Channel RMS measurement different from DVM AC RMS Measurement on MDO3024 Scope

I'm using and MDO3024 scope. I am using the measurements function to make RMS measurements of several channels. I also am using the DVM AD RMS function to make measurements of one of the channels.

Under some conditions the DVM and the channel RMS mean agree. (Within 0.5%) Under other conditions they differ. (By 5% to 8%) I am also using an external DMM to make yet another measurement of that channel. It agrees with the DVM under all conditions.

What is the difference in how these measurements are made or calculated? What could account for the difference?

Hi James,

There are countless things that could cause this issue, here are just a few:

1) Are the "true RMS" DMMs? If not, their RMS readings may only be valid for specific waveforms.

2) Are your probes properly compensated?

3) What's the bandwidth of your signal? DMMs don't measure high frequencies well.

I'd need to know more about your signals to be able to help you narrow it down more.

I've attached a screen shot from my scope that shows an example of what I'm seeing.

I am using a function generator to drive current through a helmholz coil. There is a receiver coil inside the helmholz coil, picking up the field. 
In the screenshot:
Channel 4 is the current into the helmholz coil
Channel 3 is the output of a field probe (1 mV / 1 mGauss)
Channel 1 is the output of the sense coil.

The sense coil signal has a lot of environmental noise on it. The channels are all bandwidth limited and the traces are averaged over several hundred waveforms. (Averaging is set to 512, but I am freezing the screen before that number is reached)

The scopes DVM says "Under Bandwidth". Interesting since the driving waveform is 10 kHz. I lower it to 9500 Hz, No change.

Notice that the DVM AC RMS reading and the channel 1 measurement reading do differ by a bit. Since they are using the exact same input signal, I would have expected them to be very close. But apparently the DVM and the MEASURE process the signal differently. I'm curios what the difference is.

I do have an external meter I added to the setup to be an arbitrator. It didn't help.

It appears that your CH1 waveform has some high-frequency noise on it.  This will be included in the channel's RMS measurement, but would likely be filtered away in the DVM AC RMS measurement.  That would likely account for the small 30uV difference between those measurements.

As Alan said, the hardware DVM will have significantly lower bandwidth than the scope, so that's a possible source of error.

Another source would be the fact that you're averaging a noisy signal. which inherently changes your measurement as the noise is part of your signal and counts to RMS too.
I also see some very fast spikes on your averaged waveform, which could be an artifact of trigger jitter+averaging which will affect your measurements.

The DVM measurement is done through hardware similar to digital DMM. It's inherently much lower bandwidth than the scope channel.
It also runs independently of the acquisition hardware, it's a live read of what's on the channel at that moment, not what was on it during the last acquisition.

The other measurements are calculated after the signal is captured by the ADC and processed (for your example, in your case, the waveforms are averaged before measurements are taken) and are calculated again when a new acquisition is displayed.

So there are, at least, 3 ways in which the measurement and DVM systems work that could explain the discrepancies you're seeing: bandwidth, post-processing and triggered/free run.

Makes sense. Thanks.