scope, with DC accuracy +/- 3%. EEVBlog BM235 is specified +/-0.7% at 60Hz.

[rstofer]

On the BM235, you may need to push SELECT a few times to get the sine wave to appear at the left of the numerics with the display as 0.000 V.

[exe]

I wonder if BM235 is capable of measuring Vrms with big DC bias. I think there were issues with some meters. Check this link: https://www.eevblog.com/forum/testgear/brymen-multimeters-fault/ .

For true rms measurements I trust oscilloscopes more as because I sort of know how they do it. For multimeters it's more complicated as they often use a dedicated chips that often have big gaps in specifications. Or specified for a limited range of signals, like there should't be dc bias, or frequency should be within its limit (which is much small than for an oscilloscope), or the signal min and max values should be within 10-90% of the range, etc.

PS scope should be in DC coupling, imho. Just by definition of "true rms". And the definition of true rms is: it's equivalent dc voltage that produces the same power for resistive loads. Clearly DC part of waveform also does the job, so it should be accounted.

UP formula I used above to calculate rms voltage was incorrect, it should be sqrt(Vdc^2+Vac^2): https://electronics.stackexchange.com/questions/240242/using-a-voltmeter-to-measure-rms-value-of-a-sine-signal-with-a-dc-offset

[exe]

Huh, I calculated Vrms manually. Looking at https://photos.app.goo.gl/awCr2RLSar5pxijB9 and using formula sqrt(Vdc^2 + Vac_peak^2/2) I get:

Vdc = 5.68 (=Vmid as measured by the scope)
Vac_peak = Vupper - Vdc = 7.47 - 5.68 = 1.79.
So, the true rms value is sqrt(5.68^2 + 1.79^2/2) = 5.819V, which doesn't really match what the scope calculated... Hmm...

I also found this thread: https://www.eevblog.com/forum/testgear/rigol-ds1054z-wrong-rms-measurement-on-long-waveform/ , but it didn't give anything new.

Anyway, be sure you use latest rigol firmware, just in case . Also let you scope warm up. Or you test board as I see values on different screenshots are different. Or, may be, it's sort of "quantization" error, may be it's better show less periods on the screen.. Scope calculations are sensitive to this.

[vk6zgo]

To start with, forget about the displayed voltage on the 'scope.

Vertically centre the trace with the input signal removed, then on "DC coupled", apply the signal & see whether the ac waveform is symmetrical about the zero line you just adjusted for.

If it is not, you have a DC component to the signal, so RMS will not be simply 0.707 times Vp, & needs to be calculated differently.

If you have such a complex waveform, but want to look at the ac component only, switch your 'scope to "ac coupled".
Your voltage readings should now be what you expect.

If not, as a sanity check go "retro"----- count the squares, then knowing the volts/div, you can determine Vpp.

Half of Vpp is Vp, which you then multiply by 0.707, giving you volts RMS.

I would suggest that you go up a bit in frequency to say, 1kHz.
50 or so Hz may be getting near the bottom of the 'scope pass band in "ac coupled" mode.

[prwiley]

Thanks again for the helpful info.

As will come as a surprise to no one, probably, it turns out that there is nothing wrong with the scope of the meter, but only with my understanding of and expectations about what I was looking at. Exe's comments re possible problems with Brymen meters go me to search and I came across this very helpful video:



"RMS Voltage for Sine and square waves, and why your DMM might not work right!" in which vk6zgo's 0.707 factor is explained.

Since I could reproduce the results in his first example for sine waves, I realized there was nothing wrong with my equipment.

I didn't really understand that BM235, when in AC mode, is always displaying RMS voltage. So, the Vrms in on the scope and the Vin measured on the meter are only going to match when I am sending Vrms out of the signal generator -- and they almost do at 1.000 Vrms out of the DG1022 the scope says 1.0-1.02 Vrms, the meter says 0.986V.

Sending 1.000 Vpp out of the generator, the scope sees 1.20 Vpp and 363 mVrms, which is as it should be and the meter sees 349 mVrms

Somehow I was expecting Vpp in on the scope to match V on the meter, and of course that's the totally wrong expectation. I think I know better now.