Measuring DC component on the AC line with multimeter and/or oscilloscope

[Gunders]

Hi,
I have a problem with a transformer that is humming quite loud now and then. Out of curiosity I started logging the mains voltage with a multimeter and an oscilloscope, I logged the voltage for 48 hours.

The problem is that the results from the multimeter and oscilloscope are quite different. The multimeter says that the DC voltage has been between -1,067 V and +1,176 V, the oscilloscope says the average voltage has been between -3,7258 V and -2,0095 V (I have even measured AVG voltage down to -4,5 V). I have problems interpreting those results. First, I have problems understanding why they are so different. My mains voltage is quite distorted, is that the reason why I get different results for the multimeter and oscilloscope? Is it right to use the oscilloscope’s AVG measurement to find the DC voltage on an AC voltage, even if the sine is distorted and asymmetric? Are there other or better techniques I could use to measure the DC component on the mains? 

The multimeter and oscilloscope I have used:
https://www.keysight.com/en/pdx-2891615-pn-34461A/digital-multimeter-6-digit-truevolt-dmm?cc=US&lc=eng
https://www.keysight.com/en/pcx-x205209/infiniivision-4000-x-series-oscilloscopes?cc=US&lc=eng

I have attached screenshots from the multimeter and oscilloscope.

[TimFox]

In a different application, with a very high peak to average ratio (pulsed positive-definite waveform), I found that different DVMs gave different results for the DC voltage, presumably because of the peak or AC overloading the DC measurement.  In your case, you have 240 V rms (680 V pk-pk) superimposed on your 1 to 5 V dc level.  You might want to add a serious low-pass filter ahead of your DC voltmeter to reduce the 50 Hz component.

[jfphp]

In the HP 8116A op and svce manual / alignment chapter you have to mesure the dc component in an ac ouput of 16V ptp : a 5hz low pass filter is necessairy and shown with 20K and 2,2 µF (fig 9.2) . Take a 250V rms cap model.

[Kleinstein]

There are several errors that can add up:
Possible overloading the DMM in peaks, offset errors from the Scope and linearity limits (mainly for the scope).
For the scope there could also be an error with the trigger, getting the period length right for calculation the DC level.

An extra low pass filter to dampen the AC part could help quite a bit.

[Ice-Tea]

I have to ask...

- How did you measure your AC line?
- Are you aware that if you clip the crocodile of your probe on one the AC lines you have a 1/2 chance of blowing up your probe/scope/self?

[BravoV]

Not using scope though, I did measure DC over AC mains (220V) when running my Philips 1600 watts hair dryer.

Two DMMs, left one Fluke 287 showing DC over AC (you can see the bigger digits DC reading over smaller digits AC reading at the LCD display), while Fluke 87V at the right, was set at DC reading, and also Watt meter.

DC readings both were very similar.

[JohnPi]

The meter will clamp its input (likely asymmetrically) to to protect its circuits if you try to read a larger (240 VAC) on a low voltage DC range. You would need to use a ~ 300 VDC range to be sure the result is correct.

to prove your measurement is wrong, swap the meter leads -- you'll get the same DC result (not the -ve of it).

A transformer can't really generate a DC voltage offset, but certain loads (½ wave rectified) may cause a DC component of the load current which can appear as a DC voltage component across the resistance of the wiring.

[Gunders]

Thank you for your comments and inputs. Highly appreciated.

Not using scope though, I did measure DC over AC mains (220V) when running my Philips 1600 watts hair dryer.

Two DMMs, left one Fluke 287 showing DC over AC (you can see the bigger digits DC reading over smaller digits AC reading at the LCD display), while Fluke 87V at the right, was set at DC reading, and also Watt meter.

DC readings both were very similar.

I have seen your video (thank you for the nice video and test/demonstration) and I used it as a guideline for my measurements. As I can see it, have done the same as you. The only difference is that I have used a benchtop meter instead of a handheld meter. I have a handheld Fluke 177 as well; I will try to do the same measurements later today with the Fluke.

I have to ask...

- How did you measure your AC line?
- Are you aware that if you clip the crocodile of your probe on one the AC lines you have a 1/2 chance of blowing up your probe/scope/self?

Regarding how I have tested this, my test setup.
I have used a 3-outlet power strip with ON/OFF switch. The outlets are “polarized” and the power cords I have used are “polarized”, only one way to plug the cords in the power strip. I have cut a power cord in half and made flying leads for Live and Neutral. The power cord also has a two pole ON/OFF switch to turn ON/OFF the mains voltage for the flying leads.
First, I turned the power strip OFF. Plugged the multimeter, oscilloscope and test cable in the power strip. Next step was to connect the test probes to the flying leads; connecting the probe’s ground lead to Neutral and the probe tip to Live. For the multimeter, Live was connected to HI input and Neutral to LO input. Then, I turned the power strip ON (test cord switch still OFF), turned on the multimeter and oscilloscope, waited for them to startup and finishing their boot sequence. The last step was to turn the test cord ON so there would be live voltage at the flying leads.
The probe for the oscilloscope was a 100:1 probe.

The meter will clamp its input (likely asymmetrically) to to protect its circuits if you try to read a larger (240 VAC) on a low voltage DC range. You would need to use a ~ 300 VDC range to be sure the result is correct.

to prove your measurement is wrong, swap the meter leads -- you'll get the same DC result (not the -ve of it).

A transformer can't really generate a DC voltage offset, but certain loads (½ wave rectified) may cause a DC component of the load current which can appear as a DC voltage component across the resistance of the wiring.

I am not quite sure whether I understand your comment about the meter clamping the voltage. I have set the multimeter to use the 1000 V range. If I used the 100 V range the meter will only report Overload.

[Ice-Tea]

Regarding how I have tested this, my test setup.
I have used a 3-outlet power strip with ON/OFF switch. The outlets are “polarized” and the power cords I have used are “polarized”, only one way to plug the cords in the power strip. I have cut a power cord in half and made flying leads for Live and Neutral. The power cord also has a two pole ON/OFF switch to turn ON/OFF the mains voltage for the flying leads.
First, I turned the power strip OFF. Plugged the multimeter, oscilloscope and test cable in the power strip. Next step was to connect the test probes to the flying leads; connecting the probe’s ground lead to Neutral and the probe tip to Live. For the multimeter, Live was connected to HI input and Neutral to LO input. Then, I turned the power strip ON (test cord switch still OFF), turned on the multimeter and oscilloscope, waited for them to startup and finishing their boot sequence. The last step was to turn the test cord ON so there would be live voltage at the flying leads.
The probe for the oscilloscope was a 100:1 probe.

I would strongly suggest to not ever use your scope like that again. I'm not sure about Norway, but in large parts of the world there is no mandatory standard for assigning live/neutral to pins. Even if there were, I would not trust my gear or my life to the goodwill of an electrician or whatever hack that came before me.

Not to mention that even if you're lucky enough to *not* get it wrong, a small voltage can exist between the Neutral and earth. You don't want that equalized trough your probe leads.

Just... don't do that.

[Gunders]

I would strongly suggest to not ever use your scope like that again. I'm not sure about Norway, but in large parts of the world there is no mandatory standard for assigning live/neutral to pins. Even if there were, I would not trust my gear or my life to the goodwill of an electrician or whatever hack that came before me.

Not to mention that even if you're lucky enough to *not* get it wrong, a small voltage can exist between the Neutral and earth. You don't want that equalized trough your probe leads.

Just... don't do that.

Thank you for educating me.

You are right, Line and Neutral could be mixed multiple places, even though my fuses are marked with L1 and N and electricians are taught to be consistent with L and N.
I think it would have tripped my ground fault interrupter (RCD) ff I had a potential between N and E and shorted those two from my probe GND lead and through the oscilloscope. Here where I live RCDs became mandatory in 1975 for new buildings, and then for all buildings in 2008.

But as you say, in general a bad idea to use an oscilloscope that way. I wasn’t really aware.

[Gunders]

I think I have found out of this. Using the oscilloscope to measure the average voltage isn’t a good way to measure the DC offset when you have a high AC voltage and low DC offset. I have access to an adjustable AC source, and I have tried three cases:
•   5 VAC_RMS + 2 VDC offset
•   230 VAC_RMS + 2 VDC offset
•   230 VAC_RMS + 0 VDC offset

The multimeter shows 1.90 V for the two first cases and 0 V for the last one. The AVG measurement from the oscilloscope is 1,90 V for the 5 VAC. For 230 VAC + 2 VDC the AVG reading is -0,74 V. I tried some different settings for the oscilloscope, but the AVG readings were still -0,74 V for 230VAC. I got -2,92 V for 230 VAC and no offset, that is basically the same I have measured at home.

My conclusion regarding my measurements is that the readings from the multimeter is correct and the AVG reading from the oscilloscope isn't reliable for high VAC/VDC ratios. My second conclusion is that I should follow Ice-Tea's advice in the future: Don’t use an oscilloscope to measure directly at the mains.

[Tryer]

Hi Gunders,
I hope you are still alive.
For the measurement you wanted to make, there were a lot of right proposals.
As you found with your experiment, overloading the DMM is the main cause for a wrong measurement.
Converting what Kleinstein said in practice and not taking a risk for life, i would suggest to build the following
in front of your DMM.

Choose big size resistors or put eg 5x100K resistors in series on each side(arcing)
and a non polarity capacitor(1uf/16V is ok). The AC on the capacitor will be abt. 2Vpp and the multimeter can cope with.
The resistors will dissipate abt. 50mW both (230^2/1e6 W). The DC reading on a 10Mohm DMM will be
abt. 10% less than the real value. And you can touch the DMM side with no risk.
The maximum current you get, if you touch the "bad side", will be less than 0.5mA and that is safe.
Stay healthy,
Tryer